1 INFO-VAX	Fri, 08 Nov 2002	Volume 2002 : Issue 618       Contents:9 Re: A problem telnet to OpenVMS box and run cobol program 9 Re: A problem telnet to OpenVMS box and run cobol program 9 Re: A problem telnet to OpenVMS box and run cobol program  Re: Advanced Server and UAF  Re: Advanced Server and UAF 5 Re: Alpha network performance issues over leased line 5 Re: Alpha network performance issues over leased line 5 Re: Alpha network performance issues over leased line 5 Re: Alpha network performance issues over leased line  ANN: Updated FTP_MIRROR % Re: Cluster load balancing via TCP/IP % Re: Cluster load balancing via TCP/IP % Re: Cluster load balancing via TCP/IP % Re: Cluster load balancing via TCP/IP % Re: Cluster load balancing via TCP/IP  DECtalk error codes  Re: Email Filter Re: HP Advocacy Site Re: HP Advocacy Site Re: HP Advocacy Site Re: HP Advocacy Site Re: HP keyboard problems RE: HP keyboard problems Re: HP keyboard problems Re: HP ServiceP Re: I'm looking for a Digital RRDxx CD ROM (near BELGIUM). (+VAX 6000-6xx cards P Re: I'm looking for a Digital RRDxx CD ROM (near BELGIUM). (+VAX 6000-6xx cards + Re: Last call for C-Kermit 8.0.206 binaries % Re: mouse with a scroller for OpenVMS  Re: OSU server cache Re: OSU server cache+ Re: OT: The lighter side of VMS Engineering # set time/cluster - what does it do?   Re: Slow performance with BACKUP4 The advocacy site's instapol question is VMS related Re: VAX in comp.risks ...  Re: VAX in comp.risks ...  Re: VAX in comp.risks ... ( Re: Vax Macro Porting - Still Struggling( Re: Vax Macro Porting - Still Struggling( Re: Vax Macro Porting - Still Struggling( Re: Vax Macro Porting - Still Struggling( RE: Vax Macro Porting - Still Struggling( RE: Vax Macro Porting - Still Struggling( RE: Vax Macro Porting - Still Struggling: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT> Re: VMS machines behind DSL router---do I need PAT or just NAT( Re: VMS on Itanium not mission critical? VMS-6.1  Re: VMS-6.1  Re: VMS-6.1  Re: VMS-6.1  Re: VMS-6.1  Re: VMS-6.1 $ Re: Was OpenVMS left out on purpose?$ Re: Was OpenVMS left out on purpose?$ Re: Was OpenVMS left out on purpose?$ Re: Was OpenVMS left out on purpose?$ Re: Was OpenVMS left out on purpose? Re: XP1000 hardware problem  Re: XP1000 hardware problem   F ----------------------------------------------------------------------  # Date: Thu, 07 Nov 2002 21:01:45 GMT . From: Joseph Norris <jozefn@newbolt.sonic.net>B Subject: Re: A problem telnet to OpenVMS box and run cobol programH Message-ID: <Pine.LNX.4.40.0211071256110.26794-100000@newbolt.sonic.net>   JF,   = Thank you for adding so much to my education on this subject.  I will put this to the test.  E Thanks again. You have been very helpful. I will let you how it goes.     ; #Joseph Norris (Perl - what else is there?/Linux/CGI/Mysql) I print @c=map chr $_+100,(6,17,15,16,-68,-3,10,11,16,4,1,14,-68,12,1,14,8, C -68,4,-3,-1,7,1,14,-68,-26,11,15,1,12,4,-68,-22,11,14,14,5,15,-90);     # On Thu, 7 Nov 2002, JF Mezei wrote:    > Joseph Norris wrote:F > >     $t->login(Name=>$username, Password=>$passwd, Prompt=>'/\$/'); > > print "at dir cmd\n"; % > >     $t->print("run testcob.exe");  > 9 > Not sure what the login() function exactly does on VMS.  > L > But you should be aware that after the password has been entered, most VMSH > sites have a SET TERMINAL/INQUIRE command. This sends it varous escapeO > sequences and expect a response from the remote terminal. Only after multiple P > exchanges will SET TERM/INQUIRE give up, issue a waring message and get you toN > the % sign. During that time, any data sent to the system will be "eaten" by > SET TERM/INQUIRE.  > L > So, if the "login()" function doesn't handle the VMS login properly, it isL > probable that it would complete prematurely and then the "run testcob.exe"E > would be sent while SET TERM/INQUIRE is still eating up characters.  > E > You may wish to add another exchange between the login and the "run M > testcob.exe" to ensure yo really are at the "$" sign. (for instance, send a / > "SHOW LOG SYS$LOGIN" and wait for "LNM$JOB").  >    ------------------------------  # Date: Thu, 07 Nov 2002 21:44:33 GMT + From: jcring@switch.com (John C. Ring, Jr.) B Subject: Re: A problem telnet to OpenVMS box and run cobol program, Message-ID: <aqelcp$uug$1@usenet.switch.com>  a In article <3DCAD268.3B948C6F@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> wrote:  >Joseph Norris wrote: E >>     $t->login(Name=>$username, Password=>$passwd, Prompt=>'/\$/');  >> print "at dir cmd\n";$ >>     $t->print("run testcob.exe"); > 8 >Not sure what the login() function exactly does on VMS. > K >But you should be aware that after the password has been entered, most VMS + >sites have a SET TERMINAL/INQUIRE command.  >   >[...good suggestion deleted...]  N Assuming it isn't a CAPTIVE or RESTRICTED account, another thing he could try I is to append "/NOCOMMAND" to the username. If the "set term/inq" command  N wasn't in SYS$SYLOGIN, that'll avoid the issue all together.  It may speed up M the login process a touch as well, which could help with any time-out issues.    ------------------------------  # Date: Fri, 08 Nov 2002 17:01:42 GMT . From: Joseph Norris <jozefn@newbolt.sonic.net>B Subject: Re: A problem telnet to OpenVMS box and run cobol programH Message-ID: <Pine.LNX.4.40.0211080900460.31784-100000@newbolt.sonic.net>   Jf,   B Thanks for all of your help. With the suggestions you made my test> scripts are working fine now. I am off to tackle main project.  
 Thanks again.     ; #Joseph Norris (Perl - what else is there?/Linux/CGI/Mysql) I print @c=map chr $_+100,(6,17,15,16,-68,-3,10,11,16,4,1,14,-68,12,1,14,8, C -68,4,-3,-1,7,1,14,-68,-26,11,15,1,12,4,-68,-22,11,14,14,5,15,-90);     # On Thu, 7 Nov 2002, JF Mezei wrote:    > Joseph Norris wrote:F > >     $t->login(Name=>$username, Password=>$passwd, Prompt=>'/\$/'); > > print "at dir cmd\n"; % > >     $t->print("run testcob.exe");  > 9 > Not sure what the login() function exactly does on VMS.  > L > But you should be aware that after the password has been entered, most VMSH > sites have a SET TERMINAL/INQUIRE command. This sends it varous escapeO > sequences and expect a response from the remote terminal. Only after multiple P > exchanges will SET TERM/INQUIRE give up, issue a waring message and get you toN > the % sign. During that time, any data sent to the system will be "eaten" by > SET TERM/INQUIRE.  > L > So, if the "login()" function doesn't handle the VMS login properly, it isL > probable that it would complete prematurely and then the "run testcob.exe"E > would be sent while SET TERM/INQUIRE is still eating up characters.  > E > You may wish to add another exchange between the login and the "run M > testcob.exe" to ensure yo really are at the "$" sign. (for instance, send a / > "SHOW LOG SYS$LOGIN" and wait for "LNM$JOB").  >    ------------------------------  $ Date: Thu, 7 Nov 2002 17:30:18 -05005 From: "Brad McCusker" <brad.mccuskerNosp@Mcompaq.com> $ Subject: Re: Advanced Server and UAF/ Message-ID: <uslqbve1h2c945@corp.supernews.com>   H Larry is correct.  When we create a VMS file, some VMS user must own it.> Even if everyone maps to PWRK$DEFAULT, it is still a VMS user.  K And, I think there is a hard relationship between Administrator and SYSTEM, + or, at least there is in some circumstance.   : "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message- news:kHsmO99GZWIa@eisner.encompasserve.org... J > In article <3dca98fc$0$1451$1dc6e903@news.corecomm.net>, "Dave Gudewicz" <dgudewicz@core.com> writes:K > > Can't find this in the docs.  Are there any circumstances where an A.S. J > > (v7.3) w/LanManager *only* security needs to find anything in the UAF? > G > I thought it mapped Microsoft identities into VMS identities in order J > to know who should be the file owner.  It cannot do that without SYSUAF.   ------------------------------  $ Date: Thu, 7 Nov 2002 17:34:41 -05005 From: "Brad McCusker" <brad.mccuskerNosp@Mcompaq.com> $ Subject: Re: Advanced Server and UAF/ Message-ID: <uslqk7h89pj187@corp.supernews.com>   H We just completed the pilot of a new Advanced Server course back in Sep.G They are putting the finishing touches on it now, I anticipate it being  available sometime next spring.   A I will forward your request to the training people within Compaq.   K I spoke with Mind-IQ and they planned to offer the course once HP completed * it.  Send mail to them with your interest.  I If you can find a training contact on the HP/Compaq/OpenVMS web pages, it ( won't hurt to send your request to them.  L I'm scheduled to do a session on Advanced Server at the upcoming symposia in# Nashua, if you plan to attend that.   J And, I'll offer my full day seminar next spring in Amsterdam (yeah, I know you said USA).  
 Brad McCusker  OpenVMS Engineering   5 "Dave Gudewicz" <dgudewicz@core.com> wrote in message 2 news:3dcac2a0$0$1457$1dc6e903@news.corecomm.net...J > And if anyone knows of training for Advanced Server (in the USA), please > post.  > I couldn't find any. >  > --	 > Dave...  > I > It is noble to teach oneself, but still nobler to teach others-and less 
 > trouble. > -----Mark Twain  > 7 > "Dave Gudewicz" <dgudewicz@core.com> wrote in message 4 > news:3dca98fc$0$1451$1dc6e903@news.corecomm.net...K > > Can't find this in the docs.  Are there any circumstances where an A.S. J > > (v7.3) w/LanManager *only* security needs to find anything in the UAF? > > G > > I did read that external auth. and the combined VMS and LM security  model G > > needs to look at UAF, but the third model "LM only" is my question.  > >  > > -- > > Dave...  > > K > > It is noble to teach oneself, but still nobler to teach others-and less  > > trouble. > > -----Mark Twain  > >  > >  >  >    ------------------------------   Date: 7 Nov 2002 23:09:35 -0800 ( From: univms@bigfoot.com (Hamlyn Mootoo)> Subject: Re: Alpha network performance issues over leased line= Message-ID: <9cf7e06d.0211072309.7a14d9bb@posting.google.com>   b "Maverick" <maverick902@hotmail.com> wrote in message news:<aqe2qa$khj$1@sparta.btinternet.com>... > Hello, > K > We have a strange network performance problem.  We have a 2Mb leased line J > terminated at each end by a Cisco 2600 router.   A Digital Alpha DS20 isK > present in both locations, along with a network of NT boxes. One Alpha is L > loaded with VMS 7.1-2 and the other with 7.2-1.  The network at either end > is Switched ethernet.  >  > So we have >  > | NT boxes|   |DS20 (7.2-1)|) > ____|___________|______      Location A  >                    | >       |Baystack 350 |  >       |10/100 switch | >  >     |Cisco 2651 router|  >                   |  >                   | # >                   |    (2Mb link)  >                   |  >                   |  >     |Cisco 2621 router|  >  >       |Cisco Catalyst| >       |  3500 switch |& > _________|____________    Location B# >          |                      |  > | NT boxes|   |DS20 (7.1-2)| >  > N > All boxes can connect over IP, but performance varies dramatically depending > on the destination box.  > E > In summary, performance declines dramatically when the Alpha is the N > recipient of the data, and is OK otherwise. The alpha in location B has onlyM > recently moved there.  They used to talk quite happily to each other before M > this, when the second Alpha was in another location again.  The only things F > that have changed are the routers at each end (used to be Nortel ARNJ > routers) and the seond Alpha now plugs into a Cisco switch rather than a- > Nortel Baystack 350 it plugged into before.  > I > There appears to be no congestion on the link.  Can anyone think of any M > reason why the choice of router (Cisco) would have such an effect, but only K > on traffic to the Alphas? Or should I be looking at something else?  Some 1 > measurements on performance are given below.  .  > N > - An FTP put from the DS20 in A to the DS20 in B is dreadful at about 25kB/sL > - An FTP put from the DS20 in B to the DS20 in A is also not good at about > 50kB/sN > - An FTP put from any NT box to any NT box (over the link) is  about 170kB/s > in both directionsI > - An FTP put from an NT box in A to the DS20 in B is again about 25kB/s L > - An FTP get from an NT box on one side of the link to a DS20 on the other > is about 170kB/sM > - Local performance from NT boxes to Alpha's (both PUT and GET) seems OK at  > about 4MB/s in both locations  >   > Any help greatly appreciated!!  D How long does it take for a ping to traverse the link from an NT box! on side A to the Alpha on side B?   D Do you see the same slowdown with any other application besides FTP?  E Without any other information, the first thing I would check, is what E DNS (if any) the Alpha on the B side is pointing to, and/or how it is / resolving names, or back resolving IP to names.    Hamlyn   ------------------------------  * Date: Fri, 8 Nov 2002 09:17:24 +0100 (MET)& From: Rudolf Wingert <win@fom.fgan.de>> Subject: Re: Alpha network performance issues over leased line6 Message-ID: <200211080817.JAA22855@sinet1.fom.fgan.de>   Hello,  H I do not have any expirience with the Cat3500 switch. But I did see slowJ network performance too. Between to end nodes I did see a performance lessE then 1%. Seldom peaks with 90%. Together with our network service, we G tested all (e.g. different flow control). Nothing did help. And then we H did see, that the slow performance would be only via trunk. I don't knowC what the problem is with DECnet and trunk. Changing from trunk to a D single line, the network performance speed up to 95% constant over aD long period. AFAIK there are also some problem with full duplex. WasA this tested between the Alpha network controller and the Cat3500?    Best regards Rudolf Wingert    ------------------------------  $ Date: Fri, 8 Nov 2002 12:30:28 +02005 From: "Leonid Rosenboim" <My_1st_name@Consultant.Com> > Subject: Re: Alpha network performance issues over leased line7 Message-ID: <newscache$6k695h$7tc$1@lnews.actcom.co.il>   9 Darn, the duplex thing would have been an obvious choise, ? but I'd recommend setting 10Mbps & Half Duplex manually on both 3 the switch and the NIC before giving up this venue.   < A second thing I'd check is packet loss rate and delay time,9 as simply can be shown with a lengthy ping (>100 packets) 6 and using longer then usual packet size (1,200 bytes).  C Also, since the switch is also IP capable, do the ping test between 4 the VMS and Catalyst, then VMS to Router (local IP )3 then from VMP to local Router (T1 side IP address),  then remote router.   A This will show not only the nature of the problem, but also where 
 it begins.    H Can also do the same from the NT side, ping to local router, then remote router, H then remote switch, the VMS, but this is a no brainer cause your T to NT tests B show that the problem is *probably* on the catalust - VMS section.   -- HTHG ----------------------------------------------------------------------- I    Leonid Rosenboim     Visit:  http://www.masada2000.org/historical.html K        Consultant                Email: my first name at consultant dot com     @ "Leonid Rosenboim" <My_1st_name@Consultant.Com> wrote in message1 news:newscache$06s75h$5t9$1@lnews.actcom.co.il... @ > Well, the first thing I'd check is the PHY negotiation between8 > the Alpha and the Cat3500 switch. It sometimes happens= > that the switch things its Half duplex while the NIC thinks ) > its un full deuplex mode or vice versa.  >  >  > --I > ----------------------------------------------------------------------- K >    Leonid Rosenboim     Visit:  http://www.masada2000.org/historical.html I >        Consultant                Email: my first name at consultant dot  com  >  > 7 > "Maverick" <maverick902@hotmail.com> wrote in message , > news:aqe2qa$khj$1@sparta.btinternet.com...
 > > Hello, > > H > > We have a strange network performance problem.  We have a 2Mb leased lineL > > terminated at each end by a Cisco 2600 router.   A Digital Alpha DS20 isJ > > present in both locations, along with a network of NT boxes. One Alpha isJ > > loaded with VMS 7.1-2 and the other with 7.2-1.  The network at either end  > > is Switched ethernet.  > >  > > So we have > >   > > | NT boxes|   |DS20 (7.2-1)|+ > > ____|___________|______      Location A  > >                    | > >       |Baystack 350 |  > >       |10/100 switch | > >  > >     |Cisco 2651 router|  > >                   |  > >                   | % > >                   |    (2Mb link)  > >                   |  > >                   |  > >     |Cisco 2621 router|  > >  > >       |Cisco Catalyst| > >       |  3500 switch |( > > _________|____________    Location B% > >          |                      |   > > | NT boxes|   |DS20 (7.1-2)| > >  > > F > > All boxes can connect over IP, but performance varies dramatically > depending  > > on the destination box.  > > G > > In summary, performance declines dramatically when the Alpha is the K > > recipient of the data, and is OK otherwise. The alpha in location B has  > onlyH > > recently moved there.  They used to talk quite happily to each other > beforeH > > this, when the second Alpha was in another location again.  The only > thingsH > > that have changed are the routers at each end (used to be Nortel ARNL > > routers) and the seond Alpha now plugs into a Cisco switch rather than a/ > > Nortel Baystack 350 it plugged into before.  > > K > > There appears to be no congestion on the link.  Can anyone think of any J > > reason why the choice of router (Cisco) would have such an effect, but > onlyG > > on traffic to the Alphas? Or should I be looking at something else?  Some3 > > measurements on performance are given below.  .  > > I > > - An FTP put from the DS20 in A to the DS20 in B is dreadful at about  > 25kB/sH > > - An FTP put from the DS20 in B to the DS20 in A is also not good at about 
 > > 50kB/sH > > - An FTP put from any NT box to any NT box (over the link) is  about	 > 170kB/s  > > in both directionsK > > - An FTP put from an NT box in A to the DS20 in B is again about 25kB/s H > > - An FTP get from an NT box on one side of the link to a DS20 on the other  > > is about 170kB/sL > > - Local performance from NT boxes to Alpha's (both PUT and GET) seems OK > at! > > about 4MB/s in both locations  > > " > > Any help greatly appreciated!! > >  > >  > >  >  >    ------------------------------  # Date: Thu, 07 Nov 2002 23:19:37 GMT : From: "John Gemignani, Jr." <John.Gemignani@youknow.where>> Subject: Re: Alpha network performance issues over leased line- Message-ID: <3DCAAD55.7A759FC4@youknow.where>    Maverick wrote:  > L > > I think that I may have an idea.  Check the default route (the route for$ > > 0.0.0.0) using TCPIP SHOW ROUTE. > 
 > Hi John, > 	 > Thanks.  > H > Is that the same as UCX SHOW ROUTE? (I'm a complete bimbo where VMS isM > concerned I'm afraid).  If it is, then the route is set to the right place.   G Yes, it is the same, and I'm sorry to hear that it wasn't the problem.  5 It had me going for a while and I work on this stuff.    -John    ------------------------------  # Date: Fri, 08 Nov 2002 15:22:45 GMT - From: goathunter@goatley.com (Hunter Goatley)   Subject: ANN: Updated FTP_MIRROR/ Message-ID: <3dcbd62b.6560713@news.process.com>   J The following package has been updated on FTP.PROCESS.COM and its mirrors:      o FTP_MIRROR (Updated)   5 	FTP_MIRROR V1.0-5 is a DCL procedure that lets a VMS 3 	system mirror an FTP site (both VMS and UNIX sites 9 	are supported).  This version includes some enhancements <         supplied by Dr. Martin P.J. Zinser.  Thanks, Martin!  3 You can find this package using the following URLs:    http://www.process.com/openvms/   : ftp://ftp.process.com/vms-freeware/fileserv/ftp_mirorr.zip? http://vms.process.com/ftp/vms-freeware/fileserv/ftp_mirror.zip 6 ftp://ftp.tmk.com/vms-freeware/fileserv/ftp_mirror.zip; http://www.tmk.com/ftp/vms-freeware/fileserv/ftp_mirror.zip   ) And on the other mirrors within 24 hours.  Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  + Date: Fri, 08 Nov 2002 09:54:09 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> . Subject: Re: Cluster load balancing via TCP/IP; Message-ID: <01KOMD7T6M9Y9ZLF2M@sysdev.deutsche-boerse.com>   F > I have two Alphas in a cluster, running VMS 7.1-2, with UCX 4.2 ECO  > 2.    
 Quite old.  C > I've defined an Internet cluster address to the two systems, and  ) > both take incoming connections equally.   E IIRC, the cluster alias works as follows: the first system to get it  E keeps it; another system doesn't get it until the first system can't  . handle it.  No load balancing, no round robin.  D > But now I need to make node "A" take all of the incoming traffic, G > with node "B" ready to take the traffic if node "A" drops out of the  G > cluster.  What I don't want to do is to disable the Internet cluster  D > address on "B" right now, and hope to remember to re-establish it  > should "A" fail...   Let node "A" boot first.  F > I've searched the FAQ, Ask the Wizard, and the manuals.  Is there a 3 > way to do this, and what's the magic incantation?    ------------------------------  % Date: Fri, 08 Nov 2002 09:14:57 -0800 % From: Dean Woodward <deanw@rdrop.com> . Subject: Re: Cluster load balancing via TCP/IP( Message-ID: <3DCBF111.1030900@rdrop.com>   Phillip Helbig wrote: G > IIRC, the cluster alias works as follows: the first system to get it  G > keeps it; another system doesn't get it until the first system can't  0 > handle it.  No load balancing, no round robin.  G that's been my experience.  I load balance by having users connect via  E a DNS name, which is set up in the internal DNS server with an entry  D for each machine's IP address.  The DNS server will round-robin the " addresses it has for a given name.   ------------------------------  * Date: Fri, 8 Nov 2002 17:47:45 +0000 (UTC)+ From: david20@alpha2.mdx.ac.uk (David Webb) . Subject: Re: Cluster load balancing via TCP/IP+ Message-ID: <aqgtc0$hi4$1@aquila.mdx.ac.uk>   P In article <3DCBF111.1030900@rdrop.com>, Dean Woodward <deanw@rdrop.com> writes: >Phillip Helbig wrote:H >> IIRC, the cluster alias works as follows: the first system to get it H >> keeps it; another system doesn't get it until the first system can't 1 >> handle it.  No load balancing, no round robin.  > H >that's been my experience.  I load balance by having users connect via F >a DNS name, which is set up in the internal DNS server with an entry E >for each machine's IP address.  The DNS server will round-robin the  # >addresses it has for a given name.  > J The problem with DNS round robin is that it will still try the next system even if that system is down.G Most of the VMS TCPIP stacks support better ways of providing both load E balancing and failover using a DNS server than just using roundrobin.   E For DEC TCPIP Services see the information on the METRIC Server and   G LOAD BROKER in the Compaq TCPIP Services for OpenVMS Management manual.   
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Fri, 08 Nov 2002 19:26:14 +0100  From: munk@home.nl. Subject: Re: Cluster load balancing via TCP/IP8 Message-ID: <i1vnsu0q5vndf5nl4jf1clnu85s7mq4fpm@4ax.com>  A On Fri, 8 Nov 2002 17:47:45 +0000 (UTC), david20@alpha2.mdx.ac.uk  (David Webb) wrote:   Q >In article <3DCBF111.1030900@rdrop.com>, Dean Woodward <deanw@rdrop.com> writes:  >>Phillip Helbig wrote: I >>> IIRC, the cluster alias works as follows: the first system to get it  I >>> keeps it; another system doesn't get it until the first system can't  2 >>> handle it.  No load balancing, no round robin. >>I >>that's been my experience.  I load balance by having users connect via  G >>a DNS name, which is set up in the internal DNS server with an entry  F >>for each machine's IP address.  The DNS server will round-robin the $ >>addresses it has for a given name. >>K >The problem with DNS round robin is that it will still try the next system  >even if that system is down. H >Most of the VMS TCPIP stacks support better ways of providing both loadF >balancing and failover using a DNS server than just using roundrobin. > F >For DEC TCPIP Services see the information on the METRIC Server and  H >LOAD BROKER in the Compaq TCPIP Services for OpenVMS Management manual. >  >David Webb  >VMS and Unix team leader  >CCSS  >Middlesex University  > ; We are using the TCPIP LOADBROKER setup, and it works fine. C However your master DNS server will have to be configured to accept E dynamic DNS updates. This means that the VMS cluster will send a name F + TCPIP address to the master DNS server, and that will announce it in the usual way.  A The LOADBROKER + METRIC setup will do load balancing, and you can B register more DNS names (for different applications for instance).  > Now there is something very strange with VMS TCPIP clustering.E Normally you can have only one master DNS server, however with VMS it @ is possible to build a clustered DNS server. That is wonderfull,C because if one of the nodes in the cluster goes down, the other one 6 will still be a master DNS server. A real VMS virtue !  C Now comes the less nice part. You can not use this clustered master E DNS server with dynamic updates ! The reason is that only the cluster E member that receives the update will know about the update. The other 9 node(s) will not know of this update, because there is no : synchronisation mechanism over the cluster members for the distribution of the updates.  F So we can build clustered master DNS servers for maximum availability,? and we can build clustered production systems, also for maximum < availability, but we can't combine the two. Very unpleasant.   ------------------------------   Date: 8 Nov 2002 18:26:50 GMT ( From: bill@cs.uofs.edu (Bill Gunshannon). Subject: Re: Cluster load balancing via TCP/IP5 Message-ID: <aqgvl7$a3ttj$1@ID-135708.news.dfncis.de>   ( In article <3DCBF111.1030900@rdrop.com>,( 	Dean Woodward <deanw@rdrop.com> writes: > Phillip Helbig wrote: H >> IIRC, the cluster alias works as follows: the first system to get it H >> keeps it; another system doesn't get it until the first system can't 1 >> handle it.  No load balancing, no round robin.  > I > that's been my experience.  I load balance by having users connect via  G > a DNS name, which is set up in the internal DNS server with an entry  F > for each machine's IP address.  The DNS server will round-robin the $ > addresses it has for a given name. >   E But, because the Nameserver can't possibly have any notion of machine G usage this offers load balancing only by pure chance and in all likeli-  hood, not at all.   E The BIND documentation itself calls it "load sharing" and denies that  it is "load balancing".    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Fri, 08 Nov 2002 13:10:05 -0500  From: RCS / RI <rcs@osfn.org>  Subject: DECtalk error codes( Message-ID: <3DCBFDFC.6A6B31E7@osfn.org>  C I have a DECtalk (DTC01) that does not appear to work.  At power on A there are a series of touch tones, and then the red LED's display > 11010010  Any hints about what this indicates would be greatly appreciated.   -mikeu Michael Umbricht   --  H The Retro-Computing Society of RI, Inc.                   (401) 861-1977H 25 Eagle St Bldg 5 Ste 206                      http://www.osfn.org/rcs/H Providence RI 02908                       telnet://kronos.egr-ri.ids.net   ------------------------------  % Date: Thu, 07 Nov 2002 16:16:25 -0500 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: Email Filter / Message-ID: <3DCAD826.800A0E2D@vl.videotron.ca>   I > > I am looking for some ideals for SPAM filtering on an OpenVMS machine ? > > running Apache (Compaq Secure Web) with SMTP mail service.    Z If you have  TCPIP Services V5.1 or later, there is a built-in SPAM filetering capability.  M Look at the TCPIP Services Management guide on the VMs documentation web site  (www.openvms.digital.com:8000 )   / Also, you will find "example/template" file in:   I lib/extract=smtp_config sys$library:cpip$templates.tlb /output=myfile.txt    ------------------------------   Date: 8 Nov 2002 13:19:47 GMT , From: bill@gw5.cs.uofs.edu (Bill Gunshannon) Subject: Re: HP Advocacy Site 5 Message-ID: <aqgdli$9o6ba$2@ID-135708.news.dfncis.de>   G In article <IABy9.15550$YSz1.751@news01.bloor.is.net.cable.rogers.com>, & 	"John Smith" <a@nonymous.com> writes: > ? > "JF Mezei" <jfmezei.spamnot@vl.videotron.ca> wrote in message + > news:3DCAD646.36CB316B@vl.videotron.ca...  >>M >> Look at how the lack of a standard TCPIP stack for VMS hurt VMS during the 
 >> 1990s.  > M > That's because DECnet was considered to be the preferred protocol stack for K > VMS because it was deterministic and well tested. And because at the time 1 > unix was snake oil or a close relative thereof    F That's hilarious.  I am willing to bet that by 1990 there were alreadyC many more Unix system in use worldwide than VMS.  With many of them > doing work as serious as if not more serious than VMS systems.  M >                                                 so why would anyone want to # > use TCP/IP in a VMS environment.    N In order for VMS to fit into the computing world along with it's competitiors.  M >                                  It took quite a while for that attitude to 	 > change.    To VMS's disadvantage.    bill     --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Thu, 07 Nov 2002 16:08:25 -0500 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: HP Advocacy Site / Message-ID: <3DCAD646.36CB316B@vl.videotron.ca>    Larry Kilgallen wrote:B > It is to the disadvantage of VMS if VMS Development were to takeA > on the task of driving all the third party software vendors out  > of the VMS business.  N As far as certain middleware, I disagree. Having one standard interface to VMSN specific facilities is to the advantage of VMS because folks can then start to3 write software and expect it to run on any VMS box.   J Look at how the lack of a standard TCPIP stack for VMS hurt VMS during theH 1990s. With the arrival of TCPIP V5, this made VMS a much more palatableJ platform (even though it may have been better to just take TGV's stack and make that one standard)   I If the VMS engineers were ahead of needs, then they would have developped G those standard interfaces/middleware, which would foster development of L applications right away and make VMS a leader.  It is this void which forcesK individual ISVs to roll their own proprietary solution to fill the gap left  wide open by VMS engineers.   L And if the VMS engineers provided the wheels in a timely fashion (leadershipP mode), then ISVs could assemble the cars without having to re-invent the wheels.  L And thsi is where that unwanted port to IA64 is hurting VMS: those resourcesM devoted to IA64 port would be much more productive for VMS if they focused onNK making all sorts of "leadership" changes to bring VMS not only up to speed,n but also ahead of others.a  L VMS is labeled as a secuse system. Yet it lacks many of the secure stuff for its TCPIP stack.   ------------------------------  # Date: Thu, 07 Nov 2002 22:14:00 GMTe# From: "John Smith" <a@nonymous.com>a Subject: Re: HP Advocacy SiteeG Message-ID: <IABy9.15550$YSz1.751@news01.bloor.is.net.cable.rogers.com>e  = "JF Mezei" <jfmezei.spamnot@vl.videotron.ca> wrote in message ) news:3DCAD646.36CB316B@vl.videotron.ca...M > Larry Kilgallen wrote:D > > It is to the disadvantage of VMS if VMS Development were to takeC > > on the task of driving all the third party software vendors out  > > of the VMS business. >vL > As far as certain middleware, I disagree. Having one standard interface to VMSvG > specific facilities is to the advantage of VMS because folks can thenr start to5 > write software and expect it to run on any VMS box.P >rL > Look at how the lack of a standard TCPIP stack for VMS hurt VMS during theJ > 1990s. With the arrival of TCPIP V5, this made VMS a much more palatableL > platform (even though it may have been better to just take TGV's stack and > make that one standard)  > K > If the VMS engineers were ahead of needs, then they would have developpednI > those standard interfaces/middleware, which would foster development of G > applications right away and make VMS a leader.  It is this void whichk forcesH > individual ISVs to roll their own proprietary solution to fill the gap left > wide open by VMS engineers.  >tB > And if the VMS engineers provided the wheels in a timely fashion (leadershipoJ > mode), then ISVs could assemble the cars without having to re-invent the wheels.t > D > And thsi is where that unwanted port to IA64 is hurting VMS: those	 resources L > devoted to IA64 port would be much more productive for VMS if they focused onF > making all sorts of "leadership" changes to bring VMS not only up to speed, > but also ahead of others.  > J > VMS is labeled as a secuse system. Yet it lacks many of the secure stuff ford > its TCPIP stack.  K That's because DECnet was considered to be the preferred protocol stack forbI VMS because it was deterministic and well tested. And because at the timeoJ unix was snake oil or a close relative thereof so why would anyone want toK use TCP/IP in a VMS environment. It took quite a while for that attitude tou change.e   ------------------------------  % Date: Thu, 07 Nov 2002 18:00:51 -0500n0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: HP Advocacy Sitet/ Message-ID: <3DCAF09A.44889F39@vl.videotron.ca>b   Bob Koehler wrote:B >    Why?  We use Java for extreemly portable code.  We use nativeA >    compilers for OS-specific code and performance requirements..  8 Which is what prevents Java from being more widely used.   ------------------------------  # Date: Thu, 07 Nov 2002 22:00:58 GMTo5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>e! Subject: Re: HP keyboard problems 3 Message-ID: <uoBy9.52$ry5.1742968@news.cpqcorp.net>w  G No.  It is not wireless.  IMHO wireless should use direct-line-of-sight-H infrared signaling.  Using anything else, even low power FM, causes this risk.     % Robert Deininger wrote in message ...e= >In article <01C285B5.BA988E10@sulfer.icius.com>, Shane Smithp ><ssmith@icius.com> wrote: >.A >>Here's an interesting one. This new LK463 isn't wireless is it?o >>H >>http://www.aftenposten.no/english/local/article.jhtml?articleID=431326 >>G >>Apparently a couple of guys in Norway have HP wireless keyboards thatsH >>can type on each others' machines 150 metres (roughly yards) apart. SoG >>HP replaced them, and it's still happening. Not something you want to6 >>type passwords on... >,9 >LK463 is USB and (using an included adapter widget) PS2.0   ------------------------------  $ Date: Fri, 8 Nov 2002 10:19:25 -0800$ From: Shane Smith <ssmith@icius.com>! Subject: RE: HP keyboard problems 0 Message-ID: <01C28710.56CD3230@sulfer.icius.com>  G Very true, Fred, where security is an issue. Only trouble with infraredbD is its speed - there's a slight lag. It's not noticeable when you'reG doing office stuff, but it's a killer in most reaction based games. AndeH you /really/ don't want one of the guys walking in front of the receiver on his way to the beer.N   Shanee   -----Original Message-----: From: Fred Kleinsorge [mailto:kleinsorge@star.zko.dec.com]) Sent: Thursday, November 07, 2002 2:01 PMM To: Info-VAX@Mvb.Saic.Come! Subject: Re: HP keyboard problemst    G No.  It is not wireless.  IMHO wireless should use direct-line-of-sightnH infrared signaling.  Using anything else, even low power FM, causes this risk.i    % Robert Deininger wrote in message ...d= >In article <01C285B5.BA988E10@sulfer.icius.com>, Shane Smithx ><ssmith@icius.com> wrote: >hA >>Here's an interesting one. This new LK463 isn't wireless is it?  >>H >>http://www.aftenposten.no/english/local/article.jhtml?articleID=431326 >>G >>Apparently a couple of guys in Norway have HP wireless keyboards thatAH >>can type on each others' machines 150 metres (roughly yards) apart. SoG >>HP replaced them, and it's still happening. Not something you want to> >>type passwords on... >@9 >LK463 is USB and (using an included adapter widget) PS2.-   ------------------------------  # Date: Fri, 08 Nov 2002 18:38:11 GMTt5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> ! Subject: Re: HP keyboard problems 2 Message-ID: <nwTy9.25$2n6.597137@news.cpqcorp.net>  K The protocol that is used is probably to blame.  I'm guessing they're doingIG something really slow and stupid to try to minimize errors.  After all,'I nothing that can be done on a KB is truly "fast" - even the fastest touchd3 typist is slower than almost any signalling method.c  J In any case, I've never used one, so can't really comment on how well/poor
 they work.      D Shane Smith wrote in message <01C28710.56CD3230@sulfer.icius.com>...H >Very true, Fred, where security is an issue. Only trouble with infraredE >is its speed - there's a slight lag. It's not noticeable when you'relH >doing office stuff, but it's a killer in most reaction based games. AndI >you /really/ don't want one of the guys walking in front of the receiver  >on his way to the beer. >n >Shane >m >-----Original Message-----i; >From: Fred Kleinsorge [mailto:kleinsorge@star.zko.dec.com] * >Sent: Thursday, November 07, 2002 2:01 PM >To: Info-VAX@Mvb.Saic.Com" >Subject: Re: HP keyboard problems >1 >,H >No.  It is not wireless.  IMHO wireless should use direct-line-of-sightI >infrared signaling.  Using anything else, even low power FM, causes thisr >risk. >  > & >Robert Deininger wrote in message ...> >>In article <01C285B5.BA988E10@sulfer.icius.com>, Shane Smith >><ssmith@icius.com> wrote:  >>B >>>Here's an interesting one. This new LK463 isn't wireless is it? >>>oI >>>http://www.aftenposten.no/english/local/article.jhtml?articleID=431326e >>>eH >>>Apparently a couple of guys in Norway have HP wireless keyboards thatI >>>can type on each others' machines 150 metres (roughly yards) apart. SohH >>>HP replaced them, and it's still happening. Not something you want to >>>type passwords on...a >>: >>LK463 is USB and (using an included adapter widget) PS2. >I >I   ------------------------------  $ Date: Fri, 8 Nov 2002 17:32:49 +00002 From: John Eisenschmidt <jweisen@eisenschmidt.org> Subject: Re: HP Service 3 Message-ID: <20021108173249.A5465@eisenschmidt.org>i  % I thought I'd give you all an update.   < My tech called me at 9:35a this morning and told me he'd "beF downstairs in two minutes". I love advanced notice, especially since ID walked into a sick Metaframe and two backups that didn't happen last night.  E He has the right motherboard, so we take the system down, he replacescD it, and we go through all the firmware settings (message to HP: stopJ replacing motherboards when the CPU is bad), turns the system on, and he'sK  like "see ya". I tell him to wait and let's make sure it comes up all the m way.  @ It didn't. It died half way through the startup. He blames it onF software. "But nothing has changed since you were here Tuesday, and itE came up fine." He halts the system, does a show error (sees none) andr inits the system.s  = Don't worry, I didn't want my shadow sets to be consistant orc	 anything.6  @ It comes back up, starts a full shadow merge (7 shadow sets) andC volume rebuilds. It stops at the same point, so he halts the systemrF again. At this point I had to walk out of the room. Sure, I would haveD probably dropped the system too, but I would have liked to make that call.   C All the while, we haven't been able to ping the system. Tech bringsh@ the system up with a minimal boot, and he says "that works fine,A hardware is ok". I scratch my head and we do a SET VERIFY and runp? SYSTARTUP_VMS.COM to see where it does. It seems to get through F everything: the last couple lines are loading DECEVENT and then addingF some symbols for perl. He thinks it's maybe the symbols that have beenA there for 1 1/2 years. I ask if maybe it's a problem with the NIC E (since it's on the motherboard) because even though it loads and getsoE past TCPIP$STARTUP, I can't ping or telnet into the system. He bringseB it down to the SRM console and starts twiddling with the ewa_mode.F Now, it's on a CISCO switch that has the speed and duplex hardcoded toD 100 FULL, so even if setting the NIC to autonegotiate "fixes it", it, doesn't really fix it. He's sure this is it.  F It isn't. He gets on the phone, I have to leave the room again (as I'mA thinking to myself: this only happens to me) -- I always gets thedB technicials who seem bright and end up actually making the problemC worse. Finally he orders another new motherboard and starts to swape the old one back in.  F Guess what he found? The new board has a tiny crystal on the back nearD the network card -- it was snapped off on the new board. He found it& in the bottom of the anti-static bag.   B With the old board in, the system seems to come right up (with theB exception of not allowing console logins now), except for one tiny problem:   $ sh err, Device                           Error Count* PKC0:                                  630  B Half my shadow members are missing, and it appears that one of the' SCSI controllers is anything but happy.e  D It looks like I'll be married to my pager until next week, and I getC the added bonus of watching VMS MONITOR for the rest of today as we ) attempt to turn hamburger back into cow.    B Are you listening HP? IBM would have brought two motherboards (hadF they been silly enough to replace the whole board in the first place).C Their tech would have had an entire spares kit in the truck, and itj< would have only been one service call, not (at least) three.   Again I say:  
 	-yea IBM!	 	-boo HP!o  S Unless the Voices are Mistaken, John Eisenschmidt (jweisen@EISENSCHMIDT.ORG) Wrote:aC > My little DS10s have quite a racket going. When we got them 12-18e   [snip]   > 	-yea IBM! > 	-boo HP!M   -- d/ John W. Eisenschmidt (jweisen@eisenschmidt.org)r.   http://www.eisenschmidt.org/jweisen/pgp.html  6 [The Onion -> Horoscopes] Aquarius: (Jan. 20--Feb. 18)D      Don't just demand the best from yourself and those around you. 7      Demand coffee from yourself and those around you.     ------------------------------  $ Date: Fri, 8 Nov 2002 12:18:00 +0100" From: "Hans Vlems" <hvlems@iae.nl>Y Subject: Re: I'm looking for a Digital RRDxx CD ROM (near BELGIUM). (+VAX 6000-6xx cards  5 Message-ID: <aqg6hc$9m2k6$1@ID-143435.news.dfncis.de>   > "T.R." <tr303@notmail.com.replace.n.with.h> schreef in bericht- news:2HDy9.52302$I6.4931450@zwoll1.home.nl...t >s3 > "Shane Smith" <ssmith@icius.com> wrote in message_, > news:01C28665.8C9180C0@sulfer.icius.com...E > > Does it have to be an RRD? Or even Digital? In my experience most K > > Toshiba off-the-shelf SCSI CD roms work with VMS. I'm using a couple ofbK > > 24x's I picked up at a PC parts shop a couple of years ago. Mine workedhL > > straight out of the box, but I've heard you might have to set the parity
 > > to on. > >m	 > > Shane  > > G > Interesting! As I'm also investigating this subject, any reports from  people; > getting non DEC (toshiba?) CD drives working are welcome!h >eH This Pioneer drive works well, and is a lot faster that the DEc drive it	 replaced!n $ sh dev $1$dkb400:/full  H Disk $1$DKB400: (HELIUM), device type PIONEER CD-ROM DR-U12X, is online, file-r  L oriented device, shareable, served to cluster via MSCP Server, error logging   is enabled.   $ Error count 0 Operations completed 0  # Owner process "" Owner UIC [SYSTEM]o  6 Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W  ) Reference count 0 Default buffer size 512    Allocation class 1   ------------------------------  % Date: Fri, 08 Nov 2002 09:49:15 -0500h( From: David Froble <davef@tsoft-inc.com>Y Subject: Re: I'm looking for a Digital RRDxx CD ROM (near BELGIUM). (+VAX 6000-6xx cards o, Message-ID: <3DCBCEEB.9030806@tsoft-inc.com>   T.R. wrote:o  3 > "Shane Smith" <ssmith@icius.com> wrote in messageF, > news:01C28665.8C9180C0@sulfer.icius.com... > C >>Does it have to be an RRD? Or even Digital? In my experience most I >>Toshiba off-the-shelf SCSI CD roms work with VMS. I'm using a couple ofaI >>24x's I picked up at a PC parts shop a couple of years ago. Mine worked J >>straight out of the box, but I've heard you might have to set the parity >>to on. >> >>Shane  >> >>N > Interesting! As I'm also investigating this subject, any reports from people; > getting non DEC (toshiba?) CD drives working are welcome!     M Many of the DEC drives were in fact Toshibas.  I've got a number of Toshibas  Q that have worked on both VAXs and Alphas.  Can't remember all the model numbers. 1M   5701 or 5702 is one that comes to mind.  Should be available for about the tP same price as a Big Mac, so lose a bit of weight and get a CD-ROM drive at thew 
 same time.   Dave   ------------------------------  % Date: Fri, 08 Nov 2002 13:32:30 -0500 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>4 Subject: Re: Last call for C-Kermit 8.0.206 binaries/ Message-ID: <3DCC0339.8DB24B6F@vl.videotron.ca>    Frank da Cruz wrote:H > but there's always room for improvement.  If you can build C-Kermit onH > any platform for which we don't yet have an 8.0.206 binary, we can putH > it on the CDROM if you send it in soon.  You can find the current list > of binaries here:9  0 I'd be willing to build it for VMS 7.2 TCPIP-5.3  K However, is there really any differennce between such a build and one builtt6 with TCPIP 5.1 which you already have ? (TCPIP = UCX).  I Also, just a comment, in the list of OS at the top of the page, just use r9 "DEC Tru64, DEC VMS" instead of "DEC/Compaq/HP VMS" etc. c  " Or as a compromise, "DEC/HP VMS".   J The word Compaq doesn't belong anywhere near "VMS" and it just make things/ look far more complicated than they really are.u   ------------------------------  # Date: Thu, 07 Nov 2002 21:57:46 GMTt5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> . Subject: Re: mouse with a scroller for OpenVMS3 Message-ID: <ulBy9.51$Kx5.1708174@news.cpqcorp.net>h  , Point me to a FTP site, and I'll look at it.    D Shane Smith wrote in message <01C2859B.5623B140@sulfer.icius.com>...F >Forrest, I have a little program that came with the freeware TeraTermF >terminal emulator that might be interesting to try. You run it, pressG >keys, and note the numbers for use in TeraTerm's keyboard config file.aG >When I tried it on my LK450 I was surprised to find my keyboard driveroG >was actually generating unique codes for the extra keys, so I was ablei >to get them working.  >rH >Would you care to try it on the LK463? I won't post it here, since it's
 >a binary. >- >Shane >- >-----Original Message------9 >From: Forrest Kenney [mailto:Forrest.Kenney@hp.com.doom] , >Sent: Wednesday, November 06, 2002 12:43 PM >To: Info-VAX@Mvb.Saic.Com/ >Subject: Re: mouse with a scroller for OpenVMSo >o >  > F >    I just for the heck of it plugged an LK463 into my Win200 system. >All the >coretI >keys work as you would expect.  The windows code does not seem to have ag >goodn >idea what to do withn >h	 >    helpw >    dos >    F13-F20 >    Pf1-Pf4 >    the , key in the keypad.w >e# >    These keys map out as follows:e >o- >    Prev & next map to page up and page down- >    remove maps to delete >    inset here maps to insert >    find maps to end  >    select ???????w >sG >    All of these were mapped onto keys that are already defined in they
 >USB HumanF >Interface Device specification.  We are not using any special code we >are all >mapped C >onto existing codes.  So the Microsoft default HID driver does note >appear to map >allC >HID codes only those that it cares about.  This is not a surprise.  >h >d >Forrest Kenneyr >OpenVMS Development >D   ------------------------------  $ Date: Fri, 8 Nov 2002 13:07:51 -00002 From: "Chris Sharman" <chris.sharman@sorry.nospam> Subject: Re: OSU server cachei4 Message-ID: <aqgcs9$aaj$1$830fa7b3@news.demon.co.uk>  > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message7 news:cf15391e.0211071318.6519468d@posting.google.com... ? > "Chris Sharman" <chris.sharman@sorry.nospam> wrote in message 0 news:<aqdvju$gg8$1$8300dec7@news.demon.co.uk>...K > > We'd like to specify a short life from the request time, or no caching,  or! > > something - any suggestions ?  > H > You might find http://www.htmlgoodies.com/beyond/nocache.html helpful.  D Appears to be strictly html (& some asp). I'm serving PDF documents.   Chris    ------------------------------  $ Date: Fri, 8 Nov 2002 13:17:53 -00002 From: "Chris Sharman" <chris.sharman@sorry.nospam> Subject: Re: OSU server caches4 Message-ID: <aqgdf3$sfr$1$8300dec7@news.demon.co.uk>  I Thanks - does this affect client-side caching, because I think that's theK	 problem ? ' <winston@SSRL.SLAC.STANFORD.EDU> wrote: ) ><craigberry@nospam.telocity.com> writes:i: > >Read the chapter on caching in Alan Winston's new book. >l > Here's an excerpt: >o- > FILECACHE REFRESH minutes:seconds | secondsn > K > Specifies number of seconds for which a document can be served from cachedH > without being checked against the on-disk file system.  The default is threeuL > minutes; you could specify that as 3:00 or 180.  You could also specify it asJ > as 0; this would force an on-disk check for every file access, but you'd stillnK > get some benefit from the cache as the actual file wouldn't be read if it % > hadn't changed since the last time.sL > -------------------------------------------------------------------------- -- >eG > (This command affects every file ever served.  To deal just with your0 PDFs, or0 > for any link of which you are in control ... ) > L > -------------------------------------------------------------------------- ---- > FILECACHE EXCLUDE template >nL > Don't cache files matching this template.  You can have multiple instances ofJ > this directive, and all will apply.  By default the HTTP_PATHS.CONF file has a1 >B > FileCache exclude *; > K > directive in it.  This allows the requestor to force an on-disk lookup ofn atF > given file by putting a bare semicolon on the end of the path.  This templateL > supports full VMS wildcarding.  If you had a lot of documents that changed veryL > frequently and getting the most recent version was important, you could do >o > FileCache exclude */dynamic/*e >gE > and any directory tree, including userdirs, could have a [.DYNAMIC] G > subdirectory, within which any file would automatically be noncached.  >.L > -------------------------------------------------------------------------- ----  G Thanks for this. I don't think server-side caching is the whole problemf! though, or necessarily any of it. I I think my problem is client-side, or possibly some third-party cache. Do + these commands affect client-side caching ? I With any luck I'll get to the bookshop at the weekend, if it's made it to-
 the UK ...   Regards, Chris:   ------------------------------  % Date: Fri, 08 Nov 2002 12:40:35 -0500p0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>4 Subject: Re: OT: The lighter side of VMS Engineering/ Message-ID: <3DCBF711.D0FDCCEC@vl.videotron.ca>F   CSABA HARANGOZO wrote: > G >         This might not be new(s) to some, but if you have a few spareiD >         minutes, you might have a look into ( TYPE will do ) theseD >         files :  SYS$SYSTEM:BIRTHS.MSGHLP  and/or  DEATHS.MSGHLP .  I Considering the dates inside those files (1991 and beyond), how come thisw doesn't exist on VAX ?  H At the time this was written, wasn't VAX still the main version of VMS ?   ------------------------------  % Date: Fri, 08 Nov 2002 18:02:23 +0100-6 From: Horst Drechsel <ai05@sternwarte.uni-erlangen.de>, Subject: set time/cluster - what does it do?< Message-ID: <00A16AD9.73FF42D5.9@sternwarte.uni-erlangen.de>  	 Dear all,c  E    my environment: OpenVMS cluster under VMS 7.2-1 with XP1000 serverw+                    VMS721_UPDATE V3.0 Patchn+                    TCPIP_ECO V5.0-113 Patch-  F    my problem:     I am using NTP to synchronize my system time on theJ                    cluster server once in a while and after each boot withM                    some external network time servers, which is working fine.h  D                    After having shut down the TCPIP$NTP server afterM                    successful time adjustment, the system clock of my clusterdE                    server is keeping the correct (synchronized) time.y  E                    However, if I now try to synchronize the clocks oftK                    the other cluster nodes by issuing (on the same server):a                    $ mc sysman*            SYSMAN> set environment/cluster"            SYSMAN> config set time  H                    the system clock of the cluster server (and all otherD                    cluster nodes) fall back to some discrepant time,5 		   mostly close to the one maintained by the systemr-                    before the NTP adjustment.e  D                    Same effect happens if - after having establishedC                    NTP time synchronization over the net - I issue:n                    $ set time-  E                    on the server alone: the correct time will also be E                    reset to some different value, usually the one thesG                    system clock has been showing before the adjustment.h  H                    Are there two different clocks in the XP1000, or whatD                    does the "set time" command really do? (help docuE                    says it "resets" system time, but reset to what??)       Thanks for any info,e                           Horstc  s   --M  ****************************************************************************h)   Horst Drechsel                         eL   Dr. Remeis Observatory                 drechsel@sternwarte.uni-erlangen.deL   Astronomical Institute                             Phone: +49-951-95222-15L   University Erlangen-Nuernberg                        Fax: +49-951-95222-22*   Sternwartstr.7, D-96049 Bamberg, GermanyM  ****************************************************************************>   ------------------------------  % Date: Thu, 07 Nov 2002 16:58:07 -0500k- From: "Richard D. Piccard" <piccard@ohio.edu>e) Subject: Re: Slow performance with BACKUPd( Message-ID: <3DCAE1EF.8853A456@ohio.edu>   mckinneyj@cpva.saic.com wrote: > " > >> >>> /verify will slow it down > >>N > >> No, not realy. The backup needs eight day to copy the data to the saveset > >> and two days to verify. > >dC > > This statement about it only taking 2 days to verify the backupiH > > definitely points the problem to data retrieval from the served disk3 >  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^tJ > > to the backup disk. From this statement, the backup should only take 2I > > days to complete. There is a bottleneck not a parameter problem beingi > > too small. > >e > M > ??? sounds more like the issue is with writing not reading... the operation N > with two reads is relatively quick; the one with a read and a write is slow. >  > -- > - Jim-    N My experience with disk-to-disk backup is more than ten years old, but at thatI time, if I recall correctly, creating on-disk savesets on a disk mounted @N /SYSTEM was an excruciatingly slow business.  I think that part of the problemN is the head movement to update the various RMS data structures on the disk as M it goes along, so that the other users (which for all it knows there are lotsa. of) can continue their activity on that disk.   G Have you tried allocating the target disk drive to the backup writing   O process, mounting the target disk /FOREIGN, and writing one saveset at a time, eL in a sequence.  I know that works with removable media drives where you wantN to make a multi-volume saveset of one source disk.  A second process could be R doing likewise to the other target disk, if I recall the original post correctly.   N It seems to me that BACKUP could be smart enough to preallocate the full size P of the source disk plus its own overhead, and then truncate the file and releaseM the un-used blocks if there were any multiply-listed files (e.g., SYSCOMMON) uN that it had double-counted, or any files set /NOBACKUP.  This ought to permit # very nearly spiral writing speeds. P  O Another thought:  have you measured the write speed when backing up between theeN two disks that you are trying to use as targets?  That might permit isolating 5 the bottleneck as between the disks and the network. l  	 						RDPc   -- eB ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  $ Date: Fri, 8 Nov 2002 10:36:46 -0800$ From: Shane Smith <ssmith@icius.com>= Subject: The advocacy site's instapol question is VMS related 0 Message-ID: <01C28712.C346BFB0@sulfer.icius.com>  ? Today's instapol question on http://www.hpuseradvocacy.org/ is:o   --- D The Alpha to IA64 migration potentially offers a wider product rangeB (for example, OpenVMS Workstations) as operating systems will work across the IA64 range.  H o This possibility does not change my plans to move to another operating system on an HP platform.p  F o This possibility does not change my plans to move to another vendor.  A o This possibility offers me a pathway to continue use of currenti operating systems.  : o I will believe the migration when I see it has happened.  C o The transition from Alpha to IA64 is of no interest to me at all.  ---u  D I wish there were a "Please drop the port and develop Alpha instead"	 option...-  F Incidentally, how come there's only 3 "me too"'s on the "please market- VMS" issue? 3? THREE? That's a lousy turnout.a   Shane    ------------------------------  % Date: Fri, 08 Nov 2002 20:30:21 +1100e1 From: Paddy O'Brien <paddy.o'brien@tg.nsw.gov.au>n" Subject: Re: VAX in comp.risks ..., Message-ID: <3DCB842D.3060700@tg.nsw.gov.au>   Bill Hobbs wrote:s  ; >... though maybe not in the context we would like to read.t    G Bill, I agree and with the same love of a certain OS.  But be it VMS or G Micro$shit, to me the salient point is that we as humans should be abletF to control any computer, not be controlled.  That is what I understood% of the reason for Larry posting this.s     Regards, Paddy        G ***********************************************************************e  C "This electronic message and any attachments may contain privilegedr> and confidential information intended only for the use of the B addressees named above.  If you are not the intended recipient of C this email, please delete the message and any attachment and advisenB the sender.  You are hereby notified that any use, dissemination, 7 distribution, reproduction of this email is prohibited.f  A If you have received the email in error, please notify TransGrid  A immediately.  Any views expressed in this email are those of the -= individual sender except where the sender expressly and with cC authority states them to be the views of TransGrid.  TransGrid usesB> virus scanning software but excludes any liability for viruses contained in any attachment.  < Please note the email address for TransGrid personnel is now$ firstname.lastname@transgrid.com.au"  G ***********************************************************************h   ------------------------------  % Date: Fri, 08 Nov 2002 12:32:28 -0500E0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>" Subject: Re: VAX in comp.risks .../ Message-ID: <3DCBF52B.DDFD9D19@vl.videotron.ca>i   Nic Clews wrote:I > Perhaps this should be a warning to those who combine a development andz, > production environment on the same system.  I Since pipeline systems would be operating in real time (in this case, the K software didn't have the quick enough response to close some valve before anK disaster occured), how does VMS behave when an interactive user logs on andPD does stuff on the system with regards to the real time application ?  N If a real time process is at priority 17, and an interactive process starts at5 priority 4, can the later really impact the former ? f  F (OK, I realise that if you lock some file/record, or that if you do anS ana/disk/repair, it will lock out any process from accessing the blocked resource).   L But in terms of CPU, shouldn't a priority > 16 process really get the CPU it needs, when it needs it ?D  K (obviously, if the priority was granted to the username under which the app1L runs, and then someone logs in with that username, both will be in real-time' priorities and that would be bad news).v   ------------------------------   Date: 8 Nov 2002 11:20:27 -0600. From: briggs@encompasserve.org" Subject: Re: VAX in comp.risks ...3 Message-ID: <AXiSUPKs8eEt@eisner.encompasserve.org>u  b In article <3DCBF52B.DDFD9D19@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes: > Nic Clews wrote:J >> Perhaps this should be a warning to those who combine a development and- >> production environment on the same system.1 > K > Since pipeline systems would be operating in real time (in this case, thenM > software didn't have the quick enough response to close some valve before a M > disaster occured), how does VMS behave when an interactive user logs on and F > does stuff on the system with regards to the real time application ? > P > If a real time process is at priority 17, and an interactive process starts at7 > priority 4, can the later really impact the former ? o > H > (OK, I realise that if you lock some file/record, or that if you do anU > ana/disk/repair, it will lock out any process from accessing the blocked resource).o > ? Or if you need to page in (PFW) or page out (RWMPB).  Or if younE need to swap in (COMO).  But then, you typically put a bit of thought-@ into locking pages into the working set and turning off swapping1 when you set up a process running at priority 17.   N > But in terms of CPU, shouldn't a priority > 16 process really get the CPU it > needs, when it needs it ?d  D Yes, but what makes you think that the real time application in this case was not I/O bound?w  1 "Real time" is not a synonym for "priority >= 16"   B A "real time" system is one which has critical timing constraints.  D If you get your time card entries in on Friday at 17:30 and you needC to generate the payroll tape for the FedEx man by Saturday at 08:00a* then you are operating a real time system.  ? If you need to evaluate pressure sensors at grid coordinates A, > B and C, compare against scheduled outages in the database and< open valves D, E and F to alleviate a dangerous overpressure@ situation within 10 seconds, you are running a real time system.  D That doesn't mean that you should be running payroll at priority 16.? And that doesn't mean that your pipeline outage database cannote be on disk.s   	John Briggs   ------------------------------   Date: 8 Nov 2002 06:21:26 -0600e From: briggs@encompasserve.org1 Subject: Re: Vax Macro Porting - Still Strugglingh3 Message-ID: <zTPlD5G3PFZK@eisner.encompasserve.org>s  f In article <99c3a525.0211071221.636daae7@posting.google.com>, aaron.d.mullens@lmco.com (Aaron) writes:C > Thanks for the previous help.  I have been looking into the Alpha G > calling standards.  I am still unsure of what to do.  I was wonderingaH > if anyone could help me get started on this task by looking at a small> > segment of macro code for me and giving me some suggestions. > D > ****************************************************************** > Initial_FP: .long 00 > * >   psect $CODE,long,exe,pic,rel,shr,nowrt >   * >   .entry $Initialize_Tasking,^m<R10,R11>  A Since someone was apparently attempting to implement a home-grown A threads library here, have you considered using a vendor-suppliedC threads library instead?   	John Briggs   ------------------------------   Date: 8 Nov 2002 06:34:19 -0600o- From: Kilgallen@SpamCop.net (Larry Kilgallen)a1 Subject: Re: Vax Macro Porting - Still Strugglinge3 Message-ID: <vDI5DiPY$vFd@eisner.encompasserve.org>   T In article <zTPlD5G3PFZK@eisner.encompasserve.org>, briggs@encompasserve.org writes:h > In article <99c3a525.0211071221.636daae7@posting.google.com>, aaron.d.mullens@lmco.com (Aaron) writes:D >> Thanks for the previous help.  I have been looking into the AlphaH >> calling standards.  I am still unsure of what to do.  I was wonderingI >> if anyone could help me get started on this task by looking at a smalls? >> segment of macro code for me and giving me some suggestions.s >>  E >> ******************************************************************r >> Initial_FP: .long 0 >>  + >>   psect $CODE,long,exe,pic,rel,shr,nowrta >>    + >>   .entry $Initialize_Tasking,^m<R10,R11>o > C > Since someone was apparently attempting to implement a home-growneC > threads library here, have you considered using a vendor-suppliedn > threads library instead?  D Possibly because it would increase the difficulty of the programming effort for the port.   ------------------------------   Date: 8 Nov 2002 07:25:04 -0800 & From: aaron.d.mullens@lmco.com (Aaron)1 Subject: Re: Vax Macro Porting - Still Strugglinge= Message-ID: <99c3a525.0211080725.4bfee8d8@posting.google.com>"  E This function is called from Ada code and a stack top offset is being E passed in as a parameter.  I have been looking at the openvms callinguC standards document on the compaq openvms site.  Does anyone know ofdD other sites that show examples that may help ease the learning curve( for me on walking the stack on an Alpha?  C Yes that was a typo.  I meant to type L at the begging of _SAVE_FP.n   ------------------------------  # Date: Thu, 07 Nov 2002 23:18:46 GMTt& From: John Reagan <john.reagan@hp.com>1 Subject: Re: Vax Macro Porting - Still Struggling03 Message-ID: <qxCy9.55$TE5.1912966@news.cpqcorp.net>r   Tom Linden wrote:H? > It looks like you are just unwinding stacks to the top level.bF > Look at the stack frame layout for Alpha, and find the correspondingC > offset from the frame ptr where the old one is kept, this will becJ > the equivalent of SF$L_SAVE_FP  (you permuted L_ in your code fragment?)  F It is much more complicated than that.  The FP (R29) points to either H points to a procedure descriptor or a pointer to a quadword that points G to a procedure descritpor.  Once you have the PD, you then have to see  $ what kind it is (stack or register).  G For the stack type, you then have to dig around and find the offset of lE the register save area.  The size of the area is based on the actual iG mask of registers saved.  You have to dig around to find where the R29 m  of the previous frame was saved.  H For the register type, the previous FP is saved in another register not G the stack (that is why it is a register frame not a stack frame).  The  E register PD will tell you which other register holds the previous FP.t  J As Aaron wants, you really want to use the LIB$ Calling Standard routines.  G I don't have an example handy.  If nobody else comes up with one, I'll  8 try to find 15 minutes tomorrow and hack one up for you.     -- a John ReaganP' Compaq Pascal/{A|I}MACRO Project Leader  Hewlett-Packard Company1   ------------------------------  $ Date: Fri, 8 Nov 2002 08:22:51 -0800# From: "Tom Linden" <tom@kednos.com>D1 Subject: RE: Vax Macro Porting - Still Strugglingc9 Message-ID: <CIEJLCMNHNNDLLOOGNJIKELGGAAA.tom@kednos.com>d   >-----Original Message-----o. >From: John Reagan [mailto:john.reagan@hp.com]* >Sent: Thursday, November 07, 2002 3:19 PM >To: Info-VAX@Mvb.Saic.Com2 >Subject: Re: Vax Macro Porting - Still Struggling >n >s >Tom Linden wrote:@ >> It looks like you are just unwinding stacks to the top level.G >> Look at the stack frame layout for Alpha, and find the correspondingPD >> offset from the frame ptr where the old one is kept, this will beK >> the equivalent of SF$L_SAVE_FP  (you permuted L_ in your code fragment?)n >rG >It is much more complicated than that.  The FP (R29) points to either iI >points to a procedure descriptor or a pointer to a quadword that points 0H >to a procedure descritpor.  Once you have the PD, you then have to see % >what kind it is (stack or register).'  B Looks vaguely reminiscent of Hennessy.  I thought it might be more6 involved.  So how do you know what r29 is pointing to? >yH >For the stack type, you then have to dig around and find the offset of F >the register save area.  The size of the area is based on the actual H >mask of registers saved.  You have to dig around to find where the R29 ! >of the previous frame was saved.d >aI >For the register type, the previous FP is saved in another register not  H >the stack (that is why it is a register frame not a stack frame).  The F >register PD will tell you which other register holds the previous FP. >.K >As Aaron wants, you really want to use the LIB$ Calling Standard routines.  >yH >I don't have an example handy.  If nobody else comes up with one, I'll 9 >try to find 15 minutes tomorrow and hack one up for you." >e >  >--  >John Reagan( >Compaq Pascal/{A|I}MACRO Project Leader >Hewlett-Packard Company >v >---' >Incoming mail is certified Virus Free.w; >Checked by AVG anti-virus system (http://www.grisoft.com).1B >Version: 6.0.410 / Virus Database: 231 - Release Date: 10/31/2002 >i ---i& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.410 / Virus Database: 231 - Release Date: 10/31/2002    ------------------------------  # Date: Fri, 08 Nov 2002 17:46:09 GMTt" From:   VAXman-  @SendSpamHere.ORG1 Subject: RE: Vax Macro Porting - Still StrugglingF0 Message-ID: <00A16AAD.48D43F2A@SendSpamHere.ORG>  _ In article <CIEJLCMNHNNDLLOOGNJIKELGGAAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes:k >o >  >>-----Original Message-----/ >>From: John Reagan [mailto:john.reagan@hp.com]n+ >>Sent: Thursday, November 07, 2002 3:19 PMi >>To: Info-VAX@Mvb.Saic.Coma3 >>Subject: Re: Vax Macro Porting - Still Strugglingn >> >> >>Tom Linden wrote:1A >>> It looks like you are just unwinding stacks to the top level.>H >>> Look at the stack frame layout for Alpha, and find the correspondingE >>> offset from the frame ptr where the old one is kept, this will beaL >>> the equivalent of SF$L_SAVE_FP  (you permuted L_ in your code fragment?) >>H >>It is much more complicated than that.  The FP (R29) points to either J >>points to a procedure descriptor or a pointer to a quadword that points I >>to a procedure descritpor.  Once you have the PD, you then have to see i& >>what kind it is (stack or register). >-C >Looks vaguely reminiscent of Hennessy.  I thought it might be morer7 >involved.  So how do you know what r29 is pointing to?:  G If 0(FP) points to a quadword which points to the procedure descriptor, G the value of 0(FP) will have bits <2:0> equal to 0 because the quadwordoF must be naturally aligned.  If bits <2:0> are NOT equal to zero, 0(FP)' FP points to the procedure descriptor.    G This, and much more, is detailed in the calling standard documents.  IfeG you really want to port this code you are going to have to read and un-g derstant the calling standard.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?" i   ------------------------------  $ Date: Fri, 8 Nov 2002 09:58:23 -0800# From: "Tom Linden" <tom@kednos.com>n1 Subject: RE: Vax Macro Porting - Still Struggling 9 Message-ID: <CIEJLCMNHNNDLLOOGNJIKELJGAAA.tom@kednos.com>s   >-----Original Message-----mC >From: VAXman- @SendSpamHere.ORG [mailto:VAXman- @SendSpamHere.ORG] ( >Sent: Friday, November 08, 2002 9:46 AM >To: Info-VAX@Mvb.Saic.Com2 >Subject: RE: Vax Macro Porting - Still Struggling >s >e@ >In article <CIEJLCMNHNNDLLOOGNJIKELGGAAA.tom@kednos.com>, "Tom ! >Linden" <tom@kednos.com> writes:d >> >> >>>-----Original Message----- 0 >>>From: John Reagan [mailto:john.reagan@hp.com], >>>Sent: Thursday, November 07, 2002 3:19 PM >>>To: Info-VAX@Mvb.Saic.Com4 >>>Subject: Re: Vax Macro Porting - Still Struggling >>>s >>>h >>>Tom Linden wrote:B >>>> It looks like you are just unwinding stacks to the top level.I >>>> Look at the stack frame layout for Alpha, and find the correspondingdF >>>> offset from the frame ptr where the old one is kept, this will beC >>>> the equivalent of SF$L_SAVE_FP  (you permuted L_ in your code t >fragment?)s >>>iI >>>It is much more complicated than that.  The FP (R29) points to either SK >>>points to a procedure descriptor or a pointer to a quadword that points MJ >>>to a procedure descritpor.  Once you have the PD, you then have to see ' >>>what kind it is (stack or register).  >>D >>Looks vaguely reminiscent of Hennessy.  I thought it might be more8 >>involved.  So how do you know what r29 is pointing to? >gH >If 0(FP) points to a quadword which points to the procedure descriptor,H >the value of 0(FP) will have bits <2:0> equal to 0 because the quadwordG >must be naturally aligned.  If bits <2:0> are NOT equal to zero, 0(FP)o( >FP points to the procedure descriptor.   G Its not me porting the code, I (unwisely) jumped in.  So, Proc descripsa are never quad aligned?e >rH >This, and much more, is detailed in the calling standard documents.  IfH >you really want to port this code you are going to have to read and un- >derstant the calling standard.f >t >--w8 >VAXman- OpenVMS APE certification number: AAA-0001      >VAXman(at)TMESIS(dot)COM  >           6 >  "Well my son, life is like a beanstalk, isn't it?"  >  >---' >Incoming mail is certified Virus Free.a; >Checked by AVG anti-virus system (http://www.grisoft.com). B >Version: 6.0.410 / Virus Database: 231 - Release Date: 10/31/2002 >  ---d& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.410 / Virus Database: 231 - Release Date: 10/31/2002a   ------------------------------  + Date: Fri, 08 Nov 2002 14:33:09 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>,C Subject: VMS machines behind DSL router---do I need PAT or just NAT.; Message-ID: <01KOMHF7F7YGA23BZ7@sysdev.deutsche-boerse.com>e  H Experts, please correct me if I am wrong in any of my assumptions while " you answer the questions!  Thanks!  H I will probably soon provide internet access to my hobbyist VMS cluster B via a DSL router (mainly to save costs since there are attractive  flat-rate offers about).    E At the moment, I have fixed, routable "real" IP addresses and an ISDNs2 router which does dial-out AND dial-IN on demand.   G I am perfectly happy with this setup from a technical point of view.  I-$ am moving to DSL to decrease costs.   F With a DSL connection, I will have a single IP address visible to the ) outside world, namely that of the router.E  H One reason for the original setup was that I wanted to be able to accessG all machines from outside, even though in normal operation it is enoughcD to access the cluster alias.  Now that this works fine, I could justA point the visible IP address to the cluster alias then connect tolF another machine once I get inside.  Other reasons were the lack of DSLF at the location at the time, and the fact that I wasn't aware that theG possibility of "dynamic" DNS exists so that a changing IP address is no  real problem.   G A fourth reason is that, in the other direction, more than one address eD needs to access the internet (even with a cluster alias set up, for G example, when I send email it "comes from" a specific machine, not the oD cluster alias, i.e. the cluster alias is just for incoming, not for A outgoing, stuff).  This is the main point where I have questions.n  F Looking at http://www.garykessler.net/library/tcpip.html#IPcidr I haveH the impression that NAT (which the obvious routers I would buy appear toG support) would be enough AS LONG AS ONLY ONE MACHINE AT A TIME ACCESSES H THE INTERNET.  If, however, more than one machine at a time accesses the- internet, then I need PAT.  Is this correct? W  G How big a worry is the "performance hit" due to NAT and especially PAT - mentioned in the article above?e  F If I need PAT but NAT would be better, for whatever reasons, is there A any way to get the cluster-alias address to be used for OUTgoing U connections as well?  G In the end, I would use non-routable internal addresses, with only the aF router address being visible to the outside (and assigned dynamically H when the DSL connection is initiated), thus I wouldn't need to have any F real addresses of my own, as I have now.  For a transitional time, of H course, it would be nice to leave the current ISDN connection as is and I have the DSL functionality as well (at least for outgoing connections).  iF Is it enough to just put the DSL router on the LAN, give it one of my F "real" addresses (or could I give it a non-routable address while the H other machines on the LAN have routable ones) and use this for outgoing E connections (keeping incoming ones, for the time being, via the ISDN tG router---or of course allowing additional incoming connections through r: the DSL router, while keeping the ISDN router functional)?   ------------------------------  # Date: Fri, 08 Nov 2002 14:29:08 GMTt# From: "John Smith" <a@nonymous.com>eG Subject: Re: VMS machines behind DSL router---do I need PAT or just NAT2H Message-ID: <USPy9.18530$MGm1.1826@news02.bloor.is.net.cable.rogers.com>  F "Phillip Helbig" <HELBPHI@sysdev.deutsche-boerse.com> wrote in message5 news:01KOMHF7F7YGA23BZ7@sysdev.deutsche-boerse.com...h >'H > Looking at http://www.garykessler.net/library/tcpip.html#IPcidr I haveJ > the impression that NAT (which the obvious routers I would buy appear toI > support) would be enough AS LONG AS ONLY ONE MACHINE AT A TIME ACCESSESeJ > THE INTERNET.  If, however, more than one machine at a time accesses the. > internet, then I need PAT.  Is this correct?  H Most of the lower cost NAT routers support one 'DMZ' port address, whichK will allow only one of your machines to be exposed directly to the outside.TI Usually it's simply a case of entering in the IP address of that specific  machine into the router config.   K With NAT, only the machine entered into the DMZ will be able to be directlyiE accessed remotely. What you do on that machine to control any further I penetration into the 'private side' of your network/cluster is up to you.o  E A reasonable source for information about low-cost routers (wired andu0 wireless) is http://www.practicallynetworked.com   ------------------------------  % Date: Fri, 08 Nov 2002 10:02:32 -05002* From: "Stanley F. Quayle" <stan@stanq.com>G Subject: Re: VMS machines behind DSL router---do I need PAT or just NAT . Message-ID: <3DCB8BB8.12938.28A8413@localhost>  ? My router allows me to direct selected incoming connections to b addresses inside my network.  D For example, you could have port 1223 directed to 192.168.1.10 port E 23, and port 1323 directed to 192.168.1.10 port 23.  This will allow eD you to telnet to either machine.  Or both simultaneously.  I have a 8 Belkin cable router which allows this sort of operation.  C As for routing out ISDN and DSL simultaneously, you'll need a much hA more flexible router.  I use a Linux system, and route between a fC cable modem (Internet) and a customer's private dial-in connection  , (they use an "unroutable" set of addresses).  
 --Stan Quayleo Quayle Consulting Inc.  
 ----------C Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363  Fax: +1 614 868-1671v1 8572 North Spring Ct. NW, Pickerington, OH  43147w= Preferred address:  stan@stanq.com       http://www.stanq.come   ------------------------------  + Date: Fri, 08 Nov 2002 16:08:03 +0100 (MET)e9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>hG Subject: Re: VMS machines behind DSL router---do I need PAT or just NAT ; Message-ID: <01KOMQBGJYQMA23BZ7@sysdev.deutsche-boerse.com>i  J > Most of the lower cost NAT routers support one 'DMZ' port address, whichM > will allow only one of your machines to be exposed directly to the outside.lK > Usually it's simply a case of entering in the IP address of that specifica! > machine into the router config.  > M > With NAT, only the machine entered into the DMZ will be able to be directly-G > accessed remotely. What you do on that machine to control any furthersK > penetration into the 'private side' of your network/cluster is up to you.$  D With the cluster alias, I have all the access INTO the local networkG that I need.  The problem is that I need to allow more than one machinecI access FROM the network, also simultaneously.  Does this definitely mean e NAT instead of PAT?d   ------------------------------  + Date: Fri, 08 Nov 2002 16:13:00 +0100 (MET)t9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> G Subject: Re: VMS machines behind DSL router---do I need PAT or just NATp; Message-ID: <01KOMQHOT3NOA23BZ7@sysdev.deutsche-boerse.com>   A > My router allows me to direct selected incoming connections to r > addresses inside my network. > F > For example, you could have port 1223 directed to 192.168.1.10 port G > 23, and port 1323 directed to 192.168.1.10 port 23.  This will allow eF > you to telnet to either machine.  Or both simultaneously.  I have a : > Belkin cable router which allows this sort of operation.  E Right.  I wouldn't do this, just move all connections to the cluster tI alias.  If I really need to get to a different machine, I can connect up e and then connect locally to it.   E > As for routing out ISDN and DSL simultaneously, you'll need a much u > more flexible router.     
 Why, exactly?e   ------------------------------  % Date: Fri, 08 Nov 2002 11:04:28 -05000* From: "Stanley F. Quayle" <stan@stanq.com>G Subject: Re: VMS machines behind DSL router---do I need PAT or just NAT . Message-ID: <3DCB9A3C.27292.2C3390B@localhost>  - On 8 Nov 2002 at 16:13, Phillip Helbig wrote:nG > > As for routing out ISDN and DSL simultaneously, you'll need a much - > > more flexible router.  e > Why, exactly?   > Consumer-grade routers assume that there's only one "outside"  address.  F If you're running a commercial-grade router (Cisco, Linux, etc.), you C can make routing more complicated.  You can route packets based on XE their source and destination IP addresses and port numbers.  You can n5 also spend lots of hours messing with routing tables.   D I'm not familiar with "PAT".  But NAT (also known as IP Masquerade) D should be all you need.  All machines inside your network can reach ? out to the Internet, and only certain incoming connections are cE allowed.  This is exactly what I do at my site -- I have 20 machines a3 all accessing the Internet through one cable modem.e
 --Stan Quayley Quayle Consulting Inc.  
 ----------C Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363  Fax: +1 614 868-1671 1 8572 North Spring Ct. NW, Pickerington, OH  43147 = Preferred address:  stan@stanq.com       http://www.stanq.com    ------------------------------  + Date: Fri, 08 Nov 2002 17:12:32 +0100 (MET)-9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>nG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATy; Message-ID: <01KOMSDPGGXAA23BZ7@sysdev.deutsche-boerse.com>s  F > Yes, you want PAT. NAT allows a static translation between an insideF > address, and one of a block of "outside" routable addresses. It alsoF > allows for both inbound and outbound access. PAT allows for multipleJ > inside addresses to use a single outside address for outbound access. IfD > you have only one outside address, you'll need to use PAT to allow> > multiple nodes to make outbound connections simultaneously.   6 OK, that's what I needed (if not wanted :-( ) to hear.  I Step 2: can anyone suggest a suitable DNS router, preferably someone who  E actually has more than one VMS machine connect simultaneously to the oB internet with the connection being initiated by the VMS machine?  D Requirement: it has to be configurable via a menu accessed via a VT  terminal through a serial line.t  C Presumably, I can just get a vanilla DSL connection first then the ,
 router later.:  F Considering that VMS TCPIP now has a DHCP client, and considering the F fact that (at least some---including probably the one I will have---) F vanilla DSL connections use DHCP, is there any chance of connecting a H VMS box "directly" to some standard DSL provider?  (This is more out of < curiosity---as soon as possible I want the router solution.)   ------------------------------  + Date: Fri, 08 Nov 2002 17:15:50 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>hG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATi; Message-ID: <01KOMSN2211WA23BZ7@sysdev.deutsche-boerse.com>   @ > Consumer-grade routers assume that there's only one "outside" 
 > address.  I That is the case here.  The assumption made by consumer-grade routers is (H probably that I have several machines that need outgoing connections AT  THE SAME TIME.  H > If you're running a commercial-grade router (Cisco, Linux, etc.), you E > can make routing more complicated.  You can route packets based on eG > their source and destination IP addresses and port numbers.  You can s7 > also spend lots of hours messing with routing tables.   ! That's probably MORE than I need.o  F > I'm not familiar with "PAT".  But NAT (also known as IP Masquerade) F > should be all you need.  All machines inside your network can reach A > out to the Internet, and only certain incoming connections are  G > allowed.  This is exactly what I do at my site -- I have 20 machines  5 > all accessing the Internet through one cable modem.   I See Mark Levy's response---it sounds like I really need PAT.  Do your 20 lI machines actually access the internet at the same time, or just one at a b time?h   ------------------------------  + Date: Fri, 08 Nov 2002 17:19:24 +0100 (MET),9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>aG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATr; Message-ID: <01KOMST5IQDMA23BZ7@sysdev.deutsche-boerse.com>   D Here's Mark Levy's response, which I mentioned in my last post.  It G appears to have gone just to me, so I'm reposting it here, assuming he   won't mind.e  J ---------------8<---------------------------------------------------------  D Yes, you want PAT. NAT allows a static translation between an insideD address, and one of a block of "outside" routable addresses. It alsoD allows for both inbound and outbound access. PAT allows for multipleH inside addresses to use a single outside address for outbound access. IfB you have only one outside address, you'll need to use PAT to allow< multiple nodes to make outbound connections simultaneously.    ML   ------------------------------  * Date: Fri, 8 Nov 2002 17:19:52 +0000 (UTC)+ From: david20@alpha2.mdx.ac.uk (David Webb)iG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATd+ Message-ID: <aqgrno$h94$1@aquila.mdx.ac.uk>f  Y In article <kbRy9.10747$O71.5436@rwcrnsc53>, "Mark E. Levy" <levy@sysman-inc.com> writes:yE >Yes, you want PAT. NAT allows a static translation between an inside^L >address, and one of a block of "outside" routable addresses. It also allowsE >for both inbound and outbound access. PAT allows for multiple insideaK >addresses to use a single outside address for outbound access. If you haveoL >only one outside address, you'll need to use PAT to allow multiple nodes to* >make outbound connections simultaneously. >r This isn't quite correct.   < NAT can be configured with either static or dynamic mapping.  G With static mapping you associate each individual inside address with a D specific outside address - this also allows outside users to connect to your internal systems.  rL With dynamic mapping each inside system is dynamically allocated an outside J address from the block. These mappings timeout and can be reused by other C internal systems. This allows you to have a large number of fairly  L infrequently used internal systems using a relatively small pool of outside G addresses. Static mapping requires a one-to-one correspondence between t  internal and external addresses.I You can also mix and match static and dynamic mappings ie have one staticsN mapping for a webserver you want visible to the outside world and use the restK of your external block with dynamic mapping to your other internal systems.t    O However, as you say, if you only have one external address and multiple systemsuJ that want to communicate to the internet simultaneously then you will need3 to use Port as well as Network Address translation.C  
 David Webb VMS and Unix team leader CCSS Middlesex University i     >MLo >rG >"Phillip Helbig" <HELBPHI@sysdev.deutsche-boerse.com> wrote in messagem6 >news:01KOMQBGJYQMA23BZ7@sysdev.deutsche-boerse.com...M >> > Most of the lower cost NAT routers support one 'DMZ' port address, whichhG >> > will allow only one of your machines to be exposed directly to thee	 >outside.=E >> > Usually it's simply a case of entering in the IP address of that 	 >specific $ >> > machine into the router config. >> >G >> > With NAT, only the machine entered into the DMZ will be able to beo	 >directlyoJ >> > accessed remotely. What you do on that machine to control any furtherI >> > penetration into the 'private side' of your network/cluster is up tou >you.  >>G >> With the cluster alias, I have all the access INTO the local network0J >> that I need.  The problem is that I need to allow more than one machineK >> access FROM the network, also simultaneously.  Does this definitely meanf >> NAT instead of PAT? >> >r >a   ------------------------------  * Date: Fri, 8 Nov 2002 17:36:58 +0000 (UTC)+ From: david20@alpha2.mdx.ac.uk (David Webb)lG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATs+ Message-ID: <aqgsnq$hi3$1@aquila.mdx.ac.uk>a  [ In article <3DCB9A3C.27292.2C3390B@localhost>, "Stanley F. Quayle" <stan@stanq.com> writes: . >On 8 Nov 2002 at 16:13, Phillip Helbig wrote:H >> > As for routing out ISDN and DSL simultaneously, you'll need a much  >> > more flexible router.   >> Why, exactly? >P? >Consumer-grade routers assume that there's only one "outside" i	 >address.o >oG >If you're running a commercial-grade router (Cisco, Linux, etc.), you  D >can make routing more complicated.  You can route packets based on F >their source and destination IP addresses and port numbers.  You can 6 >also spend lots of hours messing with routing tables. >eE >I'm not familiar with "PAT".  But NAT (also known as IP Masquerade) .E >should be all you need.  All machines inside your network can reach g@ >out to the Internet, and only certain incoming connections are F >allowed.  This is exactly what I do at my site -- I have 20 machines 4 >all accessing the Internet through one cable modem. >--Stan Quayle >Quayle Consulting Inc.  >t  N In that case you are almost certainly using port translation as well as basic D NAT. This is variously known as NATP , NAT with port overloading etc   NAT just translates addresses    ie    1 Private address                    Public address   - 10.5.3.1                   ->      158.94.7.5e- 10.5.3.2                   ->      158.94.7.7h      $ NATP translates port numbers as well  2 10.5.3.1:1250               ->     158.94.7.5:15002 10.5.3.1:1300               ->     158.94.7.5:16002 10.5.3.2:1400               ->     158.94.7.5:1250  I And static NATP mappings can allow access remotely to certain services on  internal systems ege    0 10.5.3.1:25                 <-     158.94.7.5:250 10.5.3.2:80                 <-     158.94.7.5:80  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Fri, 08 Nov 2002 13:05:00 -0500h0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>G Subject: Re: VMS machines behind DSL router---do I need PAT or just NAT*/ Message-ID: <3DCBFCC9.65C77BFB@vl.videotron.ca>-   Phillip Helbig wrote:aH > Looking at http://www.garykessler.net/library/tcpip.html#IPcidr I haveJ > the impression that NAT (which the obvious routers I would buy appear toI > support) would be enough AS LONG AS ONLY ONE MACHINE AT A TIME ACCESSESiJ > THE INTERNET.  If, however, more than one machine at a time accesses the. > internet, then I need PAT.  Is this correct?  L Perhaps this is a matter of semantics, but "NAT" as is marketed by makers of7 home routers, allows multiple simultaneous connections.d  L Your router will make it look to the internet like there is only a single IP* address and a single machine on your line.  F Each of your machines on your lan will have a private address, usuallyG non-routable address (for instance 10.*.*.*). Your router will have two J interfaces. On the LAN side, it will have an address in the same subnet asO your computers. On the WAN side, it will get whatever address the ISP gives it.   H > How big a worry is the "performance hit" due to NAT and especially PAT! > mentioned in the article above?o  G Shouldn't be a problem, but when you choose your router, try to see the M reviews. I know that the one I had (netgear RT314) had comments such as "goodr. trhoughput, and lots of spare CPU" in reviews.  I Note that some DSL providers insist on downgrading service with the addedcL PPPoE layer. This slows down the line, and adds overhead on the router whichN much repackage packets. (and isn't truly "alwasy on" since you must log in and; out via PPP commands (the router does this for you though).   G > If I need PAT but NAT would be better, for whatever reasons, is thereeB > any way to get the cluster-alias address to be used for OUTgoing > connections as well?  I No matter which of your LAN's machine makes the outgoing call, the targetkJ computer will get a request coming from the one IP negotiated betwene your router and the ISP.i  K On your router, you will then define how to handle inbound connections. YouaK can direct port 80 connections to the cluster alias IP address, and port 21iK connections to a specific machine. But you cannot, from the outside, chooseoJ which machine on your lAN you want to connect to. There is a predeterminedF table in your router which decides where each type of connection goes.  M So if 2 machines want to play glames where the outside attempst to connect toBN the computer, then only one machine can play a time, and you need to configureM the router so that connection requests for port XXX needed by that game go toy the right machine.   ------------------------------  % Date: Fri, 08 Nov 2002 13:17:21 -050060 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>G Subject: Re: VMS machines behind DSL router---do I need PAT or just NATn/ Message-ID: <3DCBFFAD.53F9B192@vl.videotron.ca>   , re: having a DSL and ISDN services co-exist.  M You can then buy 2 "consumer" routers. One NAT for your DSL, and one for youriL ISDN lines. You can then setup the routing on your VMS machines. ConnectionsL to certain IP adresses will be routed through one router and all the rest to the other router.n   ------------------------------  % Date: Fri, 08 Nov 2002 13:27:31 -0500a0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>G Subject: Re: VMS machines behind DSL router---do I need PAT or just NATn. Message-ID: <3DCC020E.CB84471@vl.videotron.ca>   Phillip Helbig wrote:tJ > Step 2: can anyone suggest a suitable DNS router, preferably someone whoF > actually has more than one VMS machine connect simultaneously to theB > internet with the connection being initiated by the VMS machine?E > Requirement: it has to be configurable via a menu accessed via a VTr! > terminal through a serial line.@  L Netgear RT314 can be configured from http (poor), telnet or serial port. YouK can also "hack" it to turn it into a Zyxel machine which has more functionsu& (such as SNMP). haven't done mine yet.  M It supports FTP, TFTP and Xmodem to download/upload the rom and config files.   E Its implementation of "NAT" allows multiple machines to make outbound  connections at the same time.d  N Consumer grade routers seem to use "NAT" as terminology. Perhaps in reality is2 it "PAT", but it does the job you want them to do.  D > Presumably, I can just get a vanilla DSL connection first then the > router later.>  K I wouldn't recommend it. It will be far more trouble for your LAN to have a J machine with a variable IP address, setting up the DHCP client on only one machine etc etc.  G > fact that (at least some---including probably the one I will have---)oG > vanilla DSL connections use DHCP, is there any chance of connecting aP4 > VMS box "directly" to some standard DSL provider?   L If your DSL provider uses PPPoE, then you CANNOT plug in your VMS machine toH the DSL line directly. Those services do not support TCPIP, they supportM PPPoE, and VMS doesn't support PPPoE. You absolutely need a router capable ofoN this (the router will translate your TCPIP packet into your ISP's PPPoE packet format and vice-versa).t    H Also, if your DSL service uses PPPoE, you should contact their technicalM support and ask if the router you intend to buy is known to work. I know thateK the Netgear has "patches" that you enable via menus when connecting to some*. ISP services (notably Road Runner in the USA).   ------------------------------  $ Date: Fri, 8 Nov 2002 13:38:14 -0500% From: "John Vottero" <John@mvpsi.com>aG Subject: Re: VMS machines behind DSL router---do I need PAT or just NATe/ Message-ID: <uso14na4omvsfc@news.supernews.com>   3 Most router vendors misuse and abuse NAT/PAT names.n  K I've used a lot of cheap routers and all of them would do what you need and-F most of them call it NAT, even when they support NAT and PAT and other	 variants.m  F "Phillip Helbig" <HELBPHI@sysdev.deutsche-boerse.com> wrote in message5 news:01KOMST5IQDMA23BZ7@sysdev.deutsche-boerse.com...hE > Here's Mark Levy's response, which I mentioned in my last post.  ItaH > appears to have gone just to me, so I'm reposting it here, assuming he
 > won't mind.o >hL > ---------------8<--------------------------------------------------------- > F > Yes, you want PAT. NAT allows a static translation between an insideF > address, and one of a block of "outside" routable addresses. It alsoF > allows for both inbound and outbound access. PAT allows for multipleJ > inside addresses to use a single outside address for outbound access. IfD > you have only one outside address, you'll need to use PAT to allow= > multiple nodes to make outbound connections simultaneously.i >s > ML >t   ------------------------------   Date: 8 Nov 2002 06:51:40 -0600t- From: Kilgallen@SpamCop.net (Larry Kilgallen)h1 Subject: Re: VMS on Itanium not mission critical?a3 Message-ID: <BoC+5yZoaf8I@eisner.encompasserve.org>r  a In article <9cfnsuo1o8t7j0l4c01olu2ibam85t68g8@4ax.com>, jlsue <jlsuexxxz@screaminet.com> writes:d6 > On Thu, 07 Nov 2002 12:39:23 GMT, "Terry C. Shannon"! > <terryshannon@attbi.com> wrote:u >  >  >>L >>Dual- rail, completely redundant, two copies of VMS, etc. That didn't stopJ >>thr VAXft from going down when some bozo yankd the wrong power cord at a, >>DECUS some years back on the Left Coast... > F > Yes!  I remember that situation.  It really slowed down the check-in > process that year.  K No worse than the year the electrician wired the three phases out of order.h   ------------------------------   Date: 8 Nov 2002 13:04:37 GMT ( From: bill@cs.uofs.edu (Bill Gunshannon) Subject: VMS-6.15 Message-ID: <aqgcp4$9o6ba$1@ID-135708.news.dfncis.de>o   A quick question.t  B I am trying to set up a box real quick running VAX-VMS-6.1 and theA C Compiler.  The install of 6.1 went just fine but the C CompilermG install is not working as planned.  It appears to complete successfullyIG but none of the EXE files are present.  Is it possible that the version-? of the Layered Products is too new for this version of the OS??e   Any other suggestions??c   bill   -- oJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Fri, 08 Nov 2002 09:34:47 -0500t5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>h Subject: Re: VMS-6.12 Message-ID: <T8rLPT4UWq8HcJDRlK3I3rWiDri5@4ax.com>  	     Bill,   8     It appears that V5.7 is the latest release that will3 run on VMS V6.1, at least on Alpha.  I would not bem! suprised if it's the same on VAX.r   David R. Beattyc  E On 8 Nov 2002 13:04:37 GMT, bill@cs.uofs.edu (Bill Gunshannon) wrote:e   >A quick question. >tC >I am trying to set up a box real quick running VAX-VMS-6.1 and the B >C Compiler.  The install of 6.1 went just fine but the C CompilerH >install is not working as planned.  It appears to complete successfullyH >but none of the EXE files are present.  Is it possible that the version@ >of the Layered Products is too new for this version of the OS?? >r >Any other suggestions?? >o >billr   ------------------------------   Date: 8 NOV 2002 15:44:03 GMTd+ From: Dave Greenwood <greenwoodde@ornl.gov>v Subject: Re: VMS-6.11 Message-ID: <8NOV02.15440335@feda01.fed.ornl.gov>v  @ In a previous article, bill@cs.uofs.edu (Bill Gunshannon) wrote: > A quick question.f >  .D > I am trying to set up a box real quick running VAX-VMS-6.1 and theC > C Compiler.  The install of 6.1 went just fine but the C CompilergI > install is not working as planned.  It appears to complete successfullyoI > but none of the EXE files are present.  Is it possible that the versionrA > of the Layered Products is too new for this version of the OS??c  G According to my install manual, C v6.5 (the most recent version I have)iJ can be installed on any VAX VMS 5.5-2 through 7.3 system.  So that doesn't sound like the problem.a  H Are you sure you're looking for the correct EXE filenames?  The compilerG is SYS$SYSTEM:DECC$COMPILER (at least on alpha - I have no VAXes with C  installed).4  = Did you run the IVP to test the install?  If not, try it with-     @sys$test:decc$ivp   What does it say?2   Dave --------------9 Dave Greenwood                Email: Greenwoodde@ORNL.GOV-H Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself   ------------------------------   Date: 8 Nov 2002 16:01:30 GMT"( From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: VMS-6.15 Message-ID: <aqgn4p$9rg8k$1@ID-135708.news.dfncis.de>e  1 In article <8NOV02.15440335@feda01.fed.ornl.gov>,V. 	Dave Greenwood <greenwoodde@ornl.gov> writes:B > In a previous article, bill@cs.uofs.edu (Bill Gunshannon) wrote: >> A quick question. >>  E >> I am trying to set up a box real quick running VAX-VMS-6.1 and the D >> C Compiler.  The install of 6.1 went just fine but the C CompilerJ >> install is not working as planned.  It appears to complete successfullyJ >> but none of the EXE files are present.  Is it possible that the versionB >> of the Layered Products is too new for this version of the OS?? > I > According to my install manual, C v6.5 (the most recent version I have)uL > can be installed on any VAX VMS 5.5-2 through 7.3 system.  So that doesn't > sound like the problem.e  - Damn, I knew it wouldn't be something simple.v   > J > Are you sure you're looking for the correct EXE filenames?  The compilerI > is SYS$SYSTEM:DECC$COMPILER (at least on alpha - I have no VAXes with Ce
 > installed).a  / From SYS$SYSROOT:[SYSUPD.CC056]CCXX$STARTUP.COMg3   (which seems to be about I  get fromt he install)p  9 $ CALL SAFE_INSTALL_SHARED   SYS$SYSTEM:DECC$COMPILER.EXEi9 $ CALL SAFE_INSTALL_SHARED   SYS$MESSAGE:DECC$MSG_SHR.EXEl   $ dir sys$system:decc*.*! %DIRECT-W-NOFILES, no files foundg $ dir sys$message:decc*.*d! %DIRECT-W-NOFILES, no files foundt     > ? > Did you run the IVP to test the install?  If not, try it with   F I chose "YES" when it asked during the install.  The last thing on the. screen was a message saying it was successful.   >  >   @sys$test:decc$ivp >  > What does it say?:     @sys$test:decc$ivpH %DCL-E-OPENIN, error opening SYS$SYSROOT:[SYSTEST]DECC$IVP.COM; as input -RMS-E-FNF, file not found      B Like I said, it looks like it didn't put any of the right files on; the system (except for the startup file for the compiler!!).  G I'm baffled.  Were there any LP Kits that were known to have problems??rJ (I believe there was one documented case of a problem with the COBOL Kit.)   bill   -- sJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   m   ------------------------------  % Date: Fri, 08 Nov 2002 12:43:35 -0500 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca> Subject: Re: VMS-6.1/ Message-ID: <3DCBF7C5.F41C89F8@vl.videotron.ca>t   Bill Gunshannon wrote:D > I am trying to set up a box real quick running VAX-VMS-6.1 and theC > C Compiler.  The install of 6.1 went just fine but the C Compiler $ > install is not working as planned.  ( You need a patch for BACKUP for VMS 6.1.   ------------------------------   Date: 8 Nov 2002 18:59:00 GMTo, From: bill@gw5.cs.uofs.edu (Bill Gunshannon) Subject: Re: VMS-6.15 Message-ID: <aqh1hk$9v1bu$1@ID-135708.news.dfncis.de>.  / In article <3DCBF7C5.F41C89F8@vl.videotron.ca>,o3 	JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes:n > Bill Gunshannon wrote:E >> I am trying to set up a box real quick running VAX-VMS-6.1 and the:D >> C Compiler.  The install of 6.1 went just fine but the C Compiler% >> install is not working as planned.  > * > You need a patch for BACKUP for VMS 6.1.  ? If I scarf a copy of BACKUP.EXE from 7.1 will it work and solvea this problem??   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   0   ------------------------------  $ Date: Thu, 7 Nov 2002 16:22:15 -0500  From: John Santos <JOHN@egh.com>- Subject: Re: Was OpenVMS left out on purpose?a4 Message-ID: <1021107161715.400A-100000@Ives.egh.com>  % On Thu, 7 Nov 2002, John Smith wrote:a   > > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message: > news:m1ty9.115544$bt.179925@rwcrnsc52.ops.asp.att.net... > >o2 > > "John Smith" <a@nonymous.com> wrote in messageG > > news:QlSx9.197587$%h2.53842@news02.bloor.is.net.cable.rogers.com...0 > > >0B > > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message> > > > news:n9Rx9.106352$wG.405791@rwcrnsc51.ops.asp.att.net...N > > > > VMS was mentioned a total of ONCE at the analyst briefing last week in > > > Santat > > > > Barbara. > > >  > > >g > > > As in,# > > > a) OpenVMS is owned by HP, ore* > > > b) OpenVMS will be retired by HP, or< > > > c) OpenVMS was acquired by HP in the Compaq merger, or > > > d) something more vague? > >eN > > Umm, a bit more vague. Marcello was supposed to be at the meeting, but had > a 
 > > conflict.s > >nL > > Seems to me (and I could be wrong) that the IPF port proceeds apace, COEK > > ensures 15 years longevity (although support for non Air Force GeneralsLJ > > could end in 2011), and that NEW customers will be encouraged to adoptG > > HP-UX. That's about all I got. No info on early retirement, in facts	 > OpenVMSoB > > Engineering has been largely exempt from workforce reductions. >  >  > Terry, > N > What gets me about the 2011 date is that the FTC has a ruling or legislationJ > (at least I think it falls under the FTC) that deals with issues of whatJ > happens when your car or computer manufacturer discontinues a particularK > model. The vendor is required to stock parts for repair/replacement for a N > period of 5 years form the date of the product termination. Arguably one canH > call this 'support'. No new engineering is done, no design defects areH > remedied, they just ship you an identical part to the one that failed. > : > Now lets skip over to software and HP's words about VMS. > 2 > Support until 2011...what does this really mean? > J > Does it mean the all new development on VMS ceases in 2006, and that theK > 'support' window kicks in after that for the next 5 years. If so, all one/K > could expect from HP in 2007-2011 is a new CD of VMS and layered productsIM > frozen in time as of 2006, and a friendly voice at the end of the telephone D > line asking you for your credit card number and a ship-to address. > N > If this is what HP means by support to 2011, everyone using VMS better startM > looking for a different hardware and o/s vendor today so you can finish allt > your porting by 2006.   A I understand this is a "not before" date, not a "not after" date.   C I.E.  They haven't promised they will continue supporting VMS after @ this date, which is an entirely different thing than saying theyA won't support it after this date.  (This applies to both the 2006d2 development cut-off and the 2011 support cut-off.)  N > As to the new customers being shown by HP to the PH-UX door, I am drawn to aI > paraphrase of Henry Ford's comment about the Model T, "You can have anyuB > operating system from us that you want, as long as it's PH-UX.". > L > Given all the cues HP has been giving, if I were you I'd consider renamingJ > your newsletter to "Shannon Knows AIX", and move on to greener pastures.  D Does anyone know what other vendors are promising?  Does Sun promise@ to support Sparcs or Solaris past 2011?  Does Red Hat promise to support Linux after next year?   -- t John Santosu Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  $ Date: Thu, 7 Nov 2002 16:23:14 -0500  From: John Santos <JOHN@egh.com>- Subject: Re: Was OpenVMS left out on purpose?"4 Message-ID: <1021107162234.400B-100000@Ives.egh.com>  % On Thu, 7 Nov 2002, John Smith wrote:e   > > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message: > news:j7ty9.115642$bt.179925@rwcrnsc52.ops.asp.att.net... > >n2 > > "John Smith" <a@nonymous.com> wrote in messageH > > news:CG_x9.201780$%h2.172042@news02.bloor.is.net.cable.rogers.com... > > >d7 > > > "Shane Smith" <ssmith@icius.com> wrote in messageu0 > > > news:01C284D9.53785580@sulfer.icius.com...N > > > > And /he/ couldn't get it changed? Or didn't want to? [Insert Jay's rap > > > > here.] > > >nL > > > I don't think that web site is owned/operated by HP, so I believe that > allh/ > > > he could do is bitch like the rest of us.d > > >wL > > > Tandem/NSK isn't his department, and I'm sure Pauline Nist has her ownM > > > agenda to follow, and the only thing he/she have in common is that they  > > are N > > > both fated to sink with Itanic unless the banks and stock exchanges tell > > HP > > > to change cpu's for NSK. > >WG > > Well, they have something else in common. They both report to Scott'N > > Stallard. I doubt that bank and exchanges care about the underlying CPU soN > > long as the hardware is NonStop. IPF has lockstepping support and the portG > > is well along. And note that banks.exchanges will be among the late0
 > > adopters.0 > >r > > >aL > > > Unfortunately for VMS there is no group of customers large enough with > > thatL > > > much clout left.   DOD maybe, but even with current generation Alpha'sK > > > J-Stars will not likely need much more single-cpu power well into the  > nextG > > > decade - they can cluster to add horsepower until the APU's can't 	 > providetM > > > any more electricity. At that point, they lower the wind turbine out oft > > thenJ > > > belly for more electricity and air conditioning. They can always put > > smallerrC > > > and lighter crews on-board to save the additional weight. :-)  > >hJ > > Yeah, but the ram-air-turbine will increase the radar cross signature, > last% > > thing you want on a J-Stars bird.l >  > N > It's a converted 707 for chrissake. That's about as stealthy as a cubic mile > of chaff.n  ? Doesn't it also have an extremely powerful active radar system?x   -- 4 John Santos5 Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  # Date: Thu, 07 Nov 2002 22:07:58 GMT-# From: "John Smith" <a@nonymous.com> - Subject: Re: Was OpenVMS left out on purpose?eH Message-ID: <2vBy9.15505$YSz1.3829@news01.bloor.is.net.cable.rogers.com>  - "John Santos" <JOHN@egh.com> wrote in messageg. news:1021107161715.400A-100000@Ives.egh.com...' > On Thu, 7 Nov 2002, John Smith wrote:u >s > >o@ > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message< > > news:m1ty9.115544$bt.179925@rwcrnsc52.ops.asp.att.net... > > > 4 > > > "John Smith" <a@nonymous.com> wrote in messageI > > > news:QlSx9.197587$%h2.53842@news02.bloor.is.net.cable.rogers.com...r > > > >hD > > > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message@ > > > > news:n9Rx9.106352$wG.405791@rwcrnsc51.ops.asp.att.net...H > > > > > VMS was mentioned a total of ONCE at the analyst briefing last week ine
 > > > > Santa  > > > > > Barbara. > > > >I > > > >h > > > > As in,% > > > > a) OpenVMS is owned by HP, orc, > > > > b) OpenVMS will be retired by HP, or> > > > > c) OpenVMS was acquired by HP in the Compaq merger, or  > > > > d) something more vague? > > >oL > > > Umm, a bit more vague. Marcello was supposed to be at the meeting, but hadc > > au > > > conflict.c > > >eJ > > > Seems to me (and I could be wrong) that the IPF port proceeds apace, COEdD > > > ensures 15 years longevity (although support for non Air Force GeneralsL > > > could end in 2011), and that NEW customers will be encouraged to adoptI > > > HP-UX. That's about all I got. No info on early retirement, in factt > > OpenVMS0D > > > Engineering has been largely exempt from workforce reductions. > >i > >n
 > > Terry, > > D > > What gets me about the 2011 date is that the FTC has a ruling or legislationdL > > (at least I think it falls under the FTC) that deals with issues of whatL > > happens when your car or computer manufacturer discontinues a particularK > > model. The vendor is required to stock parts for repair/replacement foru a L > > period of 5 years form the date of the product termination. Arguably one caneJ > > call this 'support'. No new engineering is done, no design defects areJ > > remedied, they just ship you an identical part to the one that failed. > >s< > > Now lets skip over to software and HP's words about VMS. > >34 > > Support until 2011...what does this really mean? > >SL > > Does it mean the all new development on VMS ceases in 2006, and that theI > > 'support' window kicks in after that for the next 5 years. If so, all  onesD > > could expect from HP in 2007-2011 is a new CD of VMS and layered productsE > > frozen in time as of 2006, and a friendly voice at the end of thei	 telephonedF > > line asking you for your credit card number and a ship-to address. > >uJ > > If this is what HP means by support to 2011, everyone using VMS better startwK > > looking for a different hardware and o/s vendor today so you can finishr all  > > your porting by 2006., >dC > I understand this is a "not before" date, not a "not after" date.h >IE > I.E.  They haven't promised they will continue supporting VMS afteruB > this date, which is an entirely different thing than saying theyC > won't support it after this date.  (This applies to both the 200684 > development cut-off and the 2011 support cut-off.) > K > > As to the new customers being shown by HP to the PH-UX door, I am drawn- to aK > > paraphrase of Henry Ford's comment about the Model T, "You can have any D > > operating system from us that you want, as long as it's PH-UX.". > >iE > > Given all the cues HP has been giving, if I were you I'd consider3 renamingL > > your newsletter to "Shannon Knows AIX", and move on to greener pastures. >tF > Does anyone know what other vendors are promising?  Does Sun promiseB > to support Sparcs or Solaris past 2011?  Does Red Hat promise to  > support Linux after next year?    L Strictly speaking, they don't. The assumption (or leap of faith) by everyoneH is that as long as those companies are in business they will continue toK develop, support, and do everything in their power to expand the market forn. those operating systems, without a time limit.  G With HP there is no real clarity despite all the dates they have throwne around.i  J Alpha will die in 2005/06 and they will 'support' (read supply spare partsJ for) Alpha until 2011, unless customers keep clamoring for Alpha, in whichI case HP will rummage through drawers and old boxes in the warehouse for au@ few old EV69 and EV7 chips in order to build a few more, by then trailing-edge, Alpha's.P  K As to VMS, It'll be 'supported' on Alpha as an EOL product once Alpha's areuJ declared EOL in 20005/06. As to VMS on Itanic, Itanic will probably be EOL by 2006 as well.  I If there isn't sufficient uptake of IA-64 by any other company other than L HP, Intel will tell HP that they're getting out of the IA-64 business unlessI HP funds it entirely. And when that happens you'll hear the great suckingTK sound of Microsoft pulling their developers off Windows Whatever for IA-64.V  C Intel as a separate entity from HP has only about 1/2 the financial F investment in the IA-64 project (I stand to be corrected), and as suchJ doesn't have quite the 'emotional' need to see the project through if theyG figure their bread is best buttered elsewhere. This would bring HP full K circle back to a position where they would have been better off financially C either moving ahead with PA-Risc or Alpha full steam ahead. Neithero= architecture has, or would cost as much as IA-64 has or will.   K HP would have been better off performance-wise moving ahead with Alpha thanc IA-64 or PA-Risc.l  H And HP would have been better off in the market since by bringing higherJ performance products to market sooner as they would not have had to changeJ horses from Alpha to IA-64 on 3 operating systems (VMS, Tru64/PH-UX, NSK),F which would have enhance profitability rather than being a drag on it.  J To believe otherwise is foolish. Alpha had 3 proprietary operating systemsC on it from Digital/Compaq  (VMS, Tru64, NSK underway) a couple frompJ Microsoft ( Windows NT and 64-bit Windows beta), and a few 'open' ones tooI (BSD, Linux). Arguably, Alpha was more 'industry standard' or 'open' than- anything from Intel.   ------------------------------  # Date: Thu, 07 Nov 2002 22:10:31 GMTB# From: "John Smith" <a@nonymous.com>j- Subject: Re: Was OpenVMS left out on purpose?nD Message-ID: <rxBy9.9823$oRV.55@news04.bloor.is.net.cable.rogers.com>  - "John Santos" <JOHN@egh.com> wrote in messageh. news:1021107162234.400B-100000@Ives.egh.com...' > On Thu, 7 Nov 2002, John Smith wrote:h >, > >m@ > > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message< > > news:j7ty9.115642$bt.179925@rwcrnsc52.ops.asp.att.net... > > > 4 > > > "John Smith" <a@nonymous.com> wrote in messageJ > > > news:CG_x9.201780$%h2.172042@news02.bloor.is.net.cable.rogers.com... > > > >b9 > > > > "Shane Smith" <ssmith@icius.com> wrote in messages2 > > > > news:01C284D9.53785580@sulfer.icius.com...L > > > > > And /he/ couldn't get it changed? Or didn't want to? [Insert Jay's rapt > > > > > here.] > > > > I > > > > I don't think that web site is owned/operated by HP, so I believev that > > all+1 > > > > he could do is bitch like the rest of us.e > > > >jJ > > > > Tandem/NSK isn't his department, and I'm sure Pauline Nist has her own J > > > > agenda to follow, and the only thing he/she have in common is that they	 > > > areyK > > > > both fated to sink with Itanic unless the banks and stock exchangesn tell > > > HP  > > > > to change cpu's for NSK. > > >oI > > > Well, they have something else in common. They both report to ScottuI > > > Stallard. I doubt that bank and exchanges care about the underlyingi CPU soK > > > long as the hardware is NonStop. IPF has lockstepping support and thet portI > > > is well along. And note that banks.exchanges will be among the late  > > > adopters.i > > >e > > > >AI > > > > Unfortunately for VMS there is no group of customers large enougha with
 > > > thatF > > > > much clout left.   DOD maybe, but even with current generation Alpha'slI > > > > J-Stars will not likely need much more single-cpu power well intop thee > > nextI > > > > decade - they can cluster to add horsepower until the APU's can't  > > provideoL > > > > any more electricity. At that point, they lower the wind turbine out of	 > > > theaL > > > > belly for more electricity and air conditioning. They can always put
 > > > smallereE > > > > and lighter crews on-board to save the additional weight. :-)n > > >eL > > > Yeah, but the ram-air-turbine will increase the radar cross signature, > > last' > > > thing you want on a J-Stars bird.v > >e > >hK > > It's a converted 707 for chrissake. That's about as stealthy as a cubic  mile
 > > of chaff.l >sA > Doesn't it also have an extremely powerful active radar system?e    J That too makes it about as stealthy as an AWACS - sort of like a supernovaL in a pitch black sky. Rather hard to miss even if you aren't looking for it.   ------------------------------  % Date: Thu, 07 Nov 2002 17:58:45 -0500o0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>- Subject: Re: Was OpenVMS left out on purpose?M/ Message-ID: <3DCAF01C.CC0DB737@vl.videotron.ca>f   John Santos wrote:E > I.E.  They haven't promised they will continue supporting VMS afteroB > this date, which is an entirely different thing than saying theyC > won't support it after this date.  (This applies to both the 2006i4 > development cut-off and the 2011 support cut-off.)   Correct.  K But in a context where one does not trust the vendor and does not know whatsN the true intentions are, such statements are easily interpreted that Compaq/HP are set to kill VMS in 2006.  B A vendor who has full confidence in a product and fully intends toH grow/expand/improve the product doesn't need to provide "minimum support dates".   G In fact, VMS is veing treated the same way as Tru64 and MPE with statedmK "minimum support" dates, except thatthey haven't announced that those dates L are "final".  Tandem doesn't have minimum support dates, neither does HP-UX.K And of course, nobody questions Microsoft's desire to grow windows forever.i  H Instead of fixing the "Digital is killing VMS" image once and for all byL taking actions that clearly show they intend to grow and expland the OS, andK by stating how important its profits are, and how strategic the product is,:K and how it beats its competitors, HP/Compaq/Digital merely acknowledged and>H re-enforced the image of "Digital is killing VMS" because they just keepG saying essentially "don't worry, we won't kill it in the next x years".i  C Actions speak louder than words. Advertise VMS, and make it competetI agressively, and open it to all markets, and you won't need to constantlyeA alleviate fears that your real intentions are to kill off the OS.s   ------------------------------  $ Date: Fri, 8 Nov 2002 13:05:00 +0100" From: "Hans Vlems" <hvlems@iae.nl>$ Subject: Re: XP1000 hardware problem5 Message-ID: <aqg9a2$9t24g$1@ID-143435.news.dfncis.de>e  I <sms@antinode.org> schreef in bericht news:02110614040794@antinode.org...wI >    After some pretty gentle transportation, I formerly happy XP1000 hasoJ > become uncommunicative (no video output, no response at COMM 1).  Of theF > eight LEDs on the main board (D14-D21), D20 and D21 stay on.  I'd beF > grateful for a diagnosis or any suggestions for further exploration. > J > ------------------------------------------------------------------------ > E >    Steven M. Schweda               (+1) 651-699-9818  (voice, home) 5 >    382 South Warwick Street        sms@antinode.orga >    Saint Paul  MN  55105-25473  G After transportation it might be a good idea to reseat all the boards. >   Hans   ------------------------------  + Date: Fri, 08 Nov 2002 10:06:05 -0600 (CST)T From: sms@antinode.org$ Subject: Re: XP1000 hardware problem) Message-ID: <02110810060504@antinode.org>   " From: "Hans Vlems" <hvlems@iae.nl>I > After transportation it might be a good idea to reseat all the boards.    B    I tried that, power, CPU, memory, and graphics.  I also tried aH different graphics card.  I even washed some apparent flux residue off aF region of the main board near some rework at the Chinese (protective?); gizmo near the COMM 1 port.  (_I_ do nicer work than that.)o  H    I hoped to get some info on the meaning of the LEDs, in the hope that+ I could learn how far it got on its way up.i  H ------------------------------------------------------------------------  C    Steven M. Schweda               (+1) 651-699-9818  (voice, home)E3    382 South Warwick Street        sms@antinode.orgc    Saint Paul  MN  55105-2547n   ------------------------------   End of INFO-VAX 2002.618 ************************ted by makers of7 home routers, allows multiple simultaneous connections.d  L Your router will make it look to the internet like there is only a single IP* address and a single machine on your line.  F Each of your machines on your lan will have a private address, usuallyG non-routable address (for instance 10.*.*.*). Your router will have two J interfaces. On the LAN side, it will have an address in the same subnet ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª     ª    ª    ª    ª    ª    ª    ª    ª    ª    	ª    
ª    ª    ª    
ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª     ª    !ª    "ª    #ª    $ª    %ª    &ª    'ª    (ª    )ª    *ª    +ª    ,ª    -ª    .ª    /ª    0ª    1ª    2ª    3ª    4ª    5ª    6ª    7ª    8ª    9ª    :ª    ;ª    <ª    =ª    >ª    ?ª    @ª    Aª    Bª    Cª    Dª    Eª    Fª    Gª    Hª    Iª    Jª    Kª    Lª    Mª    Nª    Oª    Pª    Qª    Rª    Sª    Tª    Uª    Vª    Wª    Xª    Yª    Zª    [ª    \ª    ]ª    ^ª    _ª    `ª    aª    bª    cª    dª    eª    fª    gª    hª    iª    jª    kª    lª    mª    nª    oª    pª    qª    rª    sª    tª    uª    vª    wª    xª    yª    zª    {ª    |ª    }ª    ~ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    ª    