1 INFO-VAX	Sun, 03 Apr 2005	Volume 2005 : Issue 186       Contents: RE: cloning a system disk  Re: cloning a system disk  RE: cloning a system disk  Re: cloning a system disk  RE: cloning a system disk  RE: cloning a system disk   RE: dvd writing software on OVMS  Re: dvd writing software on OVMS  Re: dvd writing software on OVMS* Re: New HP CEO Announced - Mark Hurd (NCR)* Re: New HP CEO Announced - Mark Hurd (NCR)* Re: New HP CEO Announced - Mark Hurd (NCR)$ RE: poor disk I/O performace on ds25$ RE: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25* Re: problem with router---NAT and caching? Re: Script Kiddie tarpit wanted  Re: Script Kiddie tarpit wanted  Re: Script Kiddie tarpit wanted  Re: Script Kiddie tarpit wanted + Re: There is certainly opportunity in chaos + Re: There is certainly opportunity in chaos   F ----------------------------------------------------------------------  $ Date: Sun, 3 Apr 2005 09:48:36 -0400' From: "Main, Kerry" <kerry.main@hp.com> " Subject: RE: cloning a system diskR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB59566F@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Phillip Helbig---remove CLOTHES to reply=20 - > [mailto:helbig@astro.multiCLOTHESvax.de]=20  > Sent: March 31, 2005 10:38 AM  > To: Info-VAX@Mvb.Saic.Com   > Subject: cloning a system disk >=20? > I've mentioned this a few times here in the past, but have=20  > never gotten=20 # > a completely satisfactory answer.  >=20J > Does anyone have a fool-proof plan and/or DCL procedure for cloning a=20C > system disk?  What I want to do is use one member of a cleanly=20 J > dismounted system-disk shadow set as the system disk on another node,=20J > then, after the nodes are up and running, add a second member to each=20 > single-disk shadow set.  >=20: > I have moved SYSUAF.DAT and all other files mentioned inH > SYS$MANAGER:SYLOGICALS.TEMPLATE off the system disk, so that is not anE > issue (including things like queue names etc---of course, after the G > system is up and running, I might want to set up additional queues on 	 > it).=20  >=20G > What do I have to change before booting the new system IN THE SAME=20 < > CLUSTER?  (Obviously, I will try a standalone boot first.) >=20 > Here's what I've thought of: >=20G >    o  names of disk(s) for secondary page and/or swap files (MIGHT be  >       different)=20  >=20 >    o  SCSNODE  >=20 >    o  SCSSYSTEMID  >=20 >    o  ALLOCLASS  >=20 >    o  TAPE_ALLOCLASS >=20 >    o  IP address >=20 >    o  IP hostname  >=20E >    o  complete TCPIP configuration if the name of the adapter is=20 % >    o  different (not the case here)  >=20@ > The primary reason is to avoid repeating the installation. =20 > VMS itself I=20 G > could live with, but I don't want to re-install all layered products.  >=20@ > Since this is a fresh system, I don't care about historical=20 > baggage likeC > ACCOUNTNG.DAT, TCPIP log files etc.  However, in the future, when @ > repeating the process after upgrading VMS and/or installing=20
 > new layered H > products, it would be nice to save stuff like ACCOUNTNG.DAT, TCPIP logF > files etc before the cloning then put them back in place of a cloned8 > one.  Does anyone have a complete list of such things? >=20 >=20  	 Phillip -   C This might not be applicable, but something to keep in mind is that E batch queue jobs retain the physical device where they were submitted B to. Hence, if there are lots of batch jobs to be concerned about -
 beware.=20  G I had a Cust who was looking at a DR testing scenario and that was when D they discovered that on the DR system with different physical deviceF names, the hundreds of batch jobs they had would all bomb. They had to5 write a small script to re-submit all the batch jobs.   B As an example, do a $ show entry xx/full - see the physical device
 reference?   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Sun, 03 Apr 2005 16:32:27 +0200 & From: Paul Sture <paul.sture@decus.ch>" Subject: Re: cloning a system disk, Message-ID: <3bad3rF6hpq59U1@individual.net>   Main, Kerry wrote:   >>-----Original Message-----1 >>From: Phillip Helbig---remove CLOTHES to reply  + >>[mailto:helbig@astro.multiCLOTHESvax.de]   >>Sent: March 31, 2005 10:38 AM  >>To: Info-VAX@Mvb.Saic.Com   >>Subject: cloning a system disk >>= >>I've mentioned this a few times here in the past, but have   >>never gotten  # >>a completely satisfactory answer.  >>H >>Does anyone have a fool-proof plan and/or DCL procedure for cloning a A >>system disk?  What I want to do is use one member of a cleanly  H >>dismounted system-disk shadow set as the system disk on another node, H >>then, after the nodes are up and running, add a second member to each  >>single-disk shadow set.  >>: >>I have moved SYSUAF.DAT and all other files mentioned inH >>SYS$MANAGER:SYLOGICALS.TEMPLATE off the system disk, so that is not anE >>issue (including things like queue names etc---of course, after the G >>system is up and running, I might want to set up additional queues on  >>it).   >>E >>What do I have to change before booting the new system IN THE SAME  < >>CLUSTER?  (Obviously, I will try a standalone boot first.) >> >>Here's what I've thought of: >>G >>   o  names of disk(s) for secondary page and/or swap files (MIGHT be  >>      different)   >> >>   o  SCSNODE  >> >>   o  SCSSYSTEMID  >> >>   o  ALLOCLASS  >> >>   o  TAPE_ALLOCLASS >> >>   o  IP address >> >>   o  IP hostname  >>C >>   o  complete TCPIP configuration if the name of the adapter is  % >>   o  different (not the case here)  >>> >>The primary reason is to avoid repeating the installation.   >>VMS itself I  G >>could live with, but I don't want to re-install all layered products.  >>> >>Since this is a fresh system, I don't care about historical  >>baggage likeC >>ACCOUNTNG.DAT, TCPIP log files etc.  However, in the future, when > >>repeating the process after upgrading VMS and/or installing 
 >>new layered H >>products, it would be nice to save stuff like ACCOUNTNG.DAT, TCPIP logF >>files etc before the cloning then put them back in place of a cloned8 >>one.  Does anyone have a complete list of such things? >> >> >  >  > Phillip -  > E > This might not be applicable, but something to keep in mind is that G > batch queue jobs retain the physical device where they were submitted D > to. Hence, if there are lots of batch jobs to be concerned about -
 > beware.  > I > I had a Cust who was looking at a DR testing scenario and that was when F > they discovered that on the DR system with different physical deviceH > names, the hundreds of batch jobs they had would all bomb. They had to7 > write a small script to re-submit all the batch jobs.  > D > As an example, do a $ show entry xx/full - see the physical device > reference? > I But it is "worse" than that, depending on your point of view. The FID is  H also stored, so editing the batch file (and even making sure it has the ; same version number as the original) won't execute the job.    ------------------------------  $ Date: Sun, 3 Apr 2005 11:45:35 -0400' From: "Main, Kerry" <kerry.main@hp.com> " Subject: RE: cloning a system diskR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB595678@tayexc19.americas.cpqcorp.net>   > -----Original Message-----2 > From: Paul Sture [mailto:paul.sture@decus.ch]=20 > Sent: April 3, 2005 10:32 AM > To: Info-VAX@Mvb.Saic.Com $ > Subject: Re: cloning a system disk >=20 > Main, Kerry wrote: >=20 > >>-----Original Message-----5 > >>From: Phillip Helbig---remove CLOTHES to reply=20 / > >>[mailto:helbig@astro.multiCLOTHESvax.de]=20 ! > >>Sent: March 31, 2005 10:38 AM  > >>To: Info-VAX@Mvb.Saic.Com " > >>Subject: cloning a system disk > >>A > >>I've mentioned this a few times here in the past, but have=20  > >>never gotten=20 % > >>a completely satisfactory answer.  > >>B > >>Does anyone have a fool-proof plan and/or DCL procedure for=20 > cloning a=20E > >>system disk?  What I want to do is use one member of a cleanly=20 > > >>dismounted system-disk shadow set as the system disk on=20 > another node,=20= > >>then, after the nodes are up and running, add a second=20  > member to each=20  > >>single-disk shadow set.  > >>< > >>I have moved SYSUAF.DAT and all other files mentioned in> > >>SYS$MANAGER:SYLOGICALS.TEMPLATE off the system disk, so=20 > that is not anG > >>issue (including things like queue names etc---of course, after the B > >>system is up and running, I might want to set up additional=20 > queues on  > >>it).=20  > >>I > >>What do I have to change before booting the new system IN THE SAME=20 > > >>CLUSTER?  (Obviously, I will try a standalone boot first.) > >>  > >>Here's what I've thought of: > >>B > >>   o  names of disk(s) for secondary page and/or swap files=20 > (MIGHT be  > >>      different)=20  > >> > >>   o  SCSNODE  > >> > >>   o  SCSSYSTEMID  > >> > >>   o  ALLOCLASS  > >> > >>   o  TAPE_ALLOCLASS > >> > >>   o  IP address > >> > >>   o  IP hostname  > >>G > >>   o  complete TCPIP configuration if the name of the adapter is=20 ' > >>   o  different (not the case here)  > >>B > >>The primary reason is to avoid repeating the installation. =20 > >>VMS itself I=20 B > >>could live with, but I don't want to re-install all layered=20 > products.  > >>B > >>Since this is a fresh system, I don't care about historical=20 > >>baggage likeE > >>ACCOUNTNG.DAT, TCPIP log files etc.  However, in the future, when B > >>repeating the process after upgrading VMS and/or installing=20 > >>new layered 4 > >>products, it would be nice to save stuff like=20 > ACCOUNTNG.DAT, TCPIP logH > >>files etc before the cloning then put them back in place of a cloned: > >>one.  Does anyone have a complete list of such things? > >> > >> > >=20 > >=20
 > > Phillip -  > >=20G > > This might not be applicable, but something to keep in mind is that B > > batch queue jobs retain the physical device where they were=20 > submitted F > > to. Hence, if there are lots of batch jobs to be concerned about - > > beware.=20 > >=20@ > > I had a Cust who was looking at a DR testing scenario and=20 > that was when H > > they discovered that on the DR system with different physical deviceA > > names, the hundreds of batch jobs they had would all bomb.=20 
 > They had to 9 > > write a small script to re-submit all the batch jobs.  > >=20F > > As an example, do a $ show entry xx/full - see the physical device > > reference? > >=20B > But it is "worse" than that, depending on your point of view.=20 > The FID is=20 A > also stored, so editing the batch file (and even making sure=20  > it has the=20 = > same version number as the original) won't execute the job.  >=20  C Course, while jobs executing during a significant event are another E topic, resubmitting the jobs-that-have-not-run-yet would work as well : since they would take on the new physical device names.=20  A However, you need to know timing, parameters ect so this Cust was F looking at doing a daily midnight job which was $ show queue /full/allH /out=3Dbatchfile_date_time.txt (and ensuring that file made it to the DR site as part of the batch job).   B Not totally without issues, but they were primarily concerned withE daily's that ran everyday, but were subject to occasional updates and 
 mod's etc.   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Sun, 03 Apr 2005 12:32:22 -0400 ' From: Dave Froble <davef@tsoft-inc.com> " Subject: Re: cloning a system disk/ Message-ID: <115061utm8ej74@corp.supernews.com>    Main, Kerry wrote:  I > I had a Cust who was looking at a DR testing scenario and that was when F > they discovered that on the DR system with different physical deviceH > names, the hundreds of batch jobs they had would all bomb. They had to7 > write a small script to re-submit all the batch jobs.   B Is there anybody that doesn't have such command files?  Same with @ command file(s) to initialize all the queues at system start-up.   ------------------------------  $ Date: Sun, 3 Apr 2005 13:25:59 -0400' From: "Main, Kerry" <kerry.main@hp.com> " Subject: RE: cloning a system diskR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB59567B@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Dave Froble [mailto:davef@tsoft-inc.com]=20  > Sent: April 3, 2005 12:32 PM > To: Info-VAX@Mvb.Saic.Com $ > Subject: Re: cloning a system disk >=20 > Main, Kerry wrote: >=20@ > > I had a Cust who was looking at a DR testing scenario and=20 > that was when H > > they discovered that on the DR system with different physical deviceA > > names, the hundreds of batch jobs they had would all bomb.=20 
 > They had to 9 > > write a small script to re-submit all the batch jobs.  >=20F > Is there anybody that doesn't have such command files?  Same with=20B > command file(s) to initialize all the queues at system start-up. >=20   Dave,   @ The batch job stuff remains across reboots. You do not re-submit daily's/weekly's etc on reboot.   H Most large shops typically have all sorts of application and/or db maint@ and/or reporting batch jobs that are submitted during the day byH different business groups at different times to different queues. Hence,F re-creating batch jobs to run in the right sequence and with the right: parameters on a different system can be a major challenge.   Regards    / Kerry    ------------------------------  * Date: Sun, 3 Apr 2005 17:30:01 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)" Subject: RE: cloning a system disk$ Message-ID: <d2p96o$rnu$2@online.de>  
 In articleG <FD827B33AB0D9C4E92EACEEFEE2BA2FB59566F@tayexc19.americas.cpqcorp.net>, * "Main, Kerry" <kerry.main@hp.com> writes:   E > This might not be applicable, but something to keep in mind is that G > batch queue jobs retain the physical device where they were submitted D > to. Hence, if there are lots of batch jobs to be concerned about -	 > beware.   G Yes, something to keep in mind.  One trick is to have a batch job call  D itself with the parameters it was submitted with, but using logical H names instead of physical ones.  This will avoid this problem, and also I allow the newest version of the procedure to be used, which might not be   the one which was submitted.   ------------------------------  $ Date: Sun, 3 Apr 2005 09:58:01 -0400' From: "Main, Kerry" <kerry.main@hp.com> ) Subject: RE: dvd writing software on OVMS R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB595671@tayexc19.americas.cpqcorp.net>   > -----Original Message-----5 > From: Nitendra [mailto:ns_panwar@rediffmail.com]=20  > Sent: April 1, 2005 8:02 AM  > To: Info-VAX@Mvb.Saic.Com + > Subject: Re: dvd writing software on OVMS  >=20 > try cdrecord >=20 >=20   Also, check out:# http://home.tiscali.de/dvd4openvms/ 4 http://home.tiscali.de/dvd4openvms/dvdwrite_info.txt   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Sun, 03 Apr 2005 17:00:42 +0200 & From: Paul Sture <paul.sture@decus.ch>) Subject: Re: dvd writing software on OVMS , Message-ID: <3baeoqF6bg4psU1@individual.net>   Main, Kerry wrote:   >> >>try cdrecord >> >> >  >  > Also, check out:% > http://home.tiscali.de/dvd4openvms/ 6 > http://home.tiscali.de/dvd4openvms/dvdwrite_info.txt > G Aaaaaaarrrrrggghhhhh. No offence to Eberhard Heuser-Hofmann, but every  $ time I see the name Tiscali I choke.  F To borrow a bit of VAXMAN's language, Tiscali are fucking cheats, and H then sue you for the money you didn't pay because they didn't provide a  service.  % Please, folks, don't go near Tiscali.    ------------------------------  % Date: Sun, 03 Apr 2005 09:31:18 -0700 # From: "Tom Linden" <tom@kednos.com> ) Subject: Re: dvd writing software on OVMS ( Message-ID: <opsonz6gjtzgicya@hyrrokkin>  I On Sun, 3 Apr 2005 09:58:01 -0400, Main, Kerry <kerry.main@hp.com> wrote:    >> try cdrecord  >> >> > Also, check out:% > http://home.tiscali.de/dvd4openvms/ 6 > http://home.tiscali.de/dvd4openvms/dvdwrite_info.txt  G Is there a list of supported devices online?  I need both SCSI (PWS600)  and IDE (XP1000)   ------------------------------  $ Date: Sun, 3 Apr 2005 11:02:33 -0400# From: "John Smith" <a@nonymous.com> 3 Subject: Re: New HP CEO Announced - Mark Hurd (NCR) , Message-ID: <S7ydnQYfz8AMmM3fRVn-tA@igs.net>  
 AEF wrote: > F > Explain that anyone can make and sell Unix, but only HP has VMS, andE > how VMS offers unique and important advantages that no one else can D > offer. And that by treating the customer right, there won't be anyF > worry about being locked in to one vendor. OK, one has to be careful > on this point, I suppose.     L Hurd has the experience of being the vendor of Teradata, which in itelf is aK lock-in to an extent lesser than that of the o/s but still a lock-in if all I the product features are used. It wouldn't be much different if a company F standarized on Oracle or Tibco products and used them to their fullestL extent - migration would be a painful exercise in those situations too. ThisL is the nirvana that vendors aspire to....by having products that do what the= customer really wants, not what the trade rags say they want.      --F OpenVMS - The never advertised operating system with the dwindling ISV base.    ------------------------------  % Date: Sun, 03 Apr 2005 08:04:07 -0700 # From: "Tom Linden" <tom@kednos.com> 3 Subject: Re: New HP CEO Announced - Mark Hurd (NCR) ( Message-ID: <opsonv45y0zgicya@hyrrokkin>  G On Sun, 03 Apr 2005 00:33:55 -0500, Dave Froble <davef@tsoft-inc.com>    wrote:   > AEF wrote: > J >> Well, if VMS is really bringing in good profits like Tom Linden claims,D >> then imagine the profits if the base were expanded! Not much more> >> costs, but a lot more revenue, even if you lower the price! >> > L > Tom's numbers are similar to those stated in 1999, but I think I've read  L > that the numbers lately are about half of that.  Hey, what do I know, he   > may know something.   G My numbers come from inside and are reliable.  There are two things any D good executive will look at, (1) Top line/bottom line and (2) growthF potential.  Concerning the latter, it is my view that there is great   opportunity L in the market at the present in the areas with which you are all familiar,   DR, H reliability etc.  The weak side of course is ISV and diminishing third   party L applications.  To pursue an aggressive marketing plan for growth therefore   mustE factor in the cost of reversing this diminution.  I don't know if a    marketing/sales K organization exists within HP for VMS, but it either needs to be created or C reconstituted.  Secondly,  having established a marketing program    emphasizing the J key strengths (e.g. a DR group) a sales program can be put together, and   the J message has to get out that VMS is very much alive and well.  As you all   know, F the general perception is that VMS is dead, mid-level managers won't   promote H because it is regarded as a bad career move.  Sarbanes-Oxley creates a   great A opportunity for VMS, have a sales group focused strictly on this.   K And the name has to change.  Can you imagine IBM having continued with MVS? < How about SCE-2, Secure Computin Environment, 2nd generation       > K > How about when a customer complains about all the costs of dealing with   L > virus', being able to look him directly and say that you can provide him  G > with systems that will not be affected by any virus, trojan, or worm?  > L > A good sales point would be that the savings in not coping with internet  / > problems would more than pay for the systems.  > H > While our old friend, must have been downsized Andy boy, would go on  I > about the lack of applications, VMS as a web server should be able to   6 > communicate with systems running those applications. > ( > Don't know, not running any web stuff. >  > Dave   ------------------------------  % Date: Sun, 03 Apr 2005 12:26:07 -0400 ( From: Bill Todd <billtodd@metrocast.net>3 Subject: Re: New HP CEO Announced - Mark Hurd (NCR) = Message-ID: <YZidnUId6dS7hM3fRVn-rQ@metrocastcablevision.com>    Tom Linden wrote: I > On Sun, 03 Apr 2005 00:33:55 -0500, Dave Froble <davef@tsoft-inc.com>    > wrote:   ...   G >> Tom's numbers are similar to those stated in 1999, but I think I've  I >> read  that the numbers lately are about half of that.  Hey, what do I    >> know, he  may know something. >  > / > My numbers come from inside and are reliable.   I A great deal of the absolute garbage immediately following the Alphacide  I 'came from the inside and was reliable', Tom.  Some of it even came from  E Mark Gorham, IIRC (though I have no reason to believe that he *knew*  G that what he was spouting was utterly false:  he probably just trusted  > others whom he shouldn't have, just as you may well be doing).  D So pardon some of us for not taking the numbers you've presented on G faith.  If you could bolster them with something more substantive than  % your own trust in them it would help.    - bill   ------------------------------  $ Date: Sun, 3 Apr 2005 09:23:27 -0400' From: "Main, Kerry" <kerry.main@hp.com> - Subject: RE: poor disk I/O performace on ds25 R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB59566E@tayexc19.americas.cpqcorp.net>   > -----Original Message-----5 > From: Nitendra [mailto:ns_panwar@rediffmail.com]=20  > Sent: April 2, 2005 1:35 PM  > To: Info-VAX@Mvb.Saic.Com / > Subject: Re: poor disk I/O performace on ds25  >=20 > Dear Denny >=20E > there are three Ultra 160 scsi controller, two are on board and one  > addonPCI card. >=20H > both harddisks are connected on one of the integrated scsi controller. >=20A > on other internal scsi controller only one tape drive dat72. is D > connected and third one is not having any deivces connected to it. >=20A > there is no scsi error, all harddisk are conneted to hotswap=20  > case ba610  > shipped along with the system. >=20F > as far as hardisk model and make is concern i will have to go to the- > server location and will confirm you later.  >=20	 > Regards  >=20 >=20    D One common mistake is to forget that the VMS disk security parameterH called "highwater marking" is on by default. This means that every writeF IO will have the blocks zeroed before the blocks are allocated to thatD process. Hence, this can add some additional overhead to write IO's.  F It might not make any difference, but just in case, you can check this8 on all drives on your system with the following command:  6 $ pipe sho dev/full |search sys$pipe high/wind=3D(0,5)  @ If a drive does have it enabled, you can disable it by using the< following example (substitute your drive name for $1$dka0:):  ! $ set volume $1$dka0:/nohighwater   A You do not have to dismount the drive to do this - it can be done  online.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------   Date: 3 Apr 2005 10:25:23 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) - Subject: RE: poor disk I/O performace on ds25 3 Message-ID: <Cs2pcQljdA$B@eisner.encompasserve.org>   | In article <FD827B33AB0D9C4E92EACEEFEE2BA2FB59566E@tayexc19.americas.cpqcorp.net>, "Main, Kerry" <kerry.main@hp.com> writes:  F > One common mistake is to forget that the VMS disk security parameterJ > called "highwater marking" is on by default. This means that every writeH > IO will have the blocks zeroed before the blocks are allocated to thatF > process. Hence, this can add some additional overhead to write IO's.  5 I do not believe that is how highwater marking works.   D I believe it ensures data is zeroes only when the EOF part is moved,? so for a typical sequential write the RMS buffer will be zeroed 8 beyond the highest used bytes, which is no big overhead.  B If you have an application that explicit asks for a big allocation@ increase and then writes data near the end of that space withoutC having written data in between then the effect Kerry describes will 4 take effect.  But those applications are quite rare.  G As I recall there was a time when highwater marking was less efficient, B but the design and current implementation, as I understand it, are pretty efficient.    ------------------------------  % Date: Sun, 03 Apr 2005 09:08:38 -0700 ( From: Jeff Cameron <roktsci@comcast.net>- Subject: Re: poor disk I/O performace on ds25 / Message-ID: <BE756316.ADB0%roktsci@comcast.net>   K On 4/3/05 8:25 AM, in article Cs2pcQljdA$B@eisner.encompasserve.org, "Larry ) Kilgallen" <Kilgallen@SpamCop.net> wrote:   
 > In article  P > <FD827B33AB0D9C4E92EACEEFEE2BA2FB59566E@tayexc19.americas.cpqcorp.net>, "Main,$ > Kerry" <kerry.main@hp.com> writes: > G >> One common mistake is to forget that the VMS disk security parameter K >> called "highwater marking" is on by default. This means that every write I >> IO will have the blocks zeroed before the blocks are allocated to that G >> process. Hence, this can add some additional overhead to write IO's.  > 7 > I do not believe that is how highwater marking works.  > F > I believe it ensures data is zeroes only when the EOF part is moved,A > so for a typical sequential write the RMS buffer will be zeroed : > beyond the highest used bytes, which is no big overhead. > D > If you have an application that explicit asks for a big allocationB > increase and then writes data near the end of that space withoutE > having written data in between then the effect Kerry describes will 6 > take effect.  But those applications are quite rare. > I > As I recall there was a time when highwater marking was less efficient, D > but the design and current implementation, as I understand it, are > pretty efficient.   B I was always under the impression that the blocks were zeroed when7 allocated, to prevent allocating then reading old data.    ------------------------------   Date: 3 Apr 2005 10:04:39 -0700 + From: "Nitendra" <ns_panwar@rediffmail.com> - Subject: Re: poor disk I/O performace on ds25 C Message-ID: <1112547879.223220.132730@o13g2000cwo.googlegroups.com>   
 Dear Kerry   i will try these setting  C but the thing is all these machines are newlly shipped, they dosent % have any application running on them.   D its only copy command i am using it copy 500MB single file is copied8 from one disk to another and i am getting poor response.  1 this is case with all the 5 DS25 we have procured    Regards    ------------------------------  * Date: Sun, 3 Apr 2005 17:27:26 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)3 Subject: Re: problem with router---NAT and caching? $ Message-ID: <d2p91u$rnu$1@online.de>  D In article <d2ok2i$ssb$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    G > In article <1112477771.30a44d6f285e33f4fba0a98847877464@teranews>, JF 0 > Mezei <jfmezei.spamnot@vaxination.ca> writes:  > F > > How do you change the IP address of your VMS box ?  Do you use theG > > TCPIP$CONFIG menus, or the TCPIP> Utility ?  Do you destroy the old $ > > interface with SET NOINTERFACE ? > > I > > Have you used ifcongig -a which displays kernel (live) data to ensure H > > that the TCPIP commands actually took effect on the running system ? > < > OK, it looks like there is nothing which I am doing wrong. > 5 > Here's a step-by-step description of the last test.   8 I switched off the router for a few hours.  Same result.  I Obviously, it a) somehow remembers what MAC address has been used in the  H past and b) for some reason doesn't forward port 6000 to it, unless the : original address is used.  Zyxel experts, please stand up!   ------------------------------  % Date: Sun, 03 Apr 2005 11:46:04 +0200 + From: Karsten Nyblad <nospam@nospam.nospam> ( Subject: Re: Script Kiddie tarpit wanted= Message-ID: <424fbb88$0$67260$157c6196@dreader2.cybercity.dk>    Alan Frisbie wrote:  > Roy Omond wrote: >  >> Alan Frisbie wrote: >  > @ >>> I am thinking of creating captive accounts that these script= >>> kiddies would be likely to try (root/<null>, guest/guest, ? >>> god/god, etc.).   The captive procedure should appear to be ? >>> some variant of Unix, appear to sort-of work, while wasting < >>> as much of their time as possible.   Meanwhile it should@ >>> log everything to a file for my later amusement.   (Yes, I'm >>> that sick!)  >  > < >> Or how about a captive account which runs ELIZA (from the8 >> Turing Test in the AI community) with variable-length; >> pauses up to 2 minutes between responses ?  There surely - >> must have been a version of ELIZA for VMS.  >  > 9 > I am thinking about a program that would accept typical 7 > Unix/Linux commands and give responses that will look > > like it worked, at least to a script kiddie.  An alternative> > would be something completely alien to them, like VMS.   :-) > ; > I'm sure that the machines doing the probing are zombies, = > infected with some WinDoze virus, so I don't want to attack ; > them.   They probably just pass their results back to the < > "mother ship".   When the kid returns home from school, he? > will find my IP address in the report, along with a "working" > > login name and password.   That's when the fun starts.   :-) > : > I don't want to spend a lot of time on this, just have aA > little fun watching the kid thrash around inside the "sandbox".  > < > Of course, there is always the option of simply giving him; > a regular non-priv VMS user account, with a login message < > about using the HELP command and links to the online docs.; > We might wind up with another VMS convert that way.   :-)  > ! > Which would be more evil?   :-)  >  > Alan  I In the old days AT&T Bell labs were using a Unix box as firewall.  Their  H password file was shadowed so that nobody could se it from the outside. F   However, they had a fake password file.  The passwords in that file I were:  Why, are, you, wasting, your, and time.  When somebody downloaded  C that file, Bell labs made sure to inform the system manager on the   machine used for downloading.    ------------------------------   Date: 03 Apr 2005 11:12:56 GMT/ From: Thierry Dussuet <thierry@dussuet.lugs.ch> ( Subject: Re: Script Kiddie tarpit wanted/ Message-ID: <slrnd4vjtp.sf.thierry@MARS.Family>   6 On 2005-04-03, Paul Sture <paul.sture@decus.ch> wrote: > Thierry Dussuet wrote: > @ > from a lugs.ch address! (That's Linux User Group Switzerland). >  > Ooooh. :-)   :-)   C This is just a UUCP forwarder to be sure to get all the mails^Wspam    Thierry    ------------------------------  % Date: Sun, 03 Apr 2005 07:36:35 -0700 # From: "Tom Linden" <tom@kednos.com> ( Subject: Re: Script Kiddie tarpit wanted( Message-ID: <opsonuu9n6zgicya@hyrrokkin>  H On 02 Apr 2005 20:13:26 GMT, Thierry Dussuet <thierry@dussuet.lugs.ch>   wrote:  3 > On 2005-04-02, Tom Linden <tom@kednos.com> wrote: I >> On 01 Apr 2005 16:45:15 GMT, Thierry Dussuet <thierry@dussuet.lugs.ch> 	 >> wrote:  >>H >>> On 2005-03-31, Alan Frisbie <Usenet02_REMOVE@Flying-Disk.com> wrote:B >>>> Since I like to have remote Internet access to my VMS system,2 >>>> I have my firewall forward SSH packets to it. >>>>= >>>> However, every day or so, I am alerted by the beeping of C >>>> OPCOM messages that someone is making repeated login attempts. = >>>> I am not worried about them getting in, but I find their  >>>> feeble attempts annoying. >>>>A >>>> I am thinking of creating captive accounts that these script > >>>> kiddies would be likely to try (root/<null>, guest/guest,@ >>>> god/god, etc.).   The captive procedure should appear to be@ >>>> some variant of Unix, appear to sort-of work, while wasting= >>>> as much of their time as possible.   Meanwhile it should A >>>> log everything to a file for my later amusement.   (Yes, I'm  >>>> that sick!) >>>>? >>>> I seem to recall that someone created such a DCL procedurel? >>>> years ago, but I can't seem to find it.   Can anyone pointy@ >>>> me to something that would give me a start on this project? >>>SD >>> I started something similar some time ago, in PL/I.  The code is >>> probably asnH >>> ugly as hell (I'm happy for feedback) but the result is funny :) (it >>> doesn'tn >>> all work yet, either)E >>>r. >>> http://wigwam.ethz.ch/~dussuett/prompt.pli >>>r >>> Thierry  >>J >> Cute.  Now, how are you launching the program or what causes it to run? >b@ > I have a captive account root/root, with login.com as follows: > % > $ DEFINE/USER SYS$INPUT SYS$COMMANDc > $ RUN SYS$LOGIN:PROMPT.EXE
 > $ LOGOUT >  > And the flags:@ > Flags:  DefCLI DisWelcome DisNewMail DisMail DisReport Captive >p9 > (Haven't found out how to replace the CLI itself (yet))- >-	 > Thierry-J BTW,  you don't need to have the external declarations, they cab go insideK PROMPT as can all the procedures, since it is lexically scoped.  Also,  you7H didn't need the select statement for determining the month, but then you" weren't looking for efficiency :-)  / format_date: Procedure returns (character(12));:  # declare current_date character(16), .          current_formatted_date character(12),          month character(3),          day character(2),          hour character(2),           minutes character(2);# dcl months(12) char(3) static init(tI 'Jan','Feb','Mar','Apr','Mai','Jun','Jul','Aug','Sep','Oct','Nov','Dec');r  !          on subscriptrange begin;s8                  put list (substr (current_date, 5, 2));.                  put skip list (current_date);=                  put skip list('Error Status is '!!oncode());r                  end;m  $          current_date = datetime ();<          month = months(fixed(substr (current_date, 5, 2)));  +          day = substr (current_date, 7, 2);h,          hour = substr (current_date, 9, 2);0          minutes = substr (current_date, 11, 2);M          current_formatted_date = month !! ' ' !! day !! ' ' !! hour !! ':'  o !! mit nutes;)          return (current_formatted_date);t          end format_date;A   ------------------------------  % Date: Sun, 03 Apr 2005 16:44:57 +0200 & From: Paul Sture <paul.sture@decus.ch>( Subject: Re: Script Kiddie tarpit wanted, Message-ID: <3badr9F6hconuU1@individual.net>   Thierry Dussuet wrote:  8 > On 2005-04-03, Paul Sture <paul.sture@decus.ch> wrote: >  >>Thierry Dussuet wrote: >>@ >>from a lugs.ch address! (That's Linux User Group Switzerland). >> >>Ooooh. :-) >  >  > :-)  > E > This is just a UUCP forwarder to be sure to get all the mails^Wspam   9 Yes, use whatever tool works to keep the spammers at bay!0   ------------------------------  % Date: Sun, 03 Apr 2005 08:06:35 -0700 # From: "Tom Linden" <tom@kednos.com>s4 Subject: Re: There is certainly opportunity in chaos( Message-ID: <opsonv89iizgicya@hyrrokkin>  G On Sun, 03 Apr 2005 00:53:55 -0500, Dave Froble <davef@tsoft-inc.com>  p wrote:  L > Going afield some consider another example.  Some years ago Chrysler was  L > in trouble.  Lee Iacocoa (probably misspelled) came in from Ford, partly  I > because he didn't get the head job there, to turn things around.  For  uG > his salary, he asked for $1 a year, until Chrysler became profitable.   F AIR, that wasn't the whole story, he got $5B in loan guarantees from   CongressK and his $1/year was a ruse, he ultimately got considerable back pay later  e on.a Iacoccae   ------------------------------  % Date: Sun, 03 Apr 2005 12:42:57 -0400w' From: Dave Froble <davef@tsoft-inc.com>s4 Subject: Re: There is certainly opportunity in chaos0 Message-ID: <11506lq1pb7ga47@corp.supernews.com>   Tom Linden wrote:cI > On Sun, 03 Apr 2005 00:53:55 -0500, Dave Froble <davef@tsoft-inc.com>  x > wrote: > H >> Going afield some consider another example.  Some years ago Chrysler J >> was  in trouble.  Lee Iacocoa (probably misspelled) came in from Ford, D >> partly  because he didn't get the head job there, to turn things D >> around.  For  his salary, he asked for $1 a year, until Chrysler  >> became profitable.0 >  > H > AIR, that wasn't the whole story, he got $5B in loan guarantees from  
 > CongressF > and his $1/year was a ruse, he ultimately got considerable back pay  > later  on.	 > Iacoccae  = Yes, but he got it after he got the company profitable again.>   ------------------------------   End of INFO-VAX 2005.186 ************************