1 INFO-VAX	Sat, 12 Jun 2004	Volume 2004 : Issue 324       Contents: Re: DEC PWS RAM  Development tools for VMS  Re: Development tools for VMS  Re: Development tools for VMS  Re: Development tools for VMS  Re: Development tools for VMS 9 Re: executing //jobs in os/390 or vms posted via internet % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks % Re: Experience with solid state disks ! Re: Free surplus SW Prod Lib kits  Re: Is /verify necessary?  Re: Is /verify necessary?  Re: Is /verify necessary?  Re: Is /verify necessary? , Re: Opportunity for OVMS in mainstream press, Re: Opportunity for OVMS in mainstream press, Re: Opportunity for OVMS in mainstream press, Re: Opportunity for OVMS in mainstream press, Re: Opportunity for OVMS in mainstream press% Poll: Does HP do enough to promote... - Preventing ICMP redirects from being accepted  SWS-2.0 (Apache) & CGI ???
 SYSUAF.DAT Re: SYSUAF.DAT- tcpdump fails miserably, how to make it work? 1 Re: tcpdump fails miserably, how to make it work? 1 Re: tcpdump fails miserably, how to make it work? 3 Tru64 --> OpenVMS rather than another unix platform  Re: What might have been4 Re: Where have all the rec.travel.air regulars gone?2 Re: Why does idle telnet session get disconnected?, Re: [OT] An interesting view of Linux's cost [OT] Windows TNT  F ----------------------------------------------------------------------  % Date: Fri, 11 Jun 2004 23:16:01 +0100 ) From: Antonio Carlini <arcarlini@iee.org>  Subject: Re: DEC PWS RAM+ Message-ID: <3aqyc.306$vg2.157@newsfe6-win>    Rich Jordan wrote:   > From memory:   I'll ignore the pun :-)   E > Old PC100 memory was capable of running on PC66 systems.  Old PC133 G > memory was capable of running on PC100 systems, and some few types of = > PC133 could actually run on PC66 systems.  As manufacturing ? > improvements came in it became uneconomical to keep using the F > 'downward compatible' chips on the DIMM boards, so later PC100 wouldE > not run reliably on PC66, and newer PC133 would not run reliably on  > PC100.  > I guess I'll have to dig out some old data sheets and actually; look, but *why* would PC100 not run at 66MHz? I can see why ; trying to run it at 133MHz would fail (the answer would not : be ready in time), but why would running at a slower speed
 be a problem?    Antonio    --   --------------- - Antonio Carlini             arcarlini@iee.org    ------------------------------    Date: 11 Jun 2004 11:32:58 -0700& From: twnews@kittles.com (Thomas Wirt)" Subject: Development tools for VMS= Message-ID: <b3531425.0406111032.56a22fe1@posting.google.com>   F I am submitting the below question for one of my lead developers.  AnyD information (even brief) on the subject of VMS development tools and# environments, would be appreciated.    Thanks,    Thomas Wirt  Systems Manager    ********************  > We are currently an OpenVMS shop and write all our proprietary application E programs using DEC Basic and DEC Forms with Oracle Rdb as the backend B database. We have a Menu program also written in DEC Basic and DEC Forms @ that is used to navigate from application program to application program.E The user interface hardware is either a VT character cell terminal or  a PC running VT emulation.   A   We are in the process of evaluating new application development 
 tools that@ will allow us to produce GUI applications, or even browser basedF applications (Microsoft .NET, Oracle JDeveloper, etc).  Our hope is to findF an application development tool that can deploy an application program to@ all three types of user interfaces: character cell terminal, GUI
 screen, orF internet browser, as it will be quite a few years until we can replace all D our VT terminals with PC's.  It will also be quite a few years until we canE convert our existing legacy system of Basic/DEC Forms programs to the  new < application tool as we are an extremely small shop of only 3 application = developers and spend most if not all of our time creating new 
 applications, > instead of converting old applications over to the new format.   I have three issues:A 	1.)	Does anyone know of an application development tool that can E deploy to all three user interfaces that I mentioned (character cell,  GUI,	 browser)? A 	2.)	Does anyone know of a conversion tool that would convert DEC E Basic and DEC Forms over to another technology like .NET, Java/Swing,  or other?> 	3.)	Once we are deploying GUI applications, how do I create a@ menu system that allows the users to access the legacy DEC Forms programs as @ well as the new GUI programs?  I don't want the users to have to
 figure outF if a particular application is available from the "old" character cell menu,  or the "new" GUI menu.  " Thanks for any light you can shed!  
 Dan Herron Senior Programmer Analyst  Kittles Home Furnishings Indianapolis, Indiana  <mailto:dherron@kittles.com>   ------------------------------  % Date: Fri, 11 Jun 2004 15:45:47 -0400 2 From: "Stanley F. Quayle" <squayle@insight.rr.com>& Subject: Re: Development tools for VMS/ Message-ID: <40C9D3AB.25601.2CAB8E56@localhost>   C >   We are in the process of evaluating new application development G > tools that will allow us to produce GUI applications, or even browser > > based applications (Microsoft .NET, Oracle JDeveloper, etc).  = As a WRQ reseller [Shameless Plug (tm)], might I suggest WRQ   VeraStream?   ? You divide up your character-cell application into chunks that  B implement  business rules.  You can then create a web application E that queries your VMS system dynamically without the users having to  , learn the character-cell application at all.  
 Check out:  )   http://www.wrq.com/products/verastream/    for more info...      
   Our hope? > is to find an application development tool that can deploy an F > application program to all three types of user interfaces: characterG > cell terminal, GUI screen, or internet browser, as it will be quite a D > few years until we can replace all our VT terminals with PC's.  ItB > will also be quite a few years until we can convert our existingG > legacy system of Basic/DEC Forms programs to the new application tool H > as we are an extremely small shop of only 3 application developers andF > spend most if not all of our time creating new applications, instead8 > of converting old applications over to the new format. >  > I have three issues:C >  1.)	Does anyone know of an application development tool that can G > deploy to all three user interfaces that I mentioned (character cell,  > GUI, browser)?C >  2.)	Does anyone know of a conversion tool that would convert DEC G > Basic and DEC Forms over to another technology like .NET, Java/Swing,  > or other? @ >  3.)	Once we are deploying GUI applications, how do I create aB > menu system that allows the users to access the legacy DEC Forms
 > programs as B > well as the new GUI programs?  I don't want the users to have to > figure outH > if a particular application is available from the "old" character cell > menu, or the "new" GUI menu. > $ > Thanks for any light you can shed! >  > Dan Herron > Senior Programmer Analyst  > Kittles Home Furnishings > Indianapolis, Indiana  > <mailto:dherron@kittles.com>  
 --Stan Quayle  Quayle Consulting Inc.  
 ----------- Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363 3 8572 North Spring Ct., Pickerington, OH  43147  USA 0 stan-at-stanq-dot-com       http://www.stanq.com   ------------------------------  # Date: Fri, 11 Jun 2004 20:05:44 GMT % From: "Bob Lail" <Robert.Lail@hp.com> & Subject: Re: Development tools for VMS2 Message-ID: <sgoyc.3551$fO3.1718@news.cpqcorp.net>  3 "Thomas Wirt" <twnews@kittles.com> wrote in message 7 news:b3531425.0406111032.56a22fe1@posting.google.com... H > I am submitting the below question for one of my lead developers.  AnyF > information (even brief) on the subject of VMS development tools and% > environments, would be appreciated.  > 	 > Thanks,  > 
 > Thomas Wirt  > Systems Manager  >  > ******************** > @ > We are currently an OpenVMS shop and write all our proprietary
 > application G > programs using DEC Basic and DEC Forms with Oracle Rdb as the backend D > database. We have a Menu program also written in DEC Basic and DEC > Forms B > that is used to navigate from application program to application
 > program.G > The user interface hardware is either a VT character cell terminal or  > a PC > running VT emulation.  > C >   We are in the process of evaluating new application development  > tools thatB > will allow us to produce GUI applications, or even browser basedH > applications (Microsoft .NET, Oracle JDeveloper, etc).  Our hope is to > findH > an application development tool that can deploy an application program > toB > all three types of user interfaces: character cell terminal, GUI > screen, orH > internet browser, as it will be quite a few years until we can replace > all F > our VT terminals with PC's.  It will also be quite a few years until > we canG > convert our existing legacy system of Basic/DEC Forms programs to the  > new > > application tool as we are an extremely small shop of only 3
 > application ? > developers and spend most if not all of our time creating new  > applications, @ > instead of converting old applications over to the new format. >  > I have three issues:B > 1.) Does anyone know of an application development tool that canG > deploy to all three user interfaces that I mentioned (character cell,  > GUI, > browser)? B > 2.) Does anyone know of a conversion tool that would convert DECG > Basic and DEC Forms over to another technology like .NET, Java/Swing,  > or > other?? > 3.) Once we are deploying GUI applications, how do I create a B > menu system that allows the users to access the legacy DEC Forms
 > programs as B > well as the new GUI programs?  I don't want the users to have to > figure outH > if a particular application is available from the "old" character cell > menu,  > or the "new" GUI menu. > $ > Thanks for any light you can shed! >  > Dan Herron > Senior Programmer Analyst  > Kittles Home Furnishings > Indianapolis, Indiana  > <mailto:dherron@kittles.com>      Dan  D    Check out WRQ's new Verastream product to solve your #2 question,J DECforms to Java or .NET conversion issue. Verastream could also help with your #3 question as well.   .         http://www.wrq.com/products/verastream  H    Also check out the Data/Application Access and e-business developmentI technology HP has for OpenVMS. BridgeWorks, Attunity Connect, DECform WEB  Connector, etc for data access.   L For your Development Environment, NetBeans, and Apache Xerces XML technologyJ for development.  GTK+ is available for GUI development. You can find info) on all these things at the following url.   :        http://h71000.www7.hp.com/ebusiness/technology.html     \Bob Lail      --  
 Robert G Lail  Senior Solution Architect  Corporate Accounts Hewlett-Packard Company 
 Merrimack, NH  TEL: 603.424.6272  CEL: 603.315.0556    ------------------------------  % Date: Fri, 11 Jun 2004 18:11:54 -0500 - From: "Shael Richmond" <ksrich@bellsouth.net> & Subject: Re: Development tools for VMS8 Message-ID: <hYqyc.5421$X92.2374@bignews6.bellsouth.net>  3 "Thomas Wirt" <twnews@kittles.com> wrote in message 7 news:b3531425.0406111032.56a22fe1@posting.google.com... H > I am submitting the below question for one of my lead developers.  AnyF > information (even brief) on the subject of VMS development tools and% > environments, would be appreciated.  > 	 > Thanks,  > 
 > Thomas Wirt  > Systems Manager  >  > ******************** > @ > We are currently an OpenVMS shop and write all our proprietary
 > application G > programs using DEC Basic and DEC Forms with Oracle Rdb as the backend D > database. We have a Menu program also written in DEC Basic and DEC > Forms B > that is used to navigate from application program to application
 > program.G > The user interface hardware is either a VT character cell terminal or  > a PC > running VT emulation.  > C >   We are in the process of evaluating new application development  > tools thatB > will allow us to produce GUI applications, or even browser basedH > applications (Microsoft .NET, Oracle JDeveloper, etc).  Our hope is to > findH > an application development tool that can deploy an application program > toB > all three types of user interfaces: character cell terminal, GUI > screen, orH > internet browser, as it will be quite a few years until we can replace > all F > our VT terminals with PC's.  It will also be quite a few years until > we canG > convert our existing legacy system of Basic/DEC Forms programs to the  > new > > application tool as we are an extremely small shop of only 3
 > application ? > developers and spend most if not all of our time creating new  > applications, @ > instead of converting old applications over to the new format. >  > I have three issues:B > 1.) Does anyone know of an application development tool that canG > deploy to all three user interfaces that I mentioned (character cell,  > GUI, > browser)? B > 2.) Does anyone know of a conversion tool that would convert DECG > Basic and DEC Forms over to another technology like .NET, Java/Swing,  > or > other?? > 3.) Once we are deploying GUI applications, how do I create a B > menu system that allows the users to access the legacy DEC Forms
 > programs as B > well as the new GUI programs?  I don't want the users to have to > figure outH > if a particular application is available from the "old" character cell > menu,  > or the "new" GUI menu. >   B You can use the DECForms WebConnector product to provide a browserG interface without any code changes.  We tested it with Fortran and DBMS  and worked pretty good.    Shael    ------------------------------  % Date: Sat, 12 Jun 2004 08:00:37 +0800 , From: Paul Repacholi <prep@prep.synonet.com>& Subject: Re: Development tools for VMS- Message-ID: <87659xd4re.fsf@prep.synonet.com>   ( twnews@kittles.com (Thomas Wirt) writes:  C > I am submitting the below question for one of my lead developers. F > Any information (even brief) on the subject of VMS development tools) > and environments, would be appreciated.   @ > We are currently an OpenVMS shop and write all our proprietaryD > application programs using DEC Basic and DEC Forms with Oracle RdbE > as the backend database. We have a Menu program also written in DEC ? > Basic and DEC Forms that is used to navigate from application A > program to application program.  The user interface hardware is C > either a VT character cell terminal or a PC running VT emulation.   C >   We are in the process of evaluating new application development ? > tools that will allow us to produce GUI applications, or even F > browser based applications (Microsoft .NET, Oracle JDeveloper, etc).E > Our hope is to find an application development tool that can deploy ? > an application program to all three types of user interfaces: F > character cell terminal, GUI screen, or internet browser, as it willE > be quite a few years until we can replace all our VT terminals with C > PC's.  It will also be quite a few years until we can convert our ? > existing legacy system of Basic/DEC Forms programs to the new > > application tool as we are an extremely small shop of only 3> > application developers and spend most if not all of our timeC > creating new applications, instead of converting old applications  > over to the new format.   G Why? The `new' scheme will do nothing that you are not doing now. Well, F except run up costs, and open you up to more bugs and malware than you want to imagine.   > I have three issues:  ? > 	1.) Does anyone know of an application development tool that E > can deploy to all three user interfaces that I mentioned (character  > cell, GUI, browser)?  ? > 	2.) Does anyone know of a conversion tool that would convert ? > DEC Basic and DEC Forms over to another technology like .NET,  > Java/Swing, or other?   @ > 	3.) Once we are deploying GUI applications, how do I create aB > menu system that allows the users to access the legacy DEC FormsF > programs as well as the new GUI programs?  I don't want the users toF > have to figure out if a particular application is available from the3 > "old" character cell menu, or the "new" GUI menu.   C These are problems, not solutions... Not going `new' and `improved'  solves all 3.    You owe the doctor 25c.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 11 Jun 2004 22:42:22 -0700# From: mao@gnu.org (Andrew Makhorin) B Subject: Re: executing //jobs in os/390 or vms posted via internet< Message-ID: <4209937f.0406112142.ff838b2@posting.google.com>  ^ "Hans Vlems" <hvlems.dotweg@zonnet.nl> wrote in message news:<2iu84uFr5ou3U1@uni-berlin.de>...4 > "Andrew Makhorin" <mao@gnu.org> schreef in bericht9 > news:4209937f.0406110205.3a27f3f1@posting.google.com... J > > Does anyone know if there is a server publically available through theJ > > internet which is able to execute //jobs in os/390 or mvs environment?F > > I mean that I post a //job to that server and receive its listing. > >  > > Thank you, > >  > > Andrew Makhorin  > I > Not that I know of, but why don't you install Hercules and run MVS 3.8?   A I've got Hercules installed as well as os/360, but I need os/390.   G (I bring my apologies that I intervented into the wrong forum; in fact, $ vms and mvs are different things :+)  
 Thank you,   Andrew Makhorin    ------------------------------  % Date: Fri, 11 Jun 2004 11:14:56 -0400 * From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: Experience with solid state disks2 Message-ID: <g7SdnR5ux7rcUVTdRVn-tA@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:i2fyF5+FQqAR@eisner.encompasserve.org... @ > In article <YsedncUt0ff1r1TdRVn-sA@metrocast.net>, "Bill Todd"  <billtodd@metrocast.net> writes:   ...   E > >> I highly doubt you will be able to get better than 1 millisecond ' > >> on writes with a solid state disk.  > > L > > That could still offer several times the performance of a physical write toK > > a magnetic disk.  And I suspect that there are configurations (possibly F > > using fewer intermediaries than the one you're familiar with does) offering? > > somewhat lower latency than 1 ms.  ISTR reported one-way FC 
 communication H > > latencies as good as around 10 us., for example - and there's really littleJ > > reason for a *great* deal more overhead than that in the processing at both	 > > ends.  > >  > B > Yes - but.  With the EVA having a LUN across many disks, coupled; > with the fact the write has to go out and back and acked, B > realistically is there any OS doing writes under 1 ms to a solid
 > state disk?   L 1.  A redo log LUN likely wouldn't be spread across many disks (though stillI needs to be mirrored).  But that shouldn't be an issue anyway, because in C the EVA the round trip latency presumably is only to the controller G write-back cache and back, never seeing effects due to the backing disk  layout.   J 2.  There was also the question of whether the SSDs could be placed withinJ the EVA at all.  If they can't, internal EVA overhead (whatever it may be) is not at issue.  H There's no *intrinsic* reason why a write to a FC-connected SSD couldn'tJ complete in around 100 us. - possibly even somewhat less.  I vaguely thinkB that I remember vendor specs for some SSD latencies in this range.   > ? > I don't see how there can be an advantage or where the win is > > at.  Random read IOs, okay.  But writes to write-back cached > controllers?  J If the poster's redo logs are maintained in his EVA, then I agree - unlessJ they have the EVA's write-back caching disabled (or, possibly, are stripedD at such fine grain and with so small a portion of the EVA cache thatI back-end disk bandwidth is a problem) - or unless the EVA has rather poor  response times internally.  /   Maybe if as you allude to above, it is direct < > attached - okay.  But 3 nodes are involved the solid state? > disk has to hang off a switch so the write path would be HBA, 1 > switch, SSD and back again.  1 ms or so, right?   H As I noted, FC latency can be a couple of orders of magnitude lower thanL that.  Switch latency can be as well.  But the operative phrase is 'can be':A my impression is that latencies vary quite a lot between vendors.   A And I'm sure that I remember SSD latency specs *well* under 1 ms.    - bill   ------------------------------  # Date: Fri, 11 Jun 2004 15:40:38 GMT 1 From: Michael Austin <maustin@firstdbasource.com> . Subject: Re: Experience with solid state disks< Message-ID: <Wnkyc.5763$x65.3384@newssvr22.news.prodigy.com>   Bill Todd wrote:: > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:i2fyF5+FQqAR@eisner.encompasserve.org...  > @ >>In article <YsedncUt0ff1r1TdRVn-sA@metrocast.net>, "Bill Todd" > " > <billtodd@metrocast.net> writes: >  > ...  >  > D >>>>I highly doubt you will be able to get better than 1 millisecond& >>>>on writes with a solid state disk. >>> K >>>That could still offer several times the performance of a physical write  >  > to > J >>>a magnetic disk.  And I suspect that there are configurations (possiblyE >>>using fewer intermediaries than the one you're familiar with does)  > 
 > offering > > >>>somewhat lower latency than 1 ms.  ISTR reported one-way FC >  > communication  > G >>>latencies as good as around 10 us., for example - and there's really  >  > little > I >>>reason for a *great* deal more overhead than that in the processing at  >  > both >  >>>ends. >>>  >>B >>Yes - but.  With the EVA having a LUN across many disks, coupled; >>with the fact the write has to go out and back and acked, B >>realistically is there any OS doing writes under 1 ms to a solid
 >>state disk?  >  > N > 1.  A redo log LUN likely wouldn't be spread across many disks (though stillK > needs to be mirrored).  But that shouldn't be an issue anyway, because in E > the EVA the round trip latency presumably is only to the controller I > write-back cache and back, never seeing effects due to the backing disk 	 > layout.  >   F Bill, the EVA does not allocate luns by spindle - it allocates space. D All luns will be spread across an entire disk group - to the extent G necessary - even if using RAID0,1,5 (only options) The EVA is not your  G average storage controller. And you really don't want to create a disk  E group for each LUN.  You would then loose the ability to effectively  H manage all of your space and you go back to managing spindles.  Rule of G thumb is to make the disk groups as large as possible.  I had one site  H that had 2 groups of 120 drives (240 is the MAX) on one, they would put D the LUN for database/index and the other redo log files etc... Even I though it is extremely rare, if you lost an entire disk group, you would   still be able to recover.   L > 2.  There was also the question of whether the SSDs could be placed withinL > the EVA at all.  If they can't, internal EVA overhead (whatever it may be) > is not at issue. >   ?   I would say no.  These drives are fibre not plug-n-play SCSI.   J > There's no *intrinsic* reason why a write to a FC-connected SSD couldn'tL > complete in around 100 us. - possibly even somewhat less.  I vaguely thinkD > that I remember vendor specs for some SSD latencies in this range. >  > ? >>I don't see how there can be an advantage or where the win is > >>at.  Random read IOs, okay.  But writes to write-back cached >>controllers? >  > L > If the poster's redo logs are maintained in his EVA, then I agree - unlessL > they have the EVA's write-back caching disabled (or, possibly, are stripedF > at such fine grain and with so small a portion of the EVA cache thatK > back-end disk bandwidth is a problem) - or unless the EVA has rather poor  > response times internally. > 1 >   Maybe if as you allude to above, it is direct  > < >>attached - okay.  But 3 nodes are involved the solid state? >>disk has to hang off a switch so the write path would be HBA, 1 >>switch, SSD and back again.  1 ms or so, right?  >  > J > As I noted, FC latency can be a couple of orders of magnitude lower thanN > that.  Switch latency can be as well.  But the operative phrase is 'can be':C > my impression is that latencies vary quite a lot between vendors.  > C > And I'm sure that I remember SSD latency specs *well* under 1 ms.  >  > - bill >  >  >   
 Michael...   ------------------------------  # Date: Sat, 12 Jun 2004 00:47:08 GMT 1 From: "konabear" <konabearg-newsgroups@yahoo.com> . Subject: Re: Experience with solid state disks< Message-ID: <fosyc.7363$n65.2802@newssvr33.news.prodigy.com>  > "Michael Austin" <maustin@firstdbasource.com> wrote in message6 news:qkHxc.5363$295.1919@newssvr22.news.prodigy.com... <snip>H > But first, you need to quantify that your redo logs are the bottleneckI > before throwing hardware at it. And unfortunately, the EVA doesn't have E > any performance monitoring features.  So you are unable to actually J > quanitify anything, except what you can get out of Monitor (look at yourF > T4 disk data WRT to Resp and Queue Depth.  If you don't have a queueG > depth, then your caching would be shielding your system from any disk H > latencies that would be corrected by a SSD and therefore you would not > see any performance gains.  J Actually I disagree that queue lengths are a reliable predictable of cacheL hits.  As I/O request rates increase, the queue depth will increase as well.J Little's Law says response time is queue depth divided by rate/sec.  So asL long as the queue depth stay directly proportional to the request rate, lifeG is good.  It's when the queue length increases as a faster rate than is F proportional to the request rate, that response times start to suffer.  J I'd argue watching response times.  There will still be a response time asK the host CPU, FC transfer, and controller will take time.  However even the L best FC disks today have an average service time in the 5-6ms range.  If theE average response time less than 5ms certainly there is some amount of  caching.  I As EVAs are mentioned in this case, there is the potential for dozens and J dozens of spindles available to service the I/O requests.  Response time =L Service time when there is no competition.  So as long as individual I/Os doJ not see any competition, then it is possible for the response time to stayJ very flat for dozens or hundreds of I/Os a second even with no cache hits.  F I hadn't visited hard drive service times in a while.  Gee the averageL rotational latency of a 15K RPM drive is only 0.033ms! AND modern drives are3 clocking average seek times in the 5.1-5.5ms range.    Todd   ------------------------------  % Date: Fri, 11 Jun 2004 20:52:04 -0400 * From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: Experience with solid state disks2 Message-ID: <MYSdnV736KoZzlfdRVn-sA@metrocast.net>  > "Michael Austin" <maustin@firstdbasource.com> wrote in message6 news:Wnkyc.5763$x65.3384@newssvr22.news.prodigy.com...   ...   G > Bill, the EVA does not allocate luns by spindle - it allocates space. E > All luns will be spread across an entire disk group - to the extent H > necessary - even if using RAID0,1,5 (only options) The EVA is not your > average storage controller.    I understand that.  +  And you really don't want to create a disk  > group for each LUN.   J In the case of redo logs, you might:  not only would fine-grained stripingH waste disk utilization within the array (with the write-back caching andJ sequential access pattern, you want to write at least 1 MB or so to a disk@ before moving on to the next - though Richie once mentioned thatH distribution granularity was about a MB, so if that's the stripe segmentH size this isn't too much of an issue), but your probability of data lossK increases with the number of disks the log is striped across (the same kind K of increase that you see with RAID-5 arrays even if you're mirroring, since L IIUC mirrored replication within the EVA is distributed semi-randomly ratherG than using paired disks:  while the probability is still very small, if E there's *anything* you don't want to lose, it's likely the redo log).    - bill   ------------------------------  % Date: Fri, 11 Jun 2004 21:02:10 -0400 * From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: Experience with solid state disks2 Message-ID: <YtGdnT3QdO1-yFfdRVn-hw@metrocast.net>  < "konabear" <konabearg-newsgroups@yahoo.com> wrote in message6 news:fosyc.7363$n65.2802@newssvr33.news.prodigy.com...   ...    the average 8 > rotational latency of a 15K RPM drive is only 0.033ms!  B I think you'll find that it's 2 ms. (full rotation 4 ms.).  SoundsL suspiciously like a factor-of-60 error (assuming 15Krps rather than 15Krpm).    AND modern drives are5 > clocking average seek times in the 5.1-5.5ms range.   J Those are more like average access times for a 15Krpm drive.  Average seekH times have dropped to the 3 - 4 ms. range (even for some 10Krpm drives).   - bill   ------------------------------  % Date: Fri, 11 Jun 2004 23:05:59 +0100 ) From: Antonio Carlini <arcarlini@iee.org> * Subject: Re: Free surplus SW Prod Lib kits* Message-ID: <F0qyc.304$vg2.70@newsfe6-win>   Tom Linden wrote:   H > So I wonder If I am allowed to give away these kits.  If so, then whatB > is the difference between giving a physical vs. electronic copy?  2 If my reading of the hobbyist stuff is right, once1 you have the hobbyist licence, you are allowed to / acquire your kit anyway you like. Borrowing one / and making a copy and being given one seemed to  be expressly allowed.   1 My reading of the CONDIST/CONOLD stuff was always / that you could possess the physical CDs without . any need of a licence (or at least that seemed- to be the legal position in the UK). Actually * using anything on them required a licence.  ( There are plenty of people who sell kits% on the various auction sites and I've ( never heard of anyone having any bother.  - I've never seen anyone sell a licence PAK ...    Antonio    --   --------------- - Antonio Carlini             arcarlini@iee.org    ------------------------------    Date: 11 Jun 2004 10:45:25 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) " Subject: Re: Is /verify necessary?3 Message-ID: <geV6yWMDRggu@eisner.encompasserve.org>   h In article <8mjyc.3503$4l3.3029@news.cpqcorp.net>, hammond@not@peek.ssr.hp.com (Charlie Hammond) writes:l > In article <40c9b1c7$0$433$7f8943f3@newsreader.visi.com>, Lyndon Bartels <lbartels@pressenter.com> writes:K >>I want to rebuild a disk that's gotten quite fragmented. So my intent is  G >>to backup/image from disk to scratch disk, and the backup/image from   >>scratch back to original.  >>J >>I vaguely remember reading that when doing an image backup from disk to A >>disk the verify qualifier isn't really necessary. Is that true?  > E > (1) The likely hood that the disk dirve will write bad data without 4 > detecting and reportign this is vanishingly low.   > F > (2) I think that a better check is to actually BOOT the backup copy., > This demonstrates that the backup is good.  F I didn't see a stipulation that this was a system disk, but exercisingF all code in all images would be necessary to test whether they all got- copied correctly.  /VERIFY is so much easier.    ------------------------------    Date: 11 Jun 2004 10:43:28 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) " Subject: Re: Is /verify necessary?3 Message-ID: <v7MELAR$XJoh@eisner.encompasserve.org>   j In article <40c9b1c7$0$433$7f8943f3@newsreader.visi.com>, Lyndon Bartels <lbartels@pressenter.com> writes:K > I want to rebuild a disk that's gotten quite fragmented. So my intent is  G > to backup/image from disk to scratch disk, and the backup/image from   > scratch back to original.  > J > I vaguely remember reading that when doing an image backup from disk to A > disk the verify qualifier isn't really necessary. Is that true?   B It depends on how important it is to be sure there were no errors.  9 > It'd save me some time, and time on this task is tight.u  : Only you can make the tradeoff between speed and accuracy.  = Many years ago (as will be obvious from the following), I gotm< verification failures on a large backup.  The problem turned? out to be a bad Massbus cable or plug which was _only_ detectedd? by BACKUP/VERIFY and could only be isolated with a Massbus Timei3 Domain Reflectometer (a fairly rare piece of gear).i  ; Yes, your cabling today is different, but hardware is stille: susceptible to failures, and BACKUP/VERIFY can catch them.   ------------------------------  # Date: Fri, 11 Jun 2004 18:26:22 GMTe3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)o" Subject: Re: Is /verify necessary?2 Message-ID: <iPmyc.3536$tx3.1888@news.cpqcorp.net>  4 In article <geV6yWMDRggu@eisner.encompasserve.org>, / Kilgallen@SpamCop.net (Larry Kilgallen) writes:o  G >I didn't see a stipulation that this was a system disk, but exercisingiG >all code in all images would be necessary to test whether they all gotF >copied correctly.    C In theory, yes.  In practice, if OpenVMS boots from the backup the a1 probability of a bad backup is pretty darn small.a   -- wJ       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Sat, 12 Jun 2004 05:27:40 +0800-, From: Paul Repacholi <prep@prep.synonet.com>" Subject: Re: Is /verify necessary?- Message-ID: <87u0xhdbub.fsf@prep.synonet.com>   0 Lyndon Bartels <lbartels@pressenter.com> writes:  @ > I want to rebuild a disk that's gotten quite fragmented. So my> > intent is to backup/image from disk to scratch disk, and the- > backup/image from scratch back to original.b  F > I vaguely remember reading that when doing an image backup from diskD > to disk the verify qualifier isn't really necessary. Is that true?  9 > It'd save me some time, and time on this task is tight.b  C > Oh... for those who are thinking of asking... no, we don't have aoB > defragmentor running because 1.) this is the only disk that everE > needs defragmenting. 2.) It doesn't need it very often (maybe twicew	 > a year)C  D I nearly re-started a disk to disk backup to drop the /ver. Descided I could use a slow coffee. e  D After the error messages stopped, I got a second coffee before I didB anything else. Verfiy errors are not MEANT to happen disk to disk.     -- @< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Fri, 11 Jun 2004 17:47:43 GMTW- From: JF Mezei <jfmezei.spamnot@teksavvy.com> 5 Subject: Re: Opportunity for OVMS in mainstream presso@ Message-ID: <f5d61164b49e62b78e480e1b6938b5ca@news.teranews.com>   "Dr. Dweeb" wrote: > > - Telecom billing systems.  L Compaq had a whole web page advertising its Unix based billing solutions for the telecom sector.   M > At least previously, JPL stored rocket engine statistical and other data onl > OpenVMS/OracleRdb.  M Lots of "historical" information. But to impress people, you need to show neweM VMS sites that are continuing their investment in VMS and adding applicationssM to their VMS infrastructure. The vast majroity of VMS shops are slowly movingoV their IT off of VMS as opportinuties/applications become available on other platforms.  " > Is E*Trade is still on OpenVMS ?  M The fact that one must ask the question is an admission that there is a trendeH to move away from VMS. And only HP can fix that image through marketing.  L > Deutsche Bank Money Market Trading (was Bankers Trust) - OpenVMS (not sure > about data storage)h  M Bankers Trust used to be a very serious VMS shop with very serious involvmentoM in the DECNET/SNA products.  But I lost a contract to an organisation becausesK of a BT senior employee along with a Citibank senior employee who ruled outtK any VMS based platforms because VMS was dead and corporate policies to staya! away from VMS (or so I was told).a    L I think that the only serious public "win" was probably Etrade.  But haven't& they been gobbled up by someone else ?   ------------------------------  % Date: Fri, 11 Jun 2004 19:29:44 +0200   From: "Dr. Dweeb" <dr@dweeb.com>5 Subject: Re: Opportunity for OVMS in mainstream pressh- Message-ID: <cacq6e$2ion$1@news.cybercity.dk>r   Jan-Erik Sderholm wrote:  > JF Mezei wrote:I >> >> (email privately seperately)u >> >> Kenneth Farmer wrote:: >>> 2. What applications does OpenVMS tend to be used for?A >> accounting, web serving, SMTP server,  some office automation,e8 >> development, network monitoring, FTP and many others. >n> > (Are smtp and ftp realy "applications" ? I'd call then "base > services"...)- >  > Anyway, don't forget : >5? > - The whole industial sector, such as all kinds of productionu > monitoring systems.t >sG > - Stock exchanges (with the Swedish company "OM Technology" as one ofi
 > the largestf >   software players.) > C > - The gaming/lottery area. The Swedish state owned "Svenska Spel"  > ("Swedish Games")aD >   runs most (all?) of theirs back-end systems suporting the online > terminals on VMS.P >   > None of what follows is terribly secret, so I will name names.  E ESSNet is also swedish and has 7-8% of the world Lotto market - again K OpenVMS/OracleRdb backend.  I thought it had Sweden as well, but apparentlyl now - shows how little I know.  I SciGames in Austria is also a player in that market - AFAIK they are alsow OpenVMS/OracleRdb.   > - Telecom billing systems. > H > - What about NASA ? I thought the system that continiously logs health > dataH > from the astronauts on the space shuttle was running on VMS and Rdb !? >r  K At least previously, JPL stored rocket engine statistical and other data oni OpenVMS/OracleRdb.  I Which brings us to the obvious point that if you want to understand whereaL all the big important sites are, then you could do worse than to ask the Rdb1 Mailing list - check out www.jcc.com for details.s     Other big sites.  > DirecTV orders/billing - OpenVMS/OracleRdb - Massive TP volumeJ NY State toll roads/RF billing - OpenVMS/OracleRdb - Massive data volumes.  Is E*Trade is still on OpenVMS ?* NYMEX trades directly on OpenVMS/OracleRdbJ Deutsche Bank Money Market Trading (was Bankers Trust) - OpenVMS (not sure about data storage)l    = Wheres Kerry when we need him, he has probably got a hundredt$ names/industries to add to the list.    	 Dr. Dweeb    > Jan-Erik.o   ------------------------------  % Date: Sat, 12 Jun 2004 01:55:54 +0800a, From: Paul Repacholi <prep@prep.synonet.com>5 Subject: Re: Opportunity for OVMS in mainstream press - Message-ID: <87aczadln9.fsf@prep.synonet.com>I  / JF Mezei <jfmezei.spamnot@teksavvy.com> writes:f   > Kenneth Farmer wrote:s  G >> 3. Why has it hung on so well despite so many reports of its demise?   C > BECAUSE IT WORKS. And because it doesn't require much maintenancerF > and is not prone to viri or system directory corruption. Robust file > system that you can trust.  F "We really should learn more about the Vax. But it just sits there and works."Q  C The head IBM operator at a large .au finance company. e-mail if you 
 want details.-   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.<@                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Fri, 11 Jun 2004 22:06:15 +0200n  From: "Dr. Dweeb" <dr@dweeb.com>5 Subject: Re: Opportunity for OVMS in mainstream press - Message-ID: <cad3br$2s4m$1@news.cybercity.dk>W   JF Mezei wrote:a > "Dr. Dweeb" wrote: >>> - Telecom billing systems. >a@ > Compaq had a whole web page advertising its Unix based billing > solutions forB > the telecom sector.  >cF >> At least previously, JPL stored rocket engine statistical and other >> data on OpenVMS/OracleRdb.  >hF > Lots of "historical" information. But to impress people, you need to
 > show newB > VMS sites that are continuing their investment in VMS and adding > applicationsA > to their VMS infrastructure. The vast majroity of VMS shops arel > slowly movingnG > their IT off of VMS as opportinuties/applications become available ono > other platforms. >o  J Jeez JF, I was not trying to impress anyone.  Without bothering to look atK the original post, in my mind the information requested was not specific to 
 new sites.  # >> Is E*Trade is still on OpenVMS ?i >,G > The fact that one must ask the question is an admission that there isn	 > a trendr? > to move away from VMS. And only HP can fix that image throughi > marketing. >t  K No argument from me - have you never read my other posts ?  I lost track oftJ the ETrade situation some time ago but last I heard OpenVMS had survived aI serious attempt to oust it, thus the ?, as I did not want to propogate ant untruth.  C >> Deutsche Bank Money Market Trading (was Bankers Trust) - OpenVMSn  >> (not sure about data storage) >:D > Bankers Trust used to be a very serious VMS shop with very serious > involvmentG > in the DECNET/SNA products.  But I lost a contract to an organisationw	 > because.C > of a BT senior employee along with a Citibank senior employee whoi > ruled outmE > any VMS based platforms because VMS was dead and corporate policies 	 > to stay # > away from VMS (or so I was told).o >c >tF > I think that the only serious public "win" was probably Etrade.  But	 > haven'ts( > they been gobbled up by someone else ?   ------------------------------  % Date: Fri, 11 Jun 2004 21:50:40 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>5 Subject: Re: Opportunity for OVMS in mainstream pressy+ Message-ID: <40CA6F80.AE45A217@comcast.net>r   "Alan E. Feldman" wrote: > g > David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<40C9147C.7E2614A9@comcast.net>...V > [...]  > >cE > > VMS invented clustering almost twenty years ago. Newer clustering. > E > My VMS screensaver says VMSclusters (well, VAXclusters at the time)p, > were introduced in 1983 ==> ~21 years ago. >  > [...]  >  > JMHO  E In that case. they were likely "invented" some years earlier. 1981 isi( probably not too outrageous a statement.  B Then again, Mark Levy and I had two VAX 11-750s sharing files over DECnet circa. 1983/84.  G ...and as I recall, we did DECnet copies from RSTS/E to VAX/VMS V3.2-.6. during my tenure there.l   Dunno...   D.J.D.   ------------------------------    Date: 11 Jun 2004 16:02:00 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) . Subject: Poll: Does HP do enough to promote...3 Message-ID: <ksdYnKsL2ciQ@eisner.encompasserve.org>4  M Visit, http://www.hpuseradvocacy.org/, now, to vote on the newest Instapoll: u  I Does HP do enough to promote the relevant platforms in your organization?c  R Yes. HP does enough. r4 HP sells at the right level within my organization. ? HP can do more to inform my organization on technology issues.  H HP can do more to inform my organization on business or service issues. B HP can provide more information via presentations in the regions. ? HP needs to spend more face-to-face time with my organization. n HP needs to do other things. d HP does nothing.  G Question does not apply. My organization partners with a VAR and has no  	direct contact with HP.     ------------------------------  % Date: Fri, 11 Jun 2004 13:55:29 -0500s* From: "Jim" <jim.nospam.gould@charter.net>6 Subject: Preventing ICMP redirects from being accepted- Message-ID: <cacv71$2mg$1@news.doit.wisc.edu>    Referring to the previous post  ? How is routing table updated when ROUTED and GATED aren't used.   I I just got off the phone with HP and found out that ipsendredirect is theeH wrong command to set.  That command will prevent the kernel from SENDING these commands.u   The correct command is( sysconfig -r inet icmp_rejectcodemask=32  < This causes the kernel to ignore any ICMP redirect commands.   ------------------------------  % Date: Fri, 11 Jun 2004 12:08:09 -0400m) From: "Neil Rieck" <n.rieck@sympatico.ca>s# Subject: SWS-2.0 (Apache) & CGI ???e; Message-ID: <KNkyc.52099$8k4.1165311@news20.bellglobal.com>n  E Is anyone out there using CSWS-2.0 (a.k.a. Apache for VMS), and if soe does the CGI work properly?e  G My browser is receiving intermittent data from BASIC programs which arehD started by CGI in the scripts directory. This usually results in the' following error message in my log file:e   ********** (one line wrapped)eG [Fri Jun 11 08:46:38 2004] [error] [client 142.180.39.24] (00000870)endeB of file: Error string not specified yet: ap_content_length_filter:G apr_bucket _read() failed, referer: http://kawc09.on.bell.ca/scripts/vds sl_event_log_viewerm
 **********  E While researching this problem the only thing that I've found to casti, light on this subject is the following link:B     http://hypermail.linklord.com/new-httpd.old/2002/Aug/0374.html  
 Neil Rieck. Kitchener/Waterloo/Cambridge, Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html  4 ps. Here's my platform info just in case it matters:   AlphaServer-DS20et
 OpenVMS-7.3-1      patches:      DEC AXPVMS VMS731_ACRTL V3.0!     DEC AXPVMS VMS731_UPDATE V2.0      DEC AXPVMS VMS731_PCSI V2.0e
 TCPware-5.6-2.     patches:     drivers_v562p040.a     ssh_v562p032.a   ------------------------------    Date: 11 Jun 2004 10:52:31 -0700# From: CBarry34@hotmail.com (CBarry)t Subject: SYSUAF.DAT = Message-ID: <ed28678f.0406110952.16a4dc13@posting.google.com>3  D Can anyone tell me the easiest way to reset uai_logfail count?  I amD running 6.2 and I have several accounts that need to be reset to 0. : Can this be done in authorize or must a script be written? Appreciate any help.   ------------------------------  % Date: Fri, 11 Jun 2004 21:58:54 -0500u2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: SYSUAF.DAT + Message-ID: <40CA716D.31BCFD53@comcast.net>a  
 CBarry wrote:y > F > Can anyone tell me the easiest way to reset uai_logfail count?  I amE > running 6.2 and I have several accounts that need to be reset to 0.a< > Can this be done in authorize or must a script be written? > Appreciate any help.  G To my knowledge, the only way is to have a successful login. This countrE does not prevent login - that's the result of intrusion detection andaG evasion. If the account does not get automatically DISUSER'd, the loginc3 attempt can be repeated after LGI_BRK_LIM attempts.t  ? See the SYSGEN manual for an explanation of the relevant systemnA parameters. See the Guide to System Security for a description ofn break-in detection and evasion.e   D.J.D.   ------------------------------  + Date: Fri, 11 Jun 2004 19:06:39 +0000 (UTC)e* From: bleau@UMTOF.UMD.EDU (Lawrence Bleau)6 Subject: tcpdump fails miserably, how to make it work?0 Message-ID: <cacvrv$3n1$2@grapevine.wam.umd.edu>  J I was tring to use the tcpdump utility to trace and dumpa tcp session, butN it didn't even start to work.  I'm running VMS 7.3-2, TCPIP 5.4 ECO 1.  I did:   $ @tcpip$define_commands $ sho sym tcpdumpi,   TCPDUMP == "$SYS$SYSTEM:TCPIP$TCPDUMP.EXE"	 $ tcpdumph: %SYSTEM-F-LKWSETFUL, locked portion of working set is full $ help/mess LKWSETFUL   2  LKWSETFUL,  locked portion of working set is full  '   Facility:     SYSTEM, System Services   L   Explanation:  No more pages can be locked in the working set because thereM                 would not be enough remaining pages available dynamically forc0                 the image to continue execution.  K   User Action:  Use the DCL command SET WORKING_SET to increase the workingtI                 set limit, if the limit is not already set at the maximumeK                 allowed. If the working set limit is already at the maximum L                 allowed, correct the program that attempted to lock too many)                 pages in the working set.t
 $ sho workA   Working Set (pagelets)  /Limit=2000  /Quota=4000  /Extent=65536eH   Adjustment enabled      Authorized Quota=4000  Authorized Extent=65536  >   Working Set (8Kb pages) /Limit=125  /Quota=250  /Extent=4096F                           Authorized Quota=250  Authorized Extent=4096 $ set work/lim=40009
 $ sho workA   Working Set (pagelets)  /Limit=4000  /Quota=4000  /Extent=65536 H   Adjustment enabled      Authorized Quota=4000  Authorized Extent=65536  >   Working Set (8Kb pages) /Limit=250  /Quota=250  /Extent=4096F                           Authorized Quota=250  Authorized Extent=4096	 $ tcpdump-: %SYSTEM-F-LKWSETFUL, locked portion of working set is full  5 Clearly, tcpdump is broken.  Does it work for anyone?/   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edue   ------------------------------  # Date: Sat, 12 Jun 2004 01:28:29 GMT.1 From: "Ken Randell" <kenneth.randell@verizon.net>r: Subject: Re: tcpdump fails miserably, how to make it work?7 Message-ID: <1%syc.13028$H65.1663@nwrddc02.gnilink.net>y  8 >> Clearly, tcpdump is broken.  Does it work for anyone?   Yes.   $ show workvB   Working Set (pagelets)  /Limit=8192  /Quota=32768  /Extent=81920I   Adjustment enabled      Authorized Quota=32768  Authorized Extent=81920   ?   Working Set (8Kb pages) /Limit=512  /Quota=2048  /Extent=51203G                           Authorized Quota=2048  Authorized Extent=5120     	 $ tcpdump  tcpdump: listening on WE0d Filtering in user processaI 22:28:37.015398 192.168.1.101.1320 > KAR.23: . ack 83606174 win 8325 (DF) F 22:28:37.015398 KAR.23 > 192.168.1.101.1320: P 1:28(27) ack 0 win 43808 22:28:37.054458 KAR.49442 > verizon1.53: 4916+ (44) (DF)H 22:28:37.075941 verizon1.53 > KAR.49442: 4916 NXDomain* 0/1/0 (104) (DF)C 22:28:37.215580 192.168.1.101.1320 > KAR.23: . ack 28 win 8298 (DF)mI 22:28:37.215580 KAR.23 > 192.168.1.101.1320: P 28:307(279) ack 0 win 4380c (DF)  
 etc., etc.    I think it's your configuration.   ------------------------------  % Date: Fri, 11 Jun 2004 22:06:42 -0500e2 From: David J Dachtera <djesys.nospam@comcast.net>: Subject: Re: tcpdump fails miserably, how to make it work?+ Message-ID: <40CA7342.21CD1070@comcast.net>o   Lawrence Bleau wrote:n > L > I was tring to use the tcpdump utility to trace and dumpa tcp session, butP > it didn't even start to work.  I'm running VMS 7.3-2, TCPIP 5.4 ECO 1.  I did: >  > $ @tcpip$define_commands > $ sho sym tcpdump . >   TCPDUMP == "$SYS$SYSTEM:TCPIP$TCPDUMP.EXE" > $ tcpdumpr< > %SYSTEM-F-LKWSETFUL, locked portion of working set is full > $ help/mess LKWSETFULs > 4 >  LKWSETFUL,  locked portion of working set is full > ) >   Facility:     SYSTEM, System Services  > N >   Explanation:  No more pages can be locked in the working set because thereO >                 would not be enough remaining pages available dynamically forl2 >                 the image to continue execution. > M >   User Action:  Use the DCL command SET WORKING_SET to increase the workinghK >                 set limit, if the limit is not already set at the maximumkM >                 allowed. If the working set limit is already at the maximum N >                 allowed, correct the program that attempted to lock too many+ >                 pages in the working set.  > $ sho workC >   Working Set (pagelets)  /Limit=2000  /Quota=4000  /Extent=65536 J >   Adjustment enabled      Authorized Quota=4000  Authorized Extent=65536 > @ >   Working Set (8Kb pages) /Limit=125  /Quota=250  /Extent=4096H >                           Authorized Quota=250  Authorized Extent=4096 > $ set work/lim=4000i > $ sho workC >   Working Set (pagelets)  /Limit=4000  /Quota=4000  /Extent=65536 J >   Adjustment enabled      Authorized Quota=4000  Authorized Extent=65536 > @ >   Working Set (8Kb pages) /Limit=250  /Quota=250  /Extent=4096H >                           Authorized Quota=250  Authorized Extent=4096 > $ tcpdumpd< > %SYSTEM-F-LKWSETFUL, locked portion of working set is full > 7 > Clearly, tcpdump is broken.  Does it work for anyone?   F I know it may seem outrageous, but try increasing your WSEXTENT beyondF 32MB (65536 pagelets). 64MB (131072 pagelets) might be a good startingD point. Setting WSQUOTA to match wouldn't hurt, either, depending how" memory constrained your system is.  ( Out of curiosity, what's your PGFLQUOTA?   D.J.D.   ------------------------------  % Date: Fri, 11 Jun 2004 19:38:02 -0400e# From: "John Smith" <a@nonymous.com>s< Subject: Tru64 --> OpenVMS rather than another unix platform, Message-ID: <pI6dnbGRj8DH31fdRVn_iw@igs.net>  F I was flipping though the HP User Advocacy newsletter of April 2004  (? http://www.hpuseradvocacy.org/news/ ) and came across this gem:a   Migrating from Tru64 UNIX5  J Encompass, an independent HP user community, helps members keep up-to-dateI with the latest developments in the IT world. Encompass provides an open,mK objective knowledge exchange network and vendor advocacy. Encompass membersiJ benefit from a wealth of information resources, as well as educational and training opportunities.a  K Recently, the Encompass Enterprise UNIX Special Interest Group collaborated>J with MindIQ to present a free teleconference/webcast. Paul Manno of MindIQK discussed the transition path from Tru64 UNIX operating systems to Linux orrE HP-UX. Participants were able to learn about OS platform, file systemiE options and organization, kernel tuning, device support and migrationm
 resources.  D Due to the sold out success of the webcast, Encompass and MindIQ areK planning to partner on additional free-to-member webcasts in the future. To I learn more about how to take advantage of this opportunity and many other & Encompass member benefits, click here.      H Since people are being forced to migrate from Tru64, why doesn't HP evenI suggest OpenVMS as an alternative? It's still unix-y enough to be brandedsL unix, 'cept of course it doesn't have nearly the same number of dbms choicesK or other software choices that unix platforms have, and it doesn't have theuG same support at HP that Windows & Linux - two operating systems that HPp makes no money on - have.    ------------------------------  % Date: Sat, 12 Jun 2004 02:01:34 +0800o, From: Paul Repacholi <prep@prep.synonet.com>! Subject: Re: What might have beenn- Message-ID: <87659ydldt.fsf@prep.synonet.com>r  % "John Smith" <a@nonymous.com> writes:e  E > Your final exam assignment for today is to contrast and compare the A > foregoing with the state of the ISV market and applications fore
 > OpenVMS.  B > You have 5 minutes to complete your answer. Quoting HP officials! > will result in a failing grade.u  ; Quoting a relevent positive HP official will get you a  A+.d   (in fairy tales) -- s< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.e@                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Fri, 11 Jun 2004 18:18:07 -0500y5 From: Starshine Moonbeam <silverbells@tacoshells.com>e= Subject: Re: Where have all the rec.travel.air regulars gone?n5 Message-ID: <MPG.1b33f722f2de026098b542@news.alt.net>   B In article <ca9s5n$c8b$2@blackhelicopter.databasix.com>, "Gary L. @ Burnore"(gburnore@databasix.com) dropped a +5 bundle of words...  8 > On Thu, 10 Jun 2004 00:29:57 -0500, Starshine Moonbeam% > <silverbells@tacoshells.com> wrote:  > E > >In article <ca8p80$q86$3@blackhelicopter.databasix.com>, "Gary L.  C > >Burnore"(gburnore@databasix.com) dropped a +5 bundle of words...n > > < > >> On Wed,  9 Jun 2004 18:20:02 +0200 (CEST), Nomen Nescio > >> <nobody@dizum.com> wrote: > >> d > >> >5 > >> >Gary L. Burnore <gburnore@databasix.com> wrote:  > >> >> > >> >>On Wed,  9 Jun 2004 05:00:01 +0200 (CEST), Nomen Nescio  > >> >><nobody@dizum.com> wrote: > >> >>  > >> >>>8 > >> >>>Gary L. Burnore <gburnore@databasix.com> writes: > >> >>>O > >> >>>>Still can't get that remailer thing right, eh coward.  Didja know yer'AM > >> >>>>over BI=20 with these?  Hehehe. So you're a coward AND a net abuser.C( > >> >>>>No wonder you post anonymously. > >> >>> > >> >>>What a pat[SPLAT]e > >> >>e > >> >>Poor [FLUSH]e > >> >G > >> >Poor convicted sex offender coward.  No matter how many times youtI > >> >post, people still see that you're a convicted sex offender coward. J > >> >See cowardly convicted sex molester molest.  Coward molester molests> > >> >often.  Molest, cowardly convicted sex molester, molest! > ><& > >> Getting to you, aren't I, coward? > >:% > >I'm thinking maybe just a little. s > @ > What's really funny is that we had a planned outage to upgradeH > circuits and several in an internal group talked about how much fun it@ > would be to use as a troll tactic.   We were going to announceG > something like 'going off the air' but we didn't even need to do THATd > to hook a couple of loons.  , They just jump in the boat. No hooks needed.    f   -- . Starshine Moonbeam   ------------------------------  + Date: Fri, 11 Jun 2004 18:56:52 +0000 (UTC)a* From: bleau@UMTOF.UMD.EDU (Lawrence Bleau); Subject: Re: Why does idle telnet session get disconnected?g0 Message-ID: <cacv9k$3n1$1@grapevine.wam.umd.edu>  E Okay, I posted a theory yesterday, and it failed miserably.  First, a 8 recap: an idle telnet sessions sometimes is disconnected  = VMS 7.3-2  TCPIP 5.4   session gets disconencted after ~2 hrsu4 VMS 7.1-2  UCX   4.2   session is never disconencted  G My theory was that the UCX 4.2 system never sent out keepalive packets,SF didn't even implement that feature, and therefore never disconnected. G Well, that's dead wrong.  A coworker did a dump of his tcp session, andtJ detected a small packet being sent every 75 secs, along with an ack.  This; exactly matches what we'd expect for the keepalive feature.l  I However, he did *not* detect any such packets from the TCPIP 5.4 system! iD This is definitely a clue.  He left his session up for hours, and is7 supposed to get such a packet after 2 hrs, but did not.s  J Could it be that keepalive behavior, when changed for 5.4, was implementedE incorrectly?  Maybe it has a bug in which it *thinks* it sent out thelI keepalive packet, but never did.  Then, when waiting for the response, iteH of course never gets one, and subsequently disconnects the presumed dead telnet session.    Here's the questions:s  $ Has anyone else been bitten by this?9 Does anyone have evidence that the keepalive logic works? K Has anyone done a dump of the tcp session to see what is being transferred?nB Could the above theory be correct?  Is there a way to disprove it?   Thanks, all.   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edu0   ------------------------------  % Date: Fri, 11 Jun 2004 19:22:23 -0400c# From: "John Smith" <a@nonymous.com> 5 Subject: Re: [OT] An interesting view of Linux's costi, Message-ID: <wLidnbp2AZAyo1fd4p2dnA@igs.net>   Karsten Nyblad wrote:39 > "Paul Sture" <nospam@sture.homeip.net> wrote in message % > news:2g991vF3559U1@uni-berlin.de...a >> Hans Vlems wrote:= >>> "Paul Sture" <nospam@sture.homeip.net> schreef in berichtn7 >>> news:c687mi$8pn7e$1@ID-132135.news.uni-berlin.de...T >>	 >> [snip]o >> >>>>E >>>> I believe you, but the point Hans and I are making is that there B >>>> are still pitfalls, upgrades can break stuff, and if we as ITF >>>> professionals find it hard work, what chance do Joe and Jane Home >>>> User stand? >>>> >>>tC >>> Glad you brought that up again Paul, because that's exactly thewG >>> issue we were discussing. It is quite possible to install a product D >>> on a linux platform and get it up and running. The two points we >>> were making were:s >>>tG >>> 1-It is yet quite another thing to keep it running after an upgrades9 >>> 2-Can a person untrained in IT lore do the same thingo >>>h= >>> The conclusion is that a linux o/s upgrade is not upwardsiG >>> compatible and that you need some knowledge about the os to make iti	 >>> work.e >>>n >>F >> I know we have already thrashed this one out, but here's an article >> by a seasoned Linux user: >>& >> "The Truth About the Linux Desktop" >>: >> http://www.tyma.com/modules/news/article.php?storyid=19 >>4 >> and he's pretty disparaging about OpenSource too: >>) >> "Why does Open Source Software Exist?"s >>: >> http://www.tyma.com/modules/news/article.php?storyid=20 >>G > These articles really show how some journalist are trolling when theysG > do not have anything to write.  In particularly the second article is E > written without checking research on the subject.  It is well-knownlB > that the typical open source writer is middle-aged, e.g., in hisE > thirties or forties.  It is also well-known that people are writingmB > open source software for many reasons.  Examples:  1) People mayE > write software to get new qualification, either to get a job, get avD > better job, or to stay qualified for their current job.  2) PeopleF > may write software because they think it is funny to write.  3) SomeF > write it because they are paid, e.g., employees at IBM, SUN, Redhat,B > SUSE, etc.  4) Yes, some may write it because they want their 15/ > minutes of fame, and what is wrong with that.o >wE > Yes, you can get money for write software, but if you can get moneytB > for writing compiler, database systems, drivers, kernels, or anyF > other sort of code that is actually fun to write then you are one ofD > the lucky few.  Why are people painting pictures for their leisureC > when they could get money painting other peoples houses?  Why arey> > people working for charities when they could work for money? >hF > When people work for free they usually do it for many reasons and noG > one can say that on reasons is better or more honorable than another.sC > Simply accept that people are working for free for the benefit ofeE > mankind, whether they work for charity or open source.  The subjecta= > on why people write open source software really is not thatcE > important.  The important thing is that there will continue to be a C > large body of people doing it.  That ensures that if you chose an17 > open source solution then it will also be maintained.r >KD > About the first article.  Yes, you may have to work some more whenE > you have a Linux desktop.  However, I think there are not that manycE > that think Linux on the desktop is competitive unless you have manynA > people doing the same task.  There is a reason why Redhat's neweE > desktop license is not sold in quantities less than 50.  People are D > starting to chance to Linux when they have hundreds of using their@ > desktop for the same task.  Then you can make a business case,> > because your saving on licenses are in thousands of dollars. >gA > Second, you cannot say anything about stability of an operatingcF > system from one user.  He may have an usual combination of hardware,E > he may have configured his laptop in a wrong way, or he may use hislG > laptop for usual tasks.  Please remember that there has also been VMSa+ > users whose machines crashes once a week.u  G Often in that case it it bad memory, thermal problem, or other hardwarexK fault. It isn't usually the operating ssytem. User code can do nasty thingsg but usually only to themselves.    ------------------------------    Date: 11 Jun 2004 21:06:18 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman) Subject: [OT] Windows TNTR= Message-ID: <b096a4ee.0406112006.637cbe3d@posting.google.com>m  8 http://www.cartalk.com/content/tps_reports/boomboom.html  B Enjoy.  This is one of the pages that might come up when you press their "boss button".   ------------------------------   End of INFO-VAX 2004.324 ************************ disk, and the backup/image from   > scratch back to original.  > J > I vaguely remember reading that when doing an image backup from disk to A > disk the verify qualifier isn't really necessary. Is that true?   B It depends on how important it is to be sure there were no errors.  9 > It'd save me some time, 