INFO-VAX Mon, 16 Jul 2007 Volume 2007 : Issue 385 Contents: Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Re: cURL 7.16.3 available Re: EDT Replacement Re: HoffmanLabs: High-performance Interconnects, MPI? Re: Internet2 Shibboleth on OpenVMS Re: Linux (was Re: How many people here use Itanium w VMS) ---------------------------------------------------------------------- Date: Sun, 15 Jul 2007 13:44:31 -0700 From: Bob Gezelter Subject: Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Message-ID: <1184532271.908646.316560@r34g2000hsd.googlegroups.com> On Jul 15, 3:28 pm, hel...@astro.multiCLOTHESvax.de (Phillip Helbig--- remove CLOTHES to reply) wrote: > I have a batch job which crashed due to this error. It had been running > for 15 days. (It is a never-ending "loop" job.) I has run much longer > many times. Once the process is started, nothing really changes, so > what actually causes the error? > > Parameter Name Current Default Minimum Maximum Unit Dynamic > -------------- ------- ------- ------- ------- ---- ------- > CLISYMTBL 1024 512 20 1024 Pages D > > Thus, no room for improvement here! Phillip, It would be interesting to look at the output of various SHOW SYMBOL commands at various points in the processing. Could something in the "loop" be adding symbols on occasion? - Bob Gezelter, http://www.rlgsc.com ------------------------------ Date: Sun, 15 Jul 2007 22:58:30 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Message-ID: In article , helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: > > >I have a batch job which crashed due to this error. It had been running >for 15 days. (It is a never-ending "loop" job.) I has run much longer >many times. Once the process is started, nothing really changes, so >what actually causes the error? > >Parameter Name Current Default Minimum Maximum Unit Dynamic >-------------- ------- ------- ------- ------- ---- ------- >CLISYMTBL 1024 512 20 1024 Pages D > >Thus, no room for improvement here! Phillip, One of the features I put in my SYMBOL utility was the capability to examine the pool region from which DCL allocates space for symbols. I'd suggest you use it to see what it going on. I would wager that after many days of creating and deleting DCL symbols, the pool is too fragmented to allocate enough space to create some symbol and it is this that is causing you to get the: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols error message. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" ------------------------------ Date: Sun, 15 Jul 2007 23:30:55 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols Message-ID: In article , VAXman- @SendSpamHere.ORG writes: > > >In article , helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: >> >> >>I have a batch job which crashed due to this error. It had been running >>for 15 days. (It is a never-ending "loop" job.) I has run much longer >>many times. Once the process is started, nothing really changes, so >>what actually causes the error? >> >>Parameter Name Current Default Minimum Maximum Unit Dynamic >>-------------- ------- ------- ------- ------- ---- ------- >>CLISYMTBL 1024 512 20 1024 Pages D >> >>Thus, no room for improvement here! > >Phillip, > >One of the features I put in my SYMBOL utility was the capability to examine >the pool region from which DCL allocates space for symbols. I'd suggest you >use it to see what it going on. > >I would wager that after many days of creating and deleting DCL symbols, the >pool is too fragmented to allocate enough space to create some symbol and it >is this that is causing you to get the: > > %DCL-W-SYMOVF, no room for symbol definitions - delete some symbols > >error message. I should have included a sample output to show the pool statistics. Here it is: $ SYMBOL/SHOW=(NOSYMBOLS,NOPROCEDURES,POOL_STATISTICS) ------------------------------------------------------------ Maximum Symbol Table Size: 385024(dec) / 05E000(hex) Bytes Amount in Use for Symbols: 23520(dec) / 005BE0(hex) Bytes Total Available Free Pool: 352520(dec) / 056108(hex) Bytes Currently Available Free Pool Found in 10 Pool Fragments Free (Pool Fragment) Size: 456(dec) / 0001C8(hex) Bytes Free (Pool Fragment) Size: 8(dec) / 000008(hex) Bytes Free (Pool Fragment) Size: 40(dec) / 000028(hex) Bytes Free (Pool Fragment) Size: 264(dec) / 000108(hex) Bytes Free (Pool Fragment) Size: 24(dec) / 000018(hex) Bytes Free (Pool Fragment) Size: 16(dec) / 000010(hex) Bytes Free (Pool Fragment) Size: 16(dec) / 000010(hex) Bytes Free (Pool Fragment) Size: 8(dec) / 000008(hex) Bytes Free (Pool Fragment) Size: 16(dec) / 000010(hex) Bytes Free (Pool Fragment) Size: 351672(dec) / 055DB8(hex) Bytes ------------------------------------------------------------ In the above, there are several regions in the pool which are insufficiently sized to even hold the supporting structure for the symbol. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" ------------------------------ Date: Sun, 15 Jul 2007 22:58:57 +0200 From: =?ISO-8859-1?Q?Jean-Fran=E7ois_Pi=E9ronne?= Subject: Re: cURL 7.16.3 available Message-ID: <469a8a92$0$23747$426a74cc@news.free.fr> Arne Vajhøj wrote: > Jean-François Piéronne wrote: >> is there any interest to include pycURL in Python for VMS? >> I remember it was request once. >> >> From http://pycurl.sourceforge.net/ >> """ >> PycURL is a Python interface to libcurl. PycURL can be used to fetch >> objects identified by a URL from a Python program, similar to the urllib >> Python module. PycURL is mature, very fast, and supports a lot of >> features. >> """ > > As a general rule: the more freeware available for VMS the better ! > :-) Ok, next Python kit will include pycurl. It will also include lxml (http://codespeak.net/lxml/) and a module which will expose the smg$ routines. JF ------------------------------ Date: Sun, 15 Jul 2007 20:47:00 +0200 From: "P. Sture" Subject: Re: EDT Replacement Message-ID: In article , "R.A.Omond" wrote: > David J Dachtera wrote: > > "P. Sture" wrote: > >>[snip] > >>While we are here, I just found an old message here of mine which claims > >>that EDT can't handle files with > 65K records. This could be the OP's > >>problem. > > > > Well, RMS would choke long before that. So, the point may be moot. > > Eh? I think you're parsing the "> 65K records" the wrong way. > Paul is meaning a file with more than 65,536 *records*, not a > file with records that are greater than 65,536 bytes long. You beat me to that; indeed what I meant, although having just tried to reproduce it, I see that V8.3 can cope with at least 200,000 records. -- Paul Sture ------------------------------ Date: 16 Jul 2007 00:30:20 GMT From: bill@triangle.cs.uofs.edu (Bill Gunshannon) Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <5fvsgsF3ebi79U1@mid.individual.net> In article <000f01c7c693$89f649e0$9de2dda0$@com>, "Paul Raulerson" writes: > Bill- you talk through your nose without facts. Try Baylor and the > University of Texas. > The particular interviewee I was speaking of is perhaps an a little extreme > compared to the norm, > but not that much, and his degree was awarded by U.T. UT where? Not all UY CS programs are accredited. Byalor, however is. So I would be interested in the credentials of someone with a BSCS from there that was incompetent. As would ABET, I imagine. > > I daresay that either of those schools is more than the equal of Scranton. I > know both of them turn out some top flight scientists, physicists, and > mathematicians. I *know* that because I *know*, personally, some of those > selfsame people. Wether one gets to be a "top flight scientist, physicist, or mathematician" is frequently based more on the name of the school than on the strength of its program. Thus the reason it was decided that CS (and now, CIS) now need accreditation. Berkeley turned out some really fine Computer Scientists, but not everyone who went through the program ended out with the same stature. It is the program you have to look at, and not the name of the school. > > Further, the school makes very little difference; it is the attitude and > ability of the professor or instructor that really count. What do you think ABET looks at when it accredits a program? > I will again point > out that schools in other countries emphasize the areas that the schools > here tend to deride or find "uninteresting", with the result their students > are better qualified, and better motivated, and WORK HARDER than the kids > unfortunate enough to get the self destructive instruction offered here. If that's the case, why do we always have such a high percentage of those foreign students you think so highly of? They must not think musch of their own schools if they travel halfway around the world to attend school here. I keep hearing about how muc better everything is in foreign copuntries. better schools, better medicine, better living conditions. And yet, everyone is trying to come here and even people who hate it here won't leave. Go figure. > > I did not point out, but will, that some kids coming out of those schools > really do have it together, > but there is a cadre of bull headed mulish stubborn people who insist > everyone who does not agree with them > is a moron or had not possibly thought through the problem. We have people like that right here. Why would you expect a cross-section of college professors to be any different. > Unfortunately, > the kids get exposed to this and it sticks. My buddies who work in Mental > Health tell me is similar to the effect an abusive parent has on a sensitive > and impressionable young child. Debilitating. > > Should sound familiar to you, judging by your public statements here. You must have had a really bad time in college if you dislike academia so much. Too bad, cause it probably means you are missing a lot of really bright people when your prejudices get in the way. One thing you do need to remember, some of the best known colleges in the country are known more for their football teams than their academics. Name means little when a school is living on past glory. Again, another of the reasons why accreditation has become important. But I doubt anything I say will convince you. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Mon, 16 Jul 2007 12:32:13 +0930 From: Mark Daniel Subject: Re: Internet2 Shibboleth on OpenVMS Message-ID: <139lnu95pqr1v08@corp.supernews.com> Main, Kerry wrote: >>-----Original Message----- >>From: Mark Daniel [mailto:mark.daniel@vsm.com.au] >>Sent: July 15, 2007 9:21 AM >>To: Info-VAX@Mvb.Saic.Com >>Subject: Internet2 Shibboleth on OpenVMS >> >>Anyone tinkering with this? >> >> http://shibboleth.internet2.edu/ >> >>(And yes, I remember it took a decade to get Kerberos onto VMS :-) Thanks for the extensive reply Kerry. In my brevity and reference to Kerberos I may have unintentionally given the impression I was asking about this wrt VMS Engineering support. My apologies. I was merely inquiring in a forum I know a colleague proposing to undertake this port (to a greater or less extent) would not use, perhaps saving him some effort. This is definitely a niche application, especially for the VMS community. If anything does come of it I'll feedback here. > Mark, > > As I suspect you know :-), there have been numerous attempts at providing multi-vendor, multi-platform distributed security solutions over the years - each with varying degrees of success. Web Service standards and offerings continue to struggle big time with this as well. > > I could not agree more with the driver to do this. In fact, thx for the pointer as I will read up more on this. > > However, those involved in projects like this need to have some background in order that they not repeat the many mistakes made in the past for these types of distributed security initiatives. A review of why DCE and other initiatives like Microsoft's public program failed (I forget their code name) are likely in order. Why did Kerberos not gain wider acceptance for example. > > There are also some very good commercial solutions with dedicated teams of SW engineers working on them that would likely be considered as the competition. (Yes, cheap is not a term for these solutions). > > At the US University I was at doing a DC and server optimization review, we quit counting after about 10 in terms of the different user authentication DB's in use at the University (HR out of syc with Banner out of sync with Parking out of sync with facilities out of sync with individual Colleges using different AD's like AFS, Microsoft AD etc). > > They were working to resolve it, but it was a real tough battle for them because the real issues with distributed common solutions in general are not technical, but rather data ownership issues. > > You get the same situation in most med to large companies today. > > This solution seems to take a federated approach, so it does sound promising, but will need to review this some more. > > Bottom line is that distributed security solutions are just really, really tough to do because security is people + process + technology and the technology is the one that typically gets blamed when people or process issues fail. > > Many Cust's have totally given up on SSO (single sign on) as a theoretical concept that will never really be reached in most environments. Hence, the new buzz word often touted about is RSO (reduced sign-on) with a consolidation of auth db's, but recognize there will continue to be some App's etc that will continue to use their own DB or repository for this. > > > Regards > > > Kerry Main > Senior Consultant > HP Services Canada > Voice: 613-592-4660 > Fax: 613-591-4477 > kerryDOTmainAThpDOTcom > (remove the DOT's and AT) > > OpenVMS - the secure, multi-site OS that just works. > > > ------------------------------ Date: Sun, 15 Jul 2007 20:43:39 -0500 From: Ron Johnson Subject: Re: Linux (was Re: How many people here use Itanium w VMS) Message-ID: On 07/15/07 18:56, JF Mezei wrote: > healyzh@aracnet.com wrote: >> Why not? We have quite a few at work that were purchased >> specifically to run Linux. > > How much software is available for Linux on that IA64 thing ? > Wouldn't running Linux on a non-mainstream platform make it > harder to get all the software you want ? > > I can understand building a one specifc application, (such as > calculating PI to the infinity) which doesn't depend on a whole > slew of middleware, and thus would not really matter on which > platform it is built on. > > But when you build more modern applications, don't they mostly > depend on a lot of middleware ? Do you have to build that > middleware yourself from sources on that IA64 thing, or are there > vendors such as RedHat, Suse etc which do provide pre=built easy > to install middleware for IA64 ? > > If you replace IA64 with Alpha, do the answers to the above > questions change ? What about Power ? > > Or is the 8086 the only mainstream architecture for which the > applications are easily available and for the others, you need to > build the apps yourself ? No. Just about *every* open source app that runs in 32-bit Linux also runs in the PPC64, Alpha, x86-64 & IA64 Linux. The only apps that don't run, or that run at reduced functionality are multimedia apps that have wrappers around (binary-only, of course) win32 codec code. See, Linux Alpha has been around since 1994/95 and Debian builds a 64-bit current version of the whole distro for Alpha. NetBSD does something similar. By the time that IA64 & AMD64 arrived on the scene, years and years of bug reports sent in by Alpha users made a lot of the code 64-bit clean, and the popularity of AMD64 has generated bug reports on any libs, middleware, apps, etc that weren't yet 64-bit clean. And most code builds on FreeBSD, too. > Would it be fair to say that Linux on IA64 has a greater breath > of pre-built applications than Windows on IA64 ? or are they > roughly the same ? Most probably. But if Windows coders had been following MSFT guidelines, most programs *should* work by simply re-targeting the compiler output. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ End of INFO-VAX 2007.385 ************************