sdroppers@pbs.uucp (Seton Droppers) (02/08/90)
In article <6098.25cd391d@pbs.uucp>, sdroppers@pbs.uucp (Seton Droppers) writes: > We are considering upgrading from VMS 5.1-1 to either 5.2 or 5.3 > (we have both distributions now). My manager, however, wants to > know what others have found after the "problems" we had upgrading > to VMS 5.0. > > We run a cluster of 2 8650s and 2 785s, along with 6 Giga-bytes > of disk storage arranged with RA82s, RA81s, and RA60s. When we > upgraded from VMS 4.7 to VMS 5.1-1 we lost capacity for about 10 > users on each 8650 (from about 95 to about 85). This was lost > through both increased memory usage and increased system > overhead, our paging went up a bit. This did not surprise me a > lot and fits with what most of the rest of you found as well. My > manager, however, felt very bad and wonders what problems are > lurking behind the curtain of VMS 5.2/5.3. He would very much > like to hear what experiences others have had with a similar > environment when you upgrade to VMS 5.2 or 5.3. He is also > interested in hearing what others have to say, 5.2 or 5.3? So far I have gotten two responses to this, which was originally posted only in comp.os.vms (and thus to INFOVAX): Pete Kaiser, of DEC, who runs an LAVC cluster called me to say he has been running VMS 5.3 since November 1989 with no problems. The upgrade went very smoothly, with the only know patch required being a secruity patch. He suggested calling Colorado to check on the security patch. He noted no change in performance. Another plus with going to 5.3 was the support for the latest release of DECwindows. Alan Weitzsacker (CANSYS%CANISIUS.BITNET%CORNELLC.cit.cornell.edu) mailed a note saying the problems he is aware of have been with the new version of BACKUP. He suggested following up on the BACKUP issue before going ahead with the installation of VMS 5.3. I am still looking for any other comments. Our department director is using this and another item as a trial baloon for using usenet, so don't hesitate to flood us with responses. He is also looking to talk with some one that has a site similar to ours (2x8650, 2x785, 6Gig disk, All-In-1 and Mixed "Time Share") that has been able to upgrade to 5.3, so let me know if it is OK for him to call (he may only want to call one or two sites, but let me know anyway). Please E-Mail your comments to me. I will attempt to summarize the responses again. Thank you very much for your time and effort. If you have tried to mail to us and it has bounced, sorry. We are still a pure UUCP site and that sometimes causes trouble. Seton R. Droppers, Systems Manager Public Broadcasting Service, Computer Services UUCP: ...{vrdxhq,csed-1,IDA.ORG}!pbs!sdroppers Voice: 703/739-5100 USMail: 1320 Braddock Place Alexandria, VA 22314-1698
sdroppers@pbs.uucp (Seton Droppers) (02/15/90)
In article <6098.25cd391d@pbs.uucp>, sdroppers@pbs.uucp (Seton Droppers) writes: > We are considering upgrading from VMS 5.1-1 to either 5.2 or 5.3 > (we have both distributions now). My manager, however, wants to > know what others have found after the "problems" we had upgrading > to VMS 5.0. > > We run a cluster of 2 8650s and 2 785s, along with 6 Giga-bytes > of disk storage arranged with RA82s, RA81s, and RA60s. When we > upgraded from VMS 4.7 to VMS 5.1-1 we lost capacity for about 10 > users on each 8650 (from about 95 to about 85). This was lost > through both increased memory usage and increased system > overhead, our paging went up a bit. This did not surprise me a > lot and fits with what most of the rest of you found as well. My > manager, however, felt very bad and wonders what problems are > lurking behind the curtain of VMS 5.2/5.3. He would very much > like to hear what experiences others have had with a similar > environment when you upgrade to VMS 5.2 or 5.3. He is also > interested in hearing what others have to say, 5.2 or 5.3? I have recieved a total of six responses, one of which included a compendim of notes on LAVC issues and VMS 5.2/5.3. Most comments had to do with Backup, LAVCs, or DECwindows. The only responses about performance noted that performance did *not* suffer going from 5.1 to either 5.2 or 5.3. Pete Kaiser of DEC called to say he has been running VMS 5.3 since November 1989 with no problems. The upgrade went very smoothly, with the only known patch required being a security patch. He noted not change in performance. Alan Weitzsacker and others passed notes along saying to read the VMS 5.2 release notes on BACKUP carefully since there can be problems if the parameters for the account BACKUP is run out of are not set up correctly. It appeared that there was some mention of other problems with TU81 tape drives, but nothing about TU78/TU79 drives, so I have assumed there are not problems with TU78/TU79 drives. Several comments were made about LAVCs, most mentioning that it is important to make sure all Q-Bus machines are up to current hardware rev. levels.Since my site does not make use of LAVCs I have not followed that to any depth. A comment was made that Q-Bus machines should change from DEQNA to DELQA ethernet cards on Q-Bus machines as needed. This was echoed by personal talks I had with both a VMS developer and Educational Specialist -- They both said "Change those boards as *soon* as you can!". Several comments were also made that DECwindows is much improved with the 5.2/5.3 releases of VMS. Again, since we do not use DECwindows I did not follow them to any great depth. General comments I have gathered from reading through all of the responses: Latest release: VMS5.3-1 Patches: Security Patch for VMS 5.3 -- Check with Software Support XQDRIVER.EXE -- CSC Patch #125 was recomended Please contact me if any of this information differs from what you have experienced. And thank you all for your time and effort: Pete Kaiser at DEC Alan Weitzsacker -- CANSYS%CANISIUS.BITNET%CORNELLC.cit.cornell.edu Jeanie Larson -- lll.llnl.gov!JLARSON Esther Lumsdon -- dtoa1.dt.navy.mil!lumsdon Todd Aven -- usgcdh!manager@eddie.mit.edu Steven Tihor -- tihor@acf4.NYU.EDU (Addresses are my best guess at lifting from paths and from lines...) P.S. I have copies of all responses, please E-Mail me if you want specific information. Seton -- Seton Droppers -- "Anything that I say is my opinion and not my employer's." Public Broadcasting Service, 1320 Braddock Pl. Alexandria, VA 22314 (UUCP: ...{vrdxhq,csed-1,ida.org}!pbs!sdroppers) (Voice: 703/739-5100) (VAX/VMS running DECUS UUCP 1.1, ANU News 5.9C)