carlton@apollo.UUCP (Carlton B. Hommel) (09/27/87)
[I received this message three times. I guess Carlton was frustrated getting it here. There was a uucp logjam at Rice. It is fixed now. -- sob] I believe that all of the posted problems with the High Performance Serial Boards (HPSMs) have been fixed in the RT-HPSM V4.0 release. I was just the project leader and release engineer, not the technical bug-killer, so please don't send me technical questions about DTR, DSR, RTS, or PDQ. Lets face it, the original HPSM releases were dogs. While the worked better than the IMUXes and JMUXes the replaced at handling ascii terminals, they were terribly buggy when it came to all the fancy uses that customers use to send input into the computer. The RT-HPSM V3.0 release never was distributed, because of a clash with RTU V3.1. (Don't ask, you don't want to know. Really, you don't.) Under intense pressure from several OEMs who threatened to walk away and buy Suns or (dare I say it?) Apollos, an emergency one-month project was put together to make the HPSM work. *ALL* of the outstanding SQR's were closed. Most were found to be due to errors in driver code, and fixed. The others were not reproducible with the new RTU software. (At one point, we had a 5700 with 8 hpsm boards feeding 64 tty lines, feeding into four 5400s with 2 hpsms each, running all week. Its amazing how much respect your testing group gets when OEMs wave dollar signs about.) We spent one week coding, one week testing, one week flogging the OEMS to install the kit, and one week waiting for responses. Everyone was happy. Now, if you *have* installed RT-HPSM V4.0, and have these wierd hangs, undropped lines, etc. let MASSCOMP know. Insist that the person you talk to on the phone actually lets a manager in Engineering know, so they can consider getting the fix into their next release. If you haven't installed V4.0, get it! Its only one floppy, and 5-15 minutes downtime. (5300's take 15 minutes, 5700s take 5 or less.) Carl Hommel apollo!c!)2DaGu
masscomp-request@soma.UUCP (10/01/87)
In article <3226@soma.bcm.tmc.edu> carlton@apollo.UUCP (Carlton B. Hommel) writes: >I believe that all of the posted problems with the High Performance >Serial Boards (HPSMs) have been fixed in the RT-HPSM V4.0 release. >I was just the project leader and release engineer, not the technical >bug-killer, so please don't send me technical questions about DTR, >DSR, RTS, or PDQ. Lets face it, the original HPSM releases were dogs. While the current release (4.0) is better, we still require frequent reboots because of hung lines. We're working with Masscomp on it locally, but still no joy. Since we talk to the Masscomp through an Equinox data PBX, the RS-232 signals get more exercise than they do at most installations. Still, it's extremely annoying to have a system that must be rebooted four or five times a week. Do I sound annoyed? I'm annoyed. >Under intense pressure from several OEMs who threatened to walk away and >buy Suns or (dare I say it?) Apollos, an emergency one-month project was >put together to make the HPSM work. *ALL* of the outstanding SQR's were >closed. Most were found to be due to errors in driver code, and fixed. >The others were not reproducible with the new RTU software. (At one point, >we had a 5700 with 8 hpsm boards feeding 64 tty lines, feeding into four >5400s with 2 hpsms each, running all week. Its amazing how much respect >your testing group gets when OEMs wave dollar signs about.) If the lines were modems that drop carrier rather than just wired TTY lines, I seriously doubt that you would have passed the test. >Everyone was happy. We aren't. >Now, if you *have* installed RT-HPSM V4.0, and have these wierd hangs, >undropped lines, etc. let MASSCOMP know. Insist that the person you talk >to on the phone actually lets a manager in Engineering know, so they can >consider getting the fix into their next release. Can you tell me who to talk to? No one in San Jose has had any luck or seems to know how to attack the problem. >If you haven't installed V4.0, get it! Its only one floppy, and 5-15 >minutes downtime. (5300's take 15 minutes, 5700s take 5 or less.) It's certainly better than what you had before. -- - Joe Buck {uunet,ucbvax,sun,decwrl,<smart-site>}!epimass.epi.com!jbuck Old internet mailers: jbuck%epimass.epi.com@uunet.uu.net