preece@ccvaxa.UUCP (08/21/85)
According to Electronics, DEC has stopped shipping microVax IIs with Ultrix because of an unspecified problem with the operating system. Are there any of you out there who know what the problem is? The article was interesting on a number of fronts. Apparently a large factor in pushing Ultrix out the door at first ship was a large GM RFQ for Unix-based workstations. According to the article, the mVax II was disqualified on that bid because it was not "Unix-based" but was optimized for VMS and had Ultrix grafted on. This seemed somewhat curious, since Apollo was one of the qualifying bidders, and their Unix is also an alternative, rather than principal, operating system offering (though Aegis may be closer to Unix than VMS). I wonder how GM defined "Unix-based." -- scott preece gould/csd - urbana ihnp4!uiucdcs!ccvaxa!preece [Disclaimer: it's not impossible that Gould bid on that GM contract; if so, this should not be construed as a request for information that bidders shouldn't be allowed to hear or as carping at the competition, but rather as the idle curiosity of a Unix person in no way representing himself as an official manifestation of Gould Electronics Computer Systems Division...]
gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (08/23/85)
> According to Electronics, DEC has stopped shipping microVax IIs > with Ultrix because of an unspecified problem with the operating > system. I don't know if this is related, but I heard today that the latest release of Ultrix-32 (with DECnet support) has totally broken TCP/IP. This is second-hand info, so if it matters to you, check for yourself. Maybe someone from DEC could comment.
avie@cmu-cs-wb1.ARPA (Avadis Tevanian) (08/23/85)
>According to Electronics, DEC has stopped shipping microVax IIs >with Ultrix because of an unspecified problem with the operating >system. Are there any of you out there who know what the problem >is? As far as I know (and I may be completely wrong), there is no problem with Ultrix. The problem is the hardware. It seems that DEC has experienced some memory problems. As it turns out, the "bad" addresses usually show up in the lower ranges... right where Unix lives. Apparently, VMS runs in high memory, so the problem only affects poor little user processes (a situation that they are willing to live with). Supposedly, a significant number of beta sites have reported this problem. Our MicroVax II hasn't had this problem, though. In fact, in several months use, I've seen only one crash caused by hardware (a machine check).
sdyer@bbncc5.UUCP (Steve Dyer) (08/23/85)
> I don't know if this is related, but I heard today that the > latest release of Ultrix-32 (with DECnet support) has totally > broken TCP/IP. This is second-hand info, so if it matters to > you, check for yourself. Maybe someone from DEC could comment. Here at BBN, we actually returned a LH/DH-11 to ACC on the belief that it was broken--this was a new machine supporting an 1822 Arpanet interface and running Ultrix 1.1 (1.1 being needed because of DMZ-32 hardware on this machine.) Naturally, we foolishly didn't suspect the software--I mean, DEC doesn't even officially support the non-DEC peripherals, so one wouldn't expect the IMP software to change. Ultrix 1.0 (sans DECnet) works fine. Turns out that the changes needed for DECnet did indeed "break" 1822 support (and maybe other devices--I haven't tried anything but DEUNA and 1822)--TCP/IP over the DEUNA works fine. We spoke to Ultrix engineering, and volunteered to help them fix this problem, since we knew that if we wanted our machines up and running quickly, this would be the only way to ensure that, since DEC does not have an Arpanet-style connection to any of its engineering VAXes. We have both running now, under a quick-and-dirty fix, and should have a DECnet-compatible fix which can be tested at DEC in a few more days. I can't comment on when it will be available from DEC. I might mention that DEC has been extremely responsive to this problem-- we've had the pleasure to speak with people who really know their stuff. I think their Ultrix software support system and the communication between support and engineering has improved by several orders of magnitude from our earlier experiences of a year ago, to the extent that I am quite happy to recommend their support service. -- /Steve Dyer {harvard,seismo}!bbnccv!bbncc5!sdyer sdyer@bbncc5.ARPA