home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sun.admin
- Path: sparky!uunet!bose!bose.com!beaulieu
- From: beaulieu@bose.com (Larry Beaulieu)
- Subject: Re: 4/490 mysterious hangs (4.1.1)
- In-Reply-To: Steinar.Haug@delab.sintef.no's message of 15 Nov 92 01:57:43 GMT
- Message-ID: <BEAULIEU.92Nov17090140@piranha.bose.com>
- Sender: news@bose.com
- Nntp-Posting-Host: piranha
- Organization: Bose Corporation
- References: <1992Nov13.000223.8928@pwcs.stpaul.gov>
- <STEINAR.HAUG.92Nov14175743@delab.sintef.no>
- Date: Tue, 17 Nov 1992 14:01:40 GMT
- Lines: 58
-
- In article <STEINAR.HAUG.92Nov14175743@delab.sintef.no> Steinar.Haug@delab.sintef.no (Steinar Haug) writes:
-
- >> We have a 64MB (memory) 4/490 with four IPI drives that we are using as
- >> a sybase engine. It is starting to mysteriously hang more and more
- >> often and i am at a loss to explain why. The console is a wyse
- >> terminal and there are no error messages on the console, no messages in
- >> /usr/adm/messages* and no indications of any kind anywhere else that i
- >> can think of looking that would explain why the machine just simply
- >> stopped working. Usually there are only a few people working on it,
- >> maybe one or two running elm and news plus a couple of other smallish
- >> programs. Of course, the sybase server is running all the time. The
- >> times of the hangs appear random, i.e., during the day, during the
- >> night, during the weekends, etc.
- >>
- >> Does anyone have any experience with this? Sun could only suggest
- >> running vmstat periodically in order to try and track down memory usage
- >> at the time of the crash and also forcing a memory dump next time it
- >> crashes in order to see what routine it is in at the time of the
- >> crash.
-
- >We've had similar experiences with Sybase on a smaller Sun (SS-1) running
- >Sybase. Our problems disappeared after installing patch 100159-01. I have
- >included the README file below. You might also want to look into patches
- >100303-02 and 100330-06 (README for these also included).
-
- When I was at my previous employer we had the exact same problem, with a
- 4/470 w/32MB memory and 4 IPI drives running 4.1.1. The usual Sybase caveats
- were followed in the installation of this system (partition a not a
- raw partition, etc). Upon a hang the system would respond to an ICMP
- echo request (ping), but would not respond to keyboard input or
- other remote requests. This would occur anywhere from once per day
- to about once per week. We installed these patches and a couple of
- others, which I do not now remember. They had no effect on the type
- or frequency of the failure. I forget the results of our crash
- dumps, but there wasn't anything obviously funky. Sun software support
- didn't have much to offer. Neither did Sybase.
-
- This system was a dedicated Sybase engine which was almost exclusively
- dedicated to local processing (few client requests) but which was
- consistently under very heavy load updating a large database. We could
- not replicate the problem on any of our other Sybase engines. If
- I remember right this was under V4.2 of Sybase.
-
- Several times we would take a corrective step which would stop the
- problem, only to have it reappear a couple of months later. This
- included reloading the O/S, the sybase executables, and the database.
-
- The problem finally went away this spring when on a lark I replaced
- the IPI controller.
-
-
-
- --
- Larry Beaulieu beaulieu@bose.com
- Bose Corporation Phone: 508-879-1916 x6479
- Framingham, MA 01701-9168 FAX: 508-820-4865
-
- All flames gleefully ignored.
-