I observed the problem even with zero NFS activity. Infact, for awhile
there were no RPC programs running at all and it still crashed. I never
explicitly tried the delack-timer-5 patch but instead compiled
2.3.99-pre7-8. The source may contain this patch, but I'm not very
patch/version savy (no flames please). I built the kernel with cpqarray and
the tlan driver static linked because at one point I had trouble with
cpqarray as a module with another revision of 2.3.99.
All the crashing problems have vanished. The system has been up for 4 days
:)
I now often get kernel messages like:
TW_REC: reject openreq 35040828/35000500 208.184.175.99/19509
TW_REC: reject openreq 41273047/35090657 208.184.175.99/12084
TW_REC: reject openreq 41273047/35090957 208.184.175.99/12084
TW_REC: reject openreq 41273047/35091362 208.184.175.99/11036
TW_REC: reject openreq 41273047/35091557 208.184.175.99/12084
TW_REC: reject openreq 41273047/35091662 208.184.175.99/11036
...
but understanding is that these messages are informational only. Right?
Also, when the system (Redhat 6.2) booted, I got:
May 10 16:37:53 kenny automount[647]: attempting to mount entry /import/i686
May 10 16:37:53 kenny automount[811]: lookup(file): lookup for i686 failed
May 10 16:37:53 kenny automount[647]: attempting to mount entry /import/mmx
May 10 16:37:53 kenny automount[647]: attempting to mount entry
/import/libtermcap.so.2
May 10 16:37:53 kenny automount[813]: lookup(file): lookup for
libtermcap.so.2 failed
May 10 16:37:53 kenny automount[647]: attempting to mount entry
/import/libc.so.6
May 10 16:37:53 kenny automount[814]: lookup(file): lookup for libc.so.6
failed
May 10 16:37:53 kenny automount[647]: attempting to mount entry
/import/libnss_files.so.2
May 10 16:37:53 kenny automount[815]: lookup(file): lookup for
libnss_files.so.2 failed
May 10 16:38:13 kenny automount[647]: attempting to mount entry
/import/libncurses.so.4
May 10 16:38:13 kenny automount[818]: lookup(file): lookup for
libncurses.so.4 failed
which I thought was kinda strange as my autofs is not configured to ever do
something like this. The above appeared only on boot so I attributed it to
something running in the /etc/init.d/* files that doesn't quite cooperate
with 2.3.xx. I have no time right now to figure this one out.
Note that if you are going to try the 2.3.xx kernels, depending on
distribution,
you may have to upgrade some of the packages on your system. See the Changes
file
in the Documentation directory of the kernel.
Best of luck,
Steven Uggowitzer
uggowitzers@who.int
World Health Organization
Geneva, Switzerland
-----Original Message-----
From: Jukka Timonen [mailto:jtimonen@evil.netppl.fi]
Sent: Monday, 15 May 2000 05:36
To: uggowitzers@internet.who.int
Cc: stoch@orc.ru
Subject: re: HARD CRASH with kernels (2.2.4,12,14,15)
> This problem occurs on our primary web server. Symptom is a completely
> frozen machine with typically no error messages. Occurs randomly every
> 2hrs to 3 days. I have seen similar reports on several newsgroups
including
> this listserv. My architecture is:
hi,
We're having same kind of symptoms on one of our server. Hardware problems
are ruled out, and the same thing happens with SMP/UP kernels. Most often
it's just a solid lock, sometimes same kind of messages appear that Steven
Uggowitzers posted on linux-kernel list.
This machine had also nfs-client running. When the nfs-mount was removed,
server stays up maybe few weeks or so. It's noticeably better than before,
but still crashes occasionally. Do you have had any luck with Andrea's
delack-timer-5 patch? I'm little suspicious about the problem being
nfs-related however, as nfs uses the udp protocol.
terveisin,
Jukka Timonen
Net People Ltd.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon May 15 2000 - 21:00:25 EST