I cc'ed your this response to the linux kernel mailing list. Maybe somebody
there can help you. My problems still aren't solved and I am not sure they
are identical to yours. I am not a kernel expert, but there are some comments
further down.
Christian Carstensen wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>
> Searching Dejanews, I've found two articles concerning some crashes on
> systems running 2.0.33 and sysklogd version 1.3. I could not find any
> further information about this topic and recently have no access to the
> above mentioned newsgroup.
> The incidents described in these postings seem to be similar to those,
> I could observe on our central server machine. Something causes a system
> hang, that effectively prevents any login on console. All systems, this
> bug ocurred on, use the AHA2940 SCSI host adapter.
> Fact is, that syslogd, when started with remote reception enabled, decides
> to lock up, when recieving a bulk of data (here it were about 40k) from
> the network. This weird problem causes the next login session, and even
> the system shut down, to hang when trying to write a logfile message.
>
> Reproduction as follows:
>
> 1. login as root on tty1
> tail -f /var/log/messages
> 2. flood syslog with messages from network until tail shows no further
> syslogd response.
> 3. try to log in on tty2
> this login should hang after you've entered your password.
> 4. stop syslogd
> 5. at this point your login on tty2 should complete.
>
> My current solution is to turn off remote reception on the affected
> machine, but as you can imagine, this is not acceptable for longer time.
> So, do you have any ideas or new informations concerning this bug?
I am not using remote logging. But I just became aware that there are some
patches
to syskolgd 1.3 (I found them in the Red Hat distribution even though I use
Slackware).
I can just pass on some advice I got and and from what I read on the mailing
list. Several p
eople have reporte problems (either reboots or hangs) with 2.0.33 (actuall
any 2.0.3x). But
there does not seem to be a common hw/feeature combination to pinpoint the
problem. Most
likely it is a problem in the networking code.
Some peolpe recommend going back to 2.0.29 with some patches (teardrop,
etc.). I haven't
done that yet.
Make sure you get the latest patch of the aic7xxx driver (5.0.8) that might
help you
with the SCSI controller.
You might also want to check out Alan Cox' latest 2.0.34 pre-patch.
Unfortunately all the kernel gurus are busy working on the 2.1 kernels and I
would not
hold my breath that what ever weird bug this is will get fixed any time soon.
Patrick
>
>
> sincerely,
> C. Carstensen
>
> -----BEGIN PGP SIGNATURE-----
> Version: 2.6.3i
> Charset: noconv
>
> iQCVAwUBNRB9fo7OF1ZkFrjNAQHhDAP/cnOuR/CW+LW2LNQHQHzrzlJAqfM+sb4s
> s1u7P9LtKIOnVu6IVE7ucn6mb54xCGvSBzZ+Umd4CFtAIvIs7xdTRTMCWAOF7oIw
> zeXZ3FUWMsq3EshsLffTUlz87+0atGF0RqGD+dry0DuYW4qt66Aj3x2Nv8TtT5nD
> YDaLQW4uP9A=
> =lhdc
> -----END PGP SIGNATURE-----
--
-------------------------------------------------------------------
Patrick D. Wildi patrick@wildi.com
http://www.wildi.com/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu