Do you have any log file rotation scripts running out of your cron every
hour? I had a similiar problem to this on BSDI a while ago that resulted
in a crash like this every night at about 00:00:06 or so (when my log
script would `kill -USR1` Squid). The solution for me was to use the gnu
malloc library and then very carefully tweak the cache_mem and associated
high/low watermark parameters.
It took a while, and didn't totally solve the problem, but I don't see it
crash nearly as often now. Note that at the time of day that my crash was
occuring there were zero users active (no one in the building) however.
-Bill
On Fri, 27 Mar 1998, Andre Albsmeier wrote:
> Hi,
>
> got a strange problem after upgrading from 1.1.20 to 1.1.21. Here
> is an extract of my syslog:
>
> Mar 27 09:00:08 curry.top.tld kern.info /kernel: pid 6821 (squid), uid 510: exited on signal 6
> Mar 27 09:00:18 curry.top.tld local4.err squid[11063]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 09:00:20 curry.top.tld local4.err squid[11063]: Ready to serve requests.
> Mar 27 10:00:09 curry.top.tld kern.info /kernel: pid 11063 (squid), uid 510: exited on signal 6
> Mar 27 10:00:20 curry.top.tld local4.err squid[11304]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 10:00:22 curry.top.tld local4.err squid[11304]: Ready to serve requests.
> Mar 27 11:00:08 curry.top.tld kern.info /kernel: pid 11304 (squid), uid 510: exited on signal 6
> Mar 27 11:00:19 curry.top.tld local4.err squid[11614]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 11:00:21 curry.top.tld local4.err squid[11614]: Ready to serve requests.
> Mar 27 12:00:08 curry.top.tld kern.info /kernel: pid 11614 (squid), uid 510: exited on signal 6
> Mar 27 12:00:18 curry.top.tld local4.err squid[11970]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 12:00:20 curry.top.tld local4.err squid[11970]: Ready to serve requests.
> Mar 27 14:00:15 curry.top.tld kern.info /kernel: pid 11970 (squid), uid 510: exited on signal 6
> Mar 27 14:00:25 curry.top.tld local4.err squid[12545]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 14:00:27 curry.top.tld local4.err squid[12545]: Ready to serve requests.
> Mar 27 15:00:08 curry.top.tld kern.info /kernel: pid 12545 (squid), uid 510: exited on signal 6
> Mar 27 15:00:19 curry.top.tld local4.err squid[13031]: Starting Squid Cache version 1.1.21 for i386-unknown-freebsd2.2.6...
> Mar 27 15:00:21 curry.top.tld local4.err squid[13031]: Ready to serve requests.
>
>
> It's funny that squid dies only every full hour and only during the daytime
> when my clients are busy. Now my question: What's happening every full hour
> so I can try to track the problem...
>
> Thanks,
>
> -Andre
>
Received on Fri Mar 27 1998 - 15:13:10 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:39:28 MST