On Tue, 22 Sep 2009 18:04:44 -0500, Luis Daniel Lucio Quiroz
<luis.daniel.lucio_at_gmail.com> wrote:
> Hi Squids,
>
> Well, I dont know if this a already discussed topic. Because we are
having
>
> many squid-reboots we need to know what users were doing.
>
> In my cache.log I have this:
> 2009/09/15 23:55:22| Ready to serve requests.
> FATAL: Received Segment Violation...dying.
>
> I'm using the line just before the "dying" message to know hour.
> Then I know in access.log first line is time in epoch format. However,
> after
> converting cache.log time with this code (perl):
> my $time = timelocal($sec,$min,$hours,$day,$month -1,$year-1900);
>
> I'm unable to find that time in access.log.
>
> Am I missing something?
Yes. Unless you are lucky and the failure is occuring during the cleanup
following a request there will be no access.log data about it. Squid
processes many requests simultaneously. any one of them could be dying at
any point. What you need to do is get the core dump data from the death and
generate a stack trace from it.
Details are in the FAQ. http://wiki.squid-cache.org/Squidfaq/BugReporting
Amos
Received on Wed Sep 23 2009 - 01:18:00 MDT
This archive was generated by hypermail 2.2.0 : Wed Sep 23 2009 - 12:00:03 MDT