Open GoogleCodeExporter opened 9 years ago
Wait. So the object is there, and ICP_HIT is returned, but the fetch fails?
Is that right?
Original comment by adrian.c...@gmail.com
on 30 May 2010 at 11:01
Seems so, it is appearing after running LUSCA for long time on heavy load (this
cluster receive up to 80k simultaneuous connections at peak time) without
restart
(usually i am restarting each night).
So it is actually very difficult to trigger it, now also i have less complaints
about
it. I will make sure at monday, if there is still any and probably will make
something to report instantly to me about such issues.
Original comment by nuclear...@gmail.com
on 30 May 2010 at 11:33
Probably it is related to issue 114 (memory corruption)?
Original comment by nuclear...@gmail.com
on 30 May 2010 at 11:41
Which version are you running?
Original comment by adrian.c...@gmail.com
on 30 May 2010 at 12:19
It is some HEAD, but no idea which revision (those one i dont touch, because
current
HEAD have some issues and i cannot push them to production servers till i will
be
able to make bugreport).
Maybe it is possible to include revision tag in squid -v ?
The only i know it is looks like fetched before 28 Apr.
Original comment by nuclear...@gmail.com
on 30 May 2010 at 12:26
I've begun including the svn revision in the reported version for the source
tarballs
found in the download section. I don't recall when I began doing this but it's
very
recently (in the last month or two.)
Sorry. :(
Original comment by adrian.c...@gmail.com
on 30 May 2010 at 2:15
Squid Cache: Version LUSCA_HEAD-r14674 with this issue
I did automated report, it happens mostly on POST requests.
Possible it is somehow related with recent memory corruption.
I will try to test revision that is marked as stable in memory corruption issue.
Original comment by nuclear...@gmail.com
on 7 Jun 2010 at 10:09
Original issue reported on code.google.com by
nuclear...@gmail.com
on 30 Apr 2010 at 12:48