> Hi, > could you please give us a few more details? Squid version (squid > -v), operating system, whether you got a binary package or rolled your > own, a few info about the setup (forward proxy? Reverse? Transparent?) > It's hard to tell from what I read so far, unless I missed something > He has said 3.0.stable10. With aufs problems as well, so probably running on some form of BSD. The rest of the Q's still need answering though. I'm particularly interested in the configure options used to build, confirmation of the OS, the rest of the backtrace from the core, and whether its a vanilla squid or patched. Amos > On 11/13/08, Marcel Grandemange <thavinci@xxxxxxxxxxxxxxx> wrote: >>>Good day. >> >> >>>Im wondering if anybody else has experienced this. >>>Since ive upgraded to squid3stable10 the proxy continuously stops >>>responding. >>>Firefox will say something along the lines of the proxy isn't setup to >>>accept connections. >> >>>I hit refresh and it loads page perfectly, then next page it loads all >>> the >>>pics half .... and so on.. >> >> >>>This has only been introduced in stable10 and isn't the link as I tested >>>with a neighboring cache running stable9, no issues. >> >>>Under further investigation system log file presented following: >> >>>Nov 13 19:37:21 thavinci kernel: pid 66367 (squid), uid 100: exited on >>>signal 6 (core dumped) >>>Nov 13 19:37:21 thavinci squid[66118]: Squid Parent: child process 66367 >>>exited due to signal 6 >>>Nov 13 19:37:24 thavinci squid[66118]: Squid Parent: child process 66370 >>>started >> >> Under Even further investigation cache.log revealed following... >> JUST before squid crashes each time there is the following entry... >> >> 2008/11/14 00:03:55| assertion failed: client_side_reply.cc:1843: >> "reqofs <= >> HTTP_REQBUF_SZ || flags.headersSent" >> >> >>>Also for those interested, a while back I had issues with the >>> performance >> of >>>squid... >>>Objects retrieved out of cache never went faster than 400K, turned out >>> when >>>I changed my cache_dir from aufs to ufs this was resolved, now objects >>> some >>>down full speed on LAN. >> >> I am pretty desperate to get this problem solved as this is affecting us >> big >> time. >> >> >> > > > -- > /kinkie >
