-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 16/01/2015 1:50 a.m., Yuri Voinov wrote: > > Hi gents, > > Sometimes I have too much assertions: > > 2015/01/08 09:40:39 kid1| assertion failed: comm.cc:178: > "fd_table[conn->fd].halfClosedReader != NULL" > > Core dump not produced but Squid restarts. So, I can't get stack > trace. > > When I set hals_closed_clients to on, I've got much more > assertions: > > 2015/01/15 18:41:31 kid1| assertion failed: comm.cc:1823: > "isOpen(fd) && !commHasHalfClosedMonitor(fd)" > > What can be wrong? This is very annoying behaviour. Also memory > cache is often cold after restarts and hit ratio is degrades. > > This looks similar adaptation bug 4057, but it is closed in 3.4.11. > This behaviour depends half-closed connections, but how? > > And how to troubleshoot? A) running Squid under GDB. There are instructions in <http://wiki.squid-cache.org/SquidFaq/BugReporting> on how to do that even if you are debugging a production server that needs no downtime. OR, B) finding out how to enable core dumps in your OS. The assertion handling calls abort() which will generate a core file if the OS allows them to be created. Amos -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (MingW32) iQEbBAEBAgAGBQJUt71IAAoJELJo5wb/XPRjVJEH+OPwizmwffDGN4nrKW/OwfI3 X4Pu15cXgie+6WWBsVBuz7FHlOv846oL6B2oSGHWTcAy2iR9E1+BZc/BtcIR+hZ+ vxjSmisDT7h9NvynVelrA4yqw2hrZeVVSNyuaaLk1ks2tvTYb6ms7EldGN2TKDUg Q1aXgofE4p+5mXrZB8vEnJrU4CrXp/8pGJ6+YM4tQhDcI+fvcy4rSi7lTxbbxXnL jsSLSnq9rNXKtBB3Ig6gVF3nQhwhl0IpGw571ORAuRQ7uvDC5d/fT/uNOpWfOFA5 4H1VR8RY/IddxEkhqdbnkYn6fGzk6qyOV3ZQKQi8p2mcyuNGfpAKbD9ZC+v28A== =uSQr -----END PGP SIGNATURE----- _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
