I had the same problems with squid-3.0.STABLExx (current 19) and saw up to
20 crashes a day.
Therefore I made a very crude workaround to avoid the crashes -- I do not
use disk cache in my config and
simply skip calling the routine "scheduleDiskRead" altogether by changing
the source code of src/store_client.cc:
********************************
# diff -u src/store_client.cc.org src/store_client.cc
--- src/store_client.cc.org 2009-09-06 13:29:38.000000000 +0200
+++ src/store_client.cc 2009-09-10 14:31:30.000000000 +0200
@@ -420,7 +420,12 @@
if (copyInto.offset >= mem->inmem_lo && copyInto.offset <
mem->endOffset())
scheduleMemRead();
else
- scheduleDiskRead();
+ {
+ /*scheduleDiskRead();*/
+ debugs(90,1,HERE << "WORKAROUND - averted calling
scheduleDiskRead()");
+ fail();
+ flags.store_copying = 0;
+ }
}
void
#
*********************************
It works for me (4 proxies with in sum about 1200 req/sec and 100 MBit/s
data flow as forward proxy with about 35k users)
but I can't really recommend it -- WARNING: squid can no longer read from
the disk cache with this "patch" ==> cache_dir null needed!
Best regards,
Martin
Ralf Hildebrandt <Ralf.Hildebrandt@xxxxxxxxxx>
21.10.2009 13:44
An
squid-users@xxxxxxxxxxxxxxx
Kopie
Thema
Re: squid3: assertion failed: store_client.cc:430:
"STORE_DISK_CLIENT == getType()"
> http://bugs.squid-cache.org/show_bug.cgi?id=2155
>
> I think Henrik was looking at it as part of store cleanups.
Oops, it's my bug :)
Maybe I should try 3.1.x now I have the dansguardian issues fixed.
--
Ralf Hildebrandt
Geschäftsbereich IT | Abteilung Netzwerk
Charité - Universitätsmedizin Berlin
Campus Benjamin Franklin
Hindenburgdamm 30 | D-12203 Berlin
Tel. +49 30 450 570 155 | Fax: +49 30 450 570 962
ralf.hildebrandt@xxxxxxxxxx | http://www.charite.de
