Keywords: FreeBSD, Squid, xmalloc failure
I seem to be having some serious run-time problems with Squid (vers
1.1.1). Im running it on a FreeBSD platform (2.2.5, 128Mb Ram, 8Gb
scsi disk).
Currently Im running with only 16 users (approx. 100,000 hits per
day). Trouble is, squid crashes every-so often with a message;
xmalloc - cannot allocate 4028 bytes
Im bemused, I'v reduced the cache_mem parameter from 96 (which I
though would be ok with 128Mb ram), down to 18. Still it crashes. I
cant seem to get much info on whats going on - other than results
from "top" - which show atypically before a self-distruct the
following sort of parameters;
>last pid: 2106; load averages: 0.05, 0.05, 0.02
> 14:33:06 30 processes: 1 running, 28 sleeping, 1 stopped CPU
>states: 2.3% user, 0.0% nice, 2.7% system, 1.6% interrupt, 93.4%
>idle Mem: 73M Active, 1812K Inact, 21M Wired, 28M Cache, 8344K Buf,
>616K Free Swap: 256M Total, 96K Used, 256M Free
>
> PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU
> COMMAND
> 2084 root 2 0 32388K 32568K select 1:05 3.24% 3.24% squid
I dont know if its a squid problem - or a FreeBSD problem. It's not
clear to me what Wired and Cache memory are - other than as Free
falls below 1M, cache slowly reduces. In any instance when the size
of squid climbs to over 33000K it usually fails.
Please - anybody out there running on FreeBSD with this sort of
configuration - or anyone with any idea - please let me know!
Regards
Phil Richardson
Network Services Manager
Computing and Network Services
University of Lincolnshire and Humberside
Phil Richardson, Network Services Manager
Computing and Network Services
University of Lincolnshire & Humberside
prichardson@humber.ac.uk
Tel: 01482 463101 / Fax: 01482 440279
Received on Wed Feb 18 1998 - 06:47:08 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:38:55 MST