Hello,
I am running Squid 2.5STABLE1 on a P4-1800 with 2GB of RAM and 2 18Gig HDs,
reiserfs. I have been for the longest time battling a problem of a
constantly growing swap space. I have been told to increase RAM, use PS to
see squid memory usage, and within a week I exhaust all swap space and need
to stop the squid process and re-start squid to free the paged memory. I am
running Linux 2.4.20, however I had the same problem with Linux 2.2.20 when
I was running Squid on an older machine.
Here is the output of PS:
root 700 0.0 0.0 6544 648 ? S Jan16 0:00
/usr/local/squid/
nobody 702 45.6 38.8 886292 805280 ? R Jan16 2556:24 (squid)
nobody 703 0.0 0.0 1348 264 ? S Jan16 0:00 (unlinkd)
nobody 705 0.0 38.8 886292 805280 ? S Jan16 0:00 (squid)
nobody 706 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 707 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 708 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 709 0.0 38.8 886292 805280 ? S Jan16 1:16 (squid)
nobody 710 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 711 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 712 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 713 0.0 38.8 886292 805280 ? S Jan16 1:16 (squid)
nobody 714 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 715 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 716 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 717 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 718 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 719 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 720 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 721 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 722 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 723 0.0 38.8 886292 805280 ? S Jan16 1:16 (squid)
nobody 724 0.0 38.8 886292 805280 ? S Jan16 1:16 (squid)
nobody 725 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 726 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 727 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 728 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 729 0.0 38.8 886292 805280 ? S Jan16 1:13 (squid)
nobody 730 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 731 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 732 0.0 38.8 886292 805280 ? S Jan16 1:16 (squid)
nobody 733 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 734 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 735 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
nobody 736 0.0 38.8 886292 805280 ? S Jan16 1:14 (squid)
nobody 737 0.0 38.8 886292 805280 ? S Jan16 1:15 (squid)
Notice squid seems to be using 886M of RAM, however "free" reports:
free
total used free shared buffers cached
Mem: 2070368 2062064 8304 0 304564 712196
-/+ buffers/cache: 1045304 1025064
Swap: 104412 41700 62712
There is 42Mb of swap in 3 days of runtime.
Cache mem is set to 256Mb. Here is the relavent CacheMgr output:
Memory usage for squid via mstats():
Total space in arena: 816748 KB
Total free: 1789 KB 0%
Memory accounted for:
Total accounted: 436218 KB
memPoolAlloc calls: -2073504971
memPoolFree calls: -2078299351
File descriptor usage for squid:
Maximum number of file descriptors: 8192
Largest file desc currently in use: 1564
Number of file desc currently in use: 1508
Files queued for open: 0
Available number of file descriptors: 6684
Reserved number of file descriptors: 100
Store Disk files open: 47
If I am to stop squid with -k shutdown, the paged memory dissapears except
for about 2Mb. I can then re-start squid again and run for a few days
before my swap gets exhausted again. I NEVER need to reboot the machine.
I tried using dlmalloc, gnumalloc and no memory pooling and in all cases I
have the same results.
Any ideas?
Thanks
Ron
_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*
http://join.msn.com/?page=features/virus
Received on Mon Jan 20 2003 - 09:33:09 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:12:44 MST