Squid Memory Useage /leak?



  • I have 2 pfsense boxes (mine and my friends).  We are running identical setups (now 2.0 release on vmware server) except I am the openvpn server and he his a client. This was same scenario when we were on 1.2.3 and physical boxes.
      I've noticed on my box for the last 6 months or so that my memory useage is always creeping up.  I always seem to reboot for one reason or another before 30 days so I never see the machine fill up.  Currently at 11 days I am 62% and climbing daily (at around 200 days it's over 80%).
      A large part of the memory usage is squid, but I went screen by screen and compared to the other machine's setup.  I can guess there is more usage on my box than my friends and the cache is just growing.  I have my memory cache size to 256mb, but I see squid is consuming more than that.  Can I expect to never hit 100%?
      Why am I worried about it now?  I plan on being away for 30 days, and am worried it will crash when I am gone.

    $ top
    last pid: 16806;  load averages:  0.00,  0.00,  0.00  up 11+19:46:27    16:45:53
    63 processes:  1 running, 62 sleeping

    Mem: 1065M Active, 472M Inact, 222M Wired, 213M Buf, 208M Free
    Swap: 4096M Total, 4096M Free

    PID USERNAME  THR PRI NICE  SIZE    RES STATE  C  TIME  WCPU COMMAND
    32136 snort      2  44    0  483M  396M nanslp  1  28:08  0.00% snort
    37338 proxy      17  44    0  609M  593M ucond  0  17:24  0.00% squid
    36053 root        1  64  20  5836K  1488K select  0  14:37  0.00% apinger
    42393 root        1  76  20  8292K  1756K wait    1  9:47  0.00% sh
    23081 root        1  44    0 66200K 19068K select  1  3:50  0.00% vmtoolsd
    23662 root        1  44    0 10996K  3724K select  1  1:19  0.00% openvpn
    50918 dhcpd      1  44    0 13056K  7600K select  1  0:53  0.00% dhcpd
    7778 root        1  49    0  8292K  1676K wait    1  0:51  0.00% sh
    53891 nobody      1  44    0 10144K  3232K select  0  0:46  0.00% dnsmasq
    43165 root        1  44    0 25768K  4536K kqread  0  0:28  0.00% lighttpd
    7277 root        1  60    0  7980K  1556K nanslp  1  0:28  0.00% cron
    23997 root        1  44    0 11768K  2388K bpf    1  0:14  0.00% tcpdump
    15465 _dhcp      1  44    0  5832K  1588K select  0  0:12  0.00% dhclient
    21826 root        1  44    0 14772K  2764K select  0  0:06  0.00% syslogd
    53342 proxy      1  44    0 59204K 12364K sbwait  0  0:06  0.00% squidGuard
    53515 root        1  76    0  104M 26540K accept  1  0:04  0.00% php
    4465 _ntp        1  44    0  5832K  1476K select  1  0:04  0.00% ntpd
    2777 root        1  49    0  104M 26868K piperd  1  0:03  0.00% php

    Anyone have any insight?



  • The squid memory usage depends on many things. On is your HARD DISK CACHE size. Every GB HDD is using ~10-15MB RAM.
    The 256MB of RAM is NOT the maximum squid is using. It is the MINIMUM squid uses. If squid needs more RAM it takes more RAM.

    But you do not have to worry. If squid is using mem and hdd it is a good sign because than squid is working as it should - caching.

    The high and low watermakrs will prevent your system to overload. If HDD or RAM reaches low water mark than squid starts to throw out "old" files. High watermark increases speed of throwing things out of mem and cache.


Locked