2

I'm on a (mt) (dv) base 3.5 server and am running ~20 small sites and I keep running into memory issues:

cat /proc/user_beancounters

Version: 2.5                                                                   
   uid  resource           held    maxheld    barrier      limit    failcnt
 30782: kmemsize        6836615    9267688   12288832   13517715   36800687
        lockedpages           0          0        600        600          0
        privvmpages      102117     149651     229036     249036     117066
        shmpages           7301       7301      17237      17237          0
        dummy                 0          0          0          0          0
        numproc              53         64        300        300          0
        physpages         50840      91157          0 2147483647          0
        vmguarpages           0          0     131072 2147483647          0
        oomguarpages      51395      91712     131072 2147483647          0
        numtcpsock           17         20        300        300          0
        numflock              7          8        480        528          0
        numpty                2          2         30         30          0
        numsiginfo            0          1       1024       1024          0
        tcpsndbuf        166248     278048    2867477    4096277   38165018
        tcprcvbuf        269584     285968    2867477    4096277          0
        othersockbuf      20124     192108    1433738    2662538          0
        dgramrcvbuf           0          0    1433738    1433738          0
        numothersock         19         21        300        300          0
        dcachesize       723292     755098    2684271    2764800          0
        numfile            2392       2531      10000      10000          0
        dummy                 0          0          0          0          0
        dummy                 0          0          0          0          0
        dummy                 0          0          0          0          0
        numiptent            14         14        200        200          0

I've run through the (mt) optimization suggestions, but haven't had much change. Any suggestions for tracking down what's wrong?

Thanks!

razorfrog
  • 21
  • 1

1 Answers1

1

Here's a Perl script which will dump the output of ps faux for logging if your VE hits its threshold for kmemsize or privvmpages: Perl Memory Monitor 1

Note that you can alter the panic_time subroutine behavior to get more information as needed.

danlefree
  • 2,873
  • 1
  • 18
  • 20