OpenVZ Forum


Home » General » Support » Strange Memory Problem
Strange Memory Problem [message #38714] Tue, 19 January 2010 11:55 Go to next message
ziceva is currently offline  ziceva
Messages: 2
Registered: January 2010
Junior Member
Hy,

I have a rather strange problem with a OpenVPS server ...
The HN is a Core 2 Duo E4500 @2.20 GHz with 3Gb of RAM ...
I've created a VPS with 1Gb of RAM and installed cPanel inside it ... and everything went smooth ... I even run a unixbench, just to see how would it perform ...
I've put a site on this server (ok, it is a rather heavy application... but the 1Gb i reserved for it should be enough) ... and that also worked fine for a few hours ... than the problem:

Now my apache server complains about not having enough memory:
/usr/bin/php: error while loading shared libraries: libXdmcp.so.6: failed to map segment from shared object: Cannot allocate memory

But this is not suported by anything I see on that server/VPS...

/proc/user_beancounters
Version: 2.5
uid resource      held     maxheld   barrier  limit  failcnt
101:kmemsize     8234055  10406978  67108864 73819750 0
    lockedpages  0        0         256      256      0
    privvmpages  61470    90406     262144   288358   0
    shmpages     23       23        21504    21504    0
    dummy        0        0         0        0        0
    numproc      42       57        500      500      0
    physpages    27087    55108     0        MAX UINT 0
    vmguarpages  0        0         262144   MAX UINT 0
    oomguarpages 27087    55108     262144   MAX UINT 0
    numtcpsock   24       49        750      750      0
    numflock     1        5         188      206      0
    numpty       1        2         16       16       0
    numsiginfo   0        11        256      256      0
    tcpsndbuf    420096   630144    3440640  5406720  0
    tcprcvbuf    393216   543296    3440640  5406720  0
    othersockbuf 67744    219792    1126080  2097152  0
    dgramrcvbuf  0        8472      262144   262144   0
    numothersock 35       40        750      750      0
    dcachesize   0        0         16777216 19293798 0
    numfile      1318     1572      9312     9312     0
    dummy        0        0         0        0        0
    dummy        0        0         0        0        0
    dummy        0        0         0        0        0
    numiptent    29       29        128      128      0


free -m
     total        used      free    shared    buffers     cached
Mem: 1024         240       783     0          0          0
-/+ buffers/cache:240       783
Swap:             0          0          0

and on the HN:
free -m
     total        used    free   shared    buffers     cached
Mem: 3011         1514    1496   0         186         828
-/+ buffers/cache:499     2511
Swap:             2732    0      2732

On the HN there is nothing else running ... on the VPS I have:
init-+-cpanellogd
     |-cpsrvd-ssl
     |-crond
     |-dbus-daemon
     |-dovecot-+-dovecot-auth
     |         |-2*[imap-login]
     |         `-2*[pop3-login]
     |-exim
     |-httpd---12*[httpd]
     |-mysqld_safe---mysqld---3*[{mysqld}]
     |-pure-authd
     |-pure-ftpd
     |-queueprocd
     |-snmpd---{snmpd}
     |-spamd---spamd
     |-sshd
     |-tailwatchd
     `-vzctl---bash---pstree

I really don't get what is happening ... everything else is working fine, my apache server is serving static pages just fine ... At first I thought it was a apache/php thing so I recompiled ... no luck there ... (restarting the HN/VPS did not help either)

Thank you

PS Both the HN and the VPS have CentOS 5 X86_64

[Updated on: Tue, 19 January 2010 12:12]

Report message to a moderator

Re: Strange Memory Problem [message #38734 is a reply to message #38714] Fri, 22 January 2010 15:02 Go to previous messageGo to next message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
Hello,

let me ask you a question, please?
Did you restart a VE after the problem had appeared?
The point I'm driving at is that failcounters (the last column of /proc/user_beancounters output) are cleared when a VE is being restarted so you might have missed that by chance. Your problem is pretty much looks like a resource shortage inside a VE and should be revealed via /proc/user_beancounters output.
Anyway (to make all suggestions within a single post) could you please increase privvmpages parameter to check that it is not a reason?
Re: Strange Memory Problem [message #38736 is a reply to message #38734] Fri, 22 January 2010 16:02 Go to previous message
ziceva is currently offline  ziceva
Messages: 2
Registered: January 2010
Junior Member
Hy,

Thank you for your reply.
I stopped invetigating the VPS solution for now (will probably start again in a few weeks 'cause it looks very promising ).

Yes, I restarted the VPS (and the HN) ... but this only explains why the the failcount are all 0 ... it does not explain why my apache complains about not having memory (after restart) although all the counters and barriers in user_beancounters say otherwise ...

Another interesting thing is that only apache complains, and only if php is used as CGI (I am using suPHP on all my servers, so I need to compile PHP as a CGI program). If I use it as a DSO, it works ...

I will be doing a fresh install of the HN and the VPS and try everithing all over again (and start with a less heavy application).
Previous Topic: all pop3 connection come from host ip
Next Topic: VPN server inside a CT?
Goto Forum:
  


Current Time: Tue Nov 19 00:53:52 GMT 2024

Total time taken to generate the page: 0.02976 seconds