[SOLVED] Very high CPU usage with 7.1.1

[MichaelB12]

Just little thing...

After apply PATCH 2 for 7.1.1 my cpu start again to overload et ~40% of cpu.. I saw that zmconfigd not running... even after the patch.

I read and read i that forum and find someone that wrote to install "nc"

So i did:

yum install nc

/etc/init.d/zimbra restart

And now.. zmconfigd is running and my cpu back to ~5% usage... i don't really understand but it work..

[PhD]

i cant see how installing nc would have resolved this... - what made you even think of installing it in the first place? - sounds like coincidence if yours was working fine

[MichaelB12]

You right is may be a coincidence... and still not sure of the cause of this problem.

Is where i read not information.. why zmconfigd is not running.. how to fix it.. ?

zmconfigd not running error - all possible fix

[espresscafe]

i'm on Release 7.1.2_GA_3268.RHEL6_64_20110804131406 UNKNOWN_64 FOSS edition.

Nothing worked to fix the problem so i reverted to a script to kill the zimbra log process that uses all the CPU. It's a bit brute force, but that's the only thing that works. I schedule that script in the cron every 15minutes.


kill_high_cpu_zimbra_log.pl
-------------------------------------------
#!/usr/bin/perl
my $loop=0;
# get process from zimbra user that uses more than 95% cpu
my $cpuhigh=&check_zimbra_log();
# if none
if ($cpuhigh<0) {
print "all good\n";
exit;
}
# loop 5 times with a sleep of 5 check the same process is still high, then kill -9
while ((&check_zimbra_log()==$cpuhigh) && ($loop<5))
{
print "HIGH : pid=$cpuhigh - loop $loop\n";
$loop++;
print "sleep 5\n";
sleep 5;
if ($loop==5) {
my $kill="kill -9 $cpuhigh";
print "$kill\n";
`$kill`;
}
}

# --------------
# return pid or -1
sub check_zimbra_log ()
{
my $cmd="ps axo pid,user,pcpu,cmd | grep zimbra | sort -n -k 3 | tail -1";
my $cmd_res=`$cmd`;
my $res=0;

if ($cmd_res=~/([0-9]+) zimbra[\s]+([0-9\.]+)(.*)/) {
$res=$1;
my $cpu=int($2);
if ($cpu>95) {
if (index($cmd_res,"java -XX:ErrorFile=/opt/zimbra/log")>0) {
return $res;
}
}
}
$return -1;
}
#------------------------

[phoenix]

Originally Posted by espresscafe

i'm on Release 7.1.2_GA_3268.RHEL6_64_20110804131406 UNKNOWN_64 FOSS edition.

You don't mention whether you've applied any of the patches to your installed version, in any case I'd suggest you upgrade to the most recent release of Zimbra and see if it still occurs.

[llauradog]

Hi,

I have a zimbra installation (Open Source server) v7.1.3 on Ubuntu 10.04.1 x64.

I am also experiencing this memory problem...

top - 17:16:04 up 3 days, 3:19, 2 users, load average: 60.31, 60.04, 60.68
Tasks: 572 total, 2 running, 570 sleeping, 0 stopped, 0 zombie
Cpu(s): 1.1%us, 2.7%sy, 0.0%ni, 83.9%id, 12.1%wa, 0.1%hi, 0.1%si, 0.0%st
Mem: 8178456k total, 7558816k used, 619640k free, 124396k buffers
Swap: 19780600k total, 17148k used, 19763452k free, 3152192k cached

This is a dedicated server with no other services running, only Zimbra Open Source edition.

The server is shutting down at random (when there is no more memory).

Any ideas?