CPU load average alert

linuxandtsm · 05-10-2012, 03:40 PM

Hi all,

Cpu load averages are very high on this 4 core linux box.
Below is the output of uptime command. How can i find more info about the processes causing this issue. I can see there are few perl command running from top command but is there any way to find more info ? and what should be done in general to solve such issue?

Code:

# uptime
 16:36:42 up 207 days,  3:20,  2 users,  load average: 5.22, 5.17, 5.16

Top commands output is below (first 5 lines)

Code:

top - 16:37:10 up 207 days,  3:21,  2 users,  load average: 5.14, 5.16, 5.16
Tasks: 163 total,   6 running, 156 sleeping,   0 stopped,   1 zombie
Cpu(s): 99.9%us,  0.1%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   3986160k total,  3955948k used,    30212k free,     3344k buffers
Swap: 25165816k total,  8114280k used, 17051536k free,  1179404k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
14590 oracle    25   0  789m 465m 3044 R 100.0 11.9 344:36.53 perl
  864 oracle    25   0 2903m 389m 2980 R 99.7 10.0   1203:14 perl
31948 oracle    25   0  284m 201m 3044 R 99.7  5.2 121:17.75 perl
22883 oracle    25   0 1252m 396m 3044 R 50.2 10.2 684:39.71 perl
11917 oracle    25   0 1730m 358m 3044 R 49.8  9.2 944:57.73 perl

Kustom42 · 05-10-2012, 03:47 PM

Code:

ps auxww | grep perl(or the pids)

That should give you which file is spawning the perl process.

linuxandtsm · 05-10-2012, 03:55 PM

Hi Kustom42, thanks for the reply.

below is the output

Code:

# ps auxww | grep perl
oracle     864 94.6 10.5 2990652 421884 ?      R    May09 1214:50 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/emdprocstats.pl 23542
oracle   11917 94.4  9.7 1790688 387540 ?      R    May09 959:24 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/emdprocstats.pl 1437
oracle   12643  0.0  0.1  24108  5896 ?        R    16:53   0:00 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/lsnrresp.pl
root     12645  0.0  0.0  61176   740 pts/0    S+   16:53   0:00 grep perl
oracle   14590 90.4 12.5 830240 500592 ?       R    10:17 357:50 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/emdprocstats.pl 30187
oracle   22883 93.3 10.6 1301864 426144 ?      R    04:26 697:17 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/emdprocstats.pl 12524
oracle   30165  0.0  0.0   9628  3252 ?        S    06:18   0:00 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/bin/emwd.pl agent /u01/app/oracle/agent10g/sysman/log/emagent.nohup
oracle   31948 80.3  6.2 331416 249028 ?       R    14:06 133:42 /u01/app/oracle/agent10g/perl/bin/perl /u01/app/oracle/agent10g/sysman/admin/scripts/emdprocstats.pl 14889

Kustom42 · 05-10-2012, 03:57 PM

Ok.. so you now know the actual perl script that is causing your utilization problem. Time to look over the script and see if there is some coding issues or if you just need better hardware.

Kustom42 · 05-10-2012, 03:59 PM

Doing some google searching on that oracle .pl file found this patch may help: updates.oracle.com/ARULink/PatchDetails/process_form?patch_num=5908032

DONT INSTALL IT UNTIL YOU READ WHAT IT DOES.

Kustom42 · 05-10-2012, 04:01 PM

Bad link, sorry. Here it is. http://docs.oracle.com/cd/B16240_01/...c.htm#CHDHJJGG

linuxandtsm · 05-10-2012, 04:15 PM

Hi Kustom42,

Thank you very much!