Redhat 4 Update 4 CPU usage 99%

Latest response

Hi

i have a situation that my Oracle database servers are running on Redhat 4 U4 over Intel Itanium server, once in a month the server CPu will reach 99% and will not come down unless ew restart the Server,

can anyone shed the light on this issue

Responses

Hi Shafeek.

Potential problems with 3rd-party applications aside, there are NUMEROUS issues you could be hitting -- running such an out-of-date version of RHEL (v4.4 dropped in 2006). I would suggest updating first to RHEL 4.9 to see if the issue is resolved, but since I doubt you'll do that, can you give us any more information?

  1. Do you have any ideas what process is hanging?
  2. Have you looked through the logs for anything anomalous?
  3. Does the sar data give any hints as to what's going on?

Hi Ryan

thanks for the response,  i was stuck in other tasks 

  1. Do you have any ideas what process is hanging?
  2.            Till now we are not able to identify the root cause fo this issue
  3. Have you looked through the logs for anything anomalous?
  4.            No, i checked all the logs but no findgs
  5. Does the sar data give any hints as to what's going on?
  6.              SAR data is showing the time of System CPU spike but not giving reelevant process causing this spike.

Dear Shafeek,

As Ryan said, there could be NUMEROUS issue. Yes updating to 4U9 kernel may solved this issue. But I doubt. You said once in a month it uses full CPU resources. But you did not mention time when it touches 99%. During pick or off-pick hour? You did not mentioned the which oracle versions you are using. 

1. Firts take a look I/O status (pick/off-pick) -- iostat

2. Check virtual memory / swap memory -- vmstat 

3. Check processor related statistics - mpstat

4. Check SGA memory status during it CPU is high

 

Please be informed this could be oracle issue as well and suggest me if I am wrong.

 

Shyfur

Shafeek:

Does this once-a-month occurrence follow a strict schedule?

If so, it could be caused by some sort of scheduled job (e.g., kicked off by cron, perhaps). Look in to that. (I.e.: check /etc/crontab, /etc/cron.*/, and /var/spool/cron/)

"SAR data is showing the time of System CPU spike but not giving reelevant process causing this spike."

Well, sar doesn't include per-process data; however, my point in asking was to see what trends you could find -- does the RAM or CPU usage (loadavg, number of processes) creep up every day, leading up to the event? Or is resource utilization steady aside from a spike at the end?

As I said before Shafeek, there are rarely magic answers with these kinds of things; without more data there's not much we can do. I recommend that you open a case with Red Hat technical support if you can.