topic Re: Inbox | Palo Alto Networks Live in General Topics

Inbox | Palo Alto Networks Live

tiwara — Fri, 11 Jul 2014 16:12:20 GMT

Inbox | Palo Alto Networks Live

Hello All,

I have faced the similar kind of issue for panorama , one Job got stuck in 40% and we are not able to commit to the panorama. Also the management CPU was 100.

Well after restart the management server the issue is resolved for clearing the job .but i took the PA support guy to clear the PID to reduce the management CPU utilization.

We have faced this issue twice in continuously two days. Can anyone please advise for the permanent fix for the issue .

We are at 5.1.6 panorama version.

Shared with Jive Anywhere

https://live.paloaltonetworks.com/inbox

Re: Inbox | Palo Alto Networks Live

kadak — Fri, 11 Jul 2014 16:50:33 GMT

Hello tiwara,

Restarting the management server will always be a temporary workaround. We need to know what made the management plane go high (CPU wise) in the first place to identify the culprit.

For that, TAC would require:

1.) Trace of the management server (debug software trace management-server). Associate the PID of the management server and find out what thread the function is being called.

2.) show system resources follow (1+shift H + C). That would give the top running resources with threads on and CPU's consumption on individual basis

3.) Core the management server process so that we can perform GDB analysis on it .

Once a case is opened, a TAC engineer would help you gather that information and then can present his/her findings to the engineering team for RCA.

FYI, there were are no known management CPU issues which are fixed in 5.1.7 and 5.1.8

Hope that helps!

Thanks and regards,

Kunal Adak

Re: Inbox | Palo Alto Networks Live

tiwara — Mon, 14 Jul 2014 11:48:58 GMT

Hello Kadak,

Thanks for the above details .

A TAC is already opned with the team ,

The case # is 00231173 . The tech support file is already updated in the request .

Well I remember there is one PID that is utilization the 95 % of the CPU .

Support team loged into the root side and kill the process , after that the CPU is normal .

Management CPU is good after that , but we still have the issue for commit , as one of the JOB id is getting stuck as requested in the origanl case.

Thanks