Smartevent cpu spikes

[ihenock101]

Hi All,

I am writing to seek assistance regarding a fresh install virtual dedicated r81.10 SmartEvent server which consistently shows CPU utilization above 90%. All servers have been running smoothly without any notable CPU-related concerns, except for the SmartEvent server which consistently exhibits CPU utilization above 90%.

Despite having already reached out TAC support, I regret to inform you that their response has been less than satisfactory in addressing this matter effectively. Hence, I am turning to your expertise and guidance in resolving this issue promptly.

Therefore, I kindly request your assistance in diagnosing the root cause of this problem and implementing appropriate measures for resolution.

FYI: top -c is attached 

Thanks,

[PhoneBoy]

A couple of things to point out here:

1. You can see the Java process in particular has a priority of 39 (lowest priority), which means it will back off when anything else needs it.
2. cpsead is used to correlate logs (particularly Firewall "connection" logs)...do you have this option enabled? See issue 4 here: https://support.checkpoint.com/results/sk/sk150452 

[the_rock]

We have customer with this sort of issue and I cant even count how many TAC cases were opened for it and no good solution yet. Literally, when it happens, server gets rebooted and it works until next time it happens and then you repeat the same.

I recall that senior TAC guy once suggested RAM increase, which they did, to 64 GB I believe, but not sure if that actually solved the problem,.

I will follow up tomorrow with one of my colleagues that was constantly involved in it and let you know.

[Chris_Atkinson]

Any clues in HCP or similar or does the environment suffer the issue described in sk181209?

Depending on the environment Java heap sizes were a thing that possibly needed some tuning (sk123417).

[the_rock]

Just checked with one of my colleagues and even though RAM was increased to 82 GB and its DEDICATED SE server, issue still happens, but not as often, so that gives you an idea.

Not being a bearer of bad news, but 100% honest.

[Chris_Atkinson]

We don't have enough context for this to be helpful. Why would you/anyone expect simply adding RAM alone would solve a CPU utilisation problem? 

[the_rock]

Thats exact advice TAC escalation guy gave for our customer while back, as they did not have anything else really left to suggest. I can understand why, as there were sooo many things and debugs done, it was honestly pointless continuing with the exact same troubleshooting.

[the_rock]

Also, as you know, sometimes after so many weeks/months of troubleshooting, hard to really give solid advice lol