High Server Loads Since Latest Release

[t9clkclnr]

this just started happening today. cpsrvd-ssl is causing my servers load to increase constantly..

Anyone experience this today. I'm on current.

[t9clkclnr]

WHM 11.15.0 cPanel 11.17.0-C19434
CENTOS Enterprise 4.6 i686 on standard - WHM X v3.1.0

[VeZoZ]

I can confirm this

Seen it on 2 different servers now spawning two cpsrvd-ssl processes using 1 CPU each. Ended up killing the processes as they had been running for over 45 minutes. Checked where they were coming from and it was within cPanel directories so have no reason to believe it was a malicious process or anything.

[Dathorn_ADT]

I've seen this on 6 of our servers so far. This only started after the most recent STABLE update. You'd think by the time STABLE got around to being updated that these problems wouldn't exist.

[VeZoZ]

Just killed one that had been running for 2.5 hours...

Has anyone tried sending in a ticket to cPanel about this?

[sneader]

Same problem here, with cPanel 11.17.0-R19434 (just updated last night)

I have opened Ticket #233328 with cPanel.

Update: Not hearing anything from cPanel for a while, I had my data center open a ticket, to get this escalated. Their cPanel ticket number is #233358

- Scott

[cPanelTodd]

If anyone is experiencing this issue, please submit a ticket at https://tickets.cpanel.net/submit/in...eqtype=tickets so we can get this resolved for you.

Please put 'cpsrvd load issue' in the subject of the ticket so it's easier to spot.

Thanks!

[sneader]

Nick answered my ticket with the following:

I have installed new cpsrvd-ssl and cpsrvd binaries on your system. These should work around the library memory destruction bug that was being encountered.

Unfortuntely, the fix only lasted for an hour... now I have TWO of these processes running, both @ 100% CPU.

18862 root 25 0 100 61:51.96 0.5 18116 10m 2368 R cpsrvd-ssl
24947 root 25 0 100 9:34.63 0.5 18116 10m 2368 R cpsrvd-ssl

With 4 CPUs, that's still only 50% CPU and load average is at 2.00, so it's not too bad... leaving it run in hopes it helps the staff at cPanel resolve this.

- Scott

[Dathorn_ADT]

Thanks for the updates. A publicly available fix would certainly be appreciated as well.

FWIW, I'm still seeing this same issue crop up on our servers. Some new ones, some that have already experienced it.

[LBJ]

We updated 3 of our servers to the latest stable yesterday and we've seen cpsrvd-ssl run rampant on all 3. There's never been an issue prior to the update.

[cpanelnick]

We updated 3 of our servers to the latest stable yesterday and we've seen cpsrvd-ssl run rampant on all 3. There's never been an issue prior to the update.

LBJ, do you tend to leave whm windows open?

[LBJ]

LBJ, do you tend to leave whm windows open?

Not usually.

Since yesterday's update on 3 of our servers, 2 were subsequently accessed only to add an account on each and then logout. The other was open for an extended period during an apache update.

All 3 had the problem though.

[cpanelnick]

Not usually.

Since yesterday's update on 3 of our servers, 2 were subsequently accessed only to add an account on each and then logout. The other was open for an extended period during an apache update.

All 3 had the problem though.

EDGE has been updated to include a work around for this problem. At this point EDGE is probably much safer then usual because we just built STABLE so it works out to be STABLE + a big merge of bugfixes and very little new functionality.

[LBJ]

EDGE has been updated to include a work around for this problem. At this point EDGE is probably much safer then usual because we just built STABLE so it works out to be STABLE + a big merge of bugfixes and very little new functionality.

Thanks for that, but is the workaround included in Edge different to...

Unfortuntely, the fix only lasted for an hour... now I have TWO of these processes running, both @ 100% CPU.

18862 root 25 0 100 61:51.96 0.5 18116 10m 2368 R cpsrvd-ssl
24947 root 25 0 100 9:34.63 0.5 18116 10m 2368 R cpsrvd-ssl

We're happy to move to Edge if it definitely fixes the problem. Have you found it to be a definite, permanent fix in production environments?

[cpanelnick]

Thanks for that, but is the workaround included in Edge different to...



We're happy to move to Edge if it definitely fixes the problem. Have you found it to be a definite, permanent fix in production environments?

11.17.1-EDGE_19717 appears to fully resolve the issue. We should know for sure in 48 hours, however it is extremely likely that it is solved.