OpenVZ Forum


Home » General » Support » Performance impact after patched kernel on centos6
Performance impact after patched kernel on centos6 [message #53129] Sat, 13 January 2018 19:06 Go to next message
viadck is currently offline  viadck
Messages: 2
Registered: January 2018
Junior Member
From: *red-88-13-137.dynamicip.rima-tde.net
Hello,

Las weekend we updated our OpenVZ nodes to the latest patched CentOS6 Kernel RHEL6 042stab127.2. After updating the kernel, we are seeing some big load spikes on a couple of nodes.

All vps servers running on our nodes are centos6, have directadmin as control panel and are managed by us for our clients. We have seen the following strange issues:

- On one node we had to disable clamav on all vps servers because of huge CPU spikes. All of the vps servers would have clamav running simultaneously, consuming huge amounts of CPU and driving the node load above 200 and making almost everything inaccessible.
- Directadmin runs a nightly cron at 0.10 which rotates logs, generates statistics etc. On this same node we have had to change the cron time inside each vps because load would spike above 100 when this process was running on all vps servers simultaneously.
- On another node, we have seen some big load spikes when csf firewall runs a directory check.

This is what we have been observing during this week. The normal load doesn't appear to have changed on our nodes after the new kernel was installed, but it appears that some processes are causing huge load spikes.

Anyone else has seen strange things happening on their centos6 nodes after installing the latest kernel?

Thanks in advance for your time!

Regards
Re: Performance impact after patched kernel on centos6 [message #53146 is a reply to message #53129] Wed, 17 January 2018 15:32 Go to previous message
wishd is currently offline  wishd
Messages: 12
Registered: June 2017
Junior Member
From: *interserver.net
I have many openvz6 (and a few openvz7) nodes. I've seen most impact on the openvz7. Not much on openvz6, but I did find a directadmin install datataskq was using very high CPU - however the system itself was a bit out of date. A da update + yum update was the solution in that case.
Previous Topic: Can't SSH to container
Next Topic: 3.10.0-693.11.6.vz7.40.4 Out of memory on one Server
Goto Forum:
  


Current Time: Fri Aug 17 17:25:03 GMT 2018