Abnormal Load average on first vm start from template #9570
-
Hi everyone! Maybe someone has faced with similar problem and can help me. First of all i'm using: When I creating template or registering it from URL, and trying to start VM for firs time, after cloudstack ui shows this template as downloaded on primary storage, load average starts to grow, and instance creating takes too long. It's can take about 40 mins, or even can totaly overload host and ony way to bring in to a health state its reboot. This load average is present on all disk nods where linstor creates related devices. After first instamce creation, all folowing tryes compliting without any problems, and vm's starts immediately. pc utility shows growing number of vgs and lvs processes. Like this: 393650 ? D 0:00 vgs -o vg_name,vg_extent_size,vg_size,vg_free --separator ; --units k --noheadings --nosuffix Nothing suspisies in Cloudstack agent logs, but in Linstor sattelite I have found many reports like this: ERROR REPORT 66C5C167-EF4D9-000006 ============================================================ Application: LINBIT® LINSTOR ============================================================ Reported error:Category: LinStorException Error message: Failed to query 'lvs' info ErrorContext: Call backtrace:
Caused by:Category: Exception Call backtrace:
END OF ERROR REPORT. I have tryed to disable apparmor and firewall, but it's not helped |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
Seting up global_filter = [ "r|^/dev/drbd|" ] in /etc/lvm/lvm.conf, solved this problem. |
Beta Was this translation helpful? Give feedback.
Seting up global_filter = [ "r|^/dev/drbd|" ] in /etc/lvm/lvm.conf, solved this problem.