ON reports successful monitoring of KVM-HV's, although it fails due to ulimit being hit
|Assignee:||Vlastimil Holer||% Done:|
|Category:||Drivers - VM|
|Target version:||Release 5.4.4|
|Affected Versions:||OpenNebula 5.2|
If you have a KVM-Hypervisor in your setup, that has reached, or is about to reach the ulimit, OpenNebula will probably fail trying to do the run_probes, but instead of throwing an error, it will simply report that the machine was successfully monitored.
While trying to open a remote shell manually, you'll get the Error:
Resource temporarily unavailable
But in the oned.log, all you just see is basically something like
Wed Jun 21 12:36:54 2017 [Z0][InM][D]: Host node18.blafasel (23) successfully monitored.
Of course, the quick fix is to increase the ulimits on the Hypervisor to more suitable values, nonetheless, it should be expected for OpenNebula to report an issue when trying to remote-execute run_probes