I've got the fully updated SCOM 2016 with latest UNIX/Linux MP and try to discover/add new CentOS 6.0/7.0 hosts.
While the process ain't as straightforward and easy as one would want, after spending some time with all the available resources I've managed to get pass all the obstacles and force SCOM to work with the CentOS hosts. By this I mean I have several hosts already discovered and being monitored properly, with all necessary steps documented and templated.
Both the discovery and agent installation workflows (respectively,
Install-SCXAgent) keep failing 90-95% of time (against same, single host) due to an internal SCOM issue. And since both steps are required in the workflow, I need to run the process many times unit it succeeds.
Both above workflows include a task named
UNIX/Linux Discovery Enumerate Agents Task that keeps failing all the time, providing error
-2130661868 that means (I assume, no description) that the task has timed out.
The task runs simple PowerShell script that enumerates few local folders on SCOM server to find UNIX/Linux Agent installation scripts and outputs a small XML containing the result. Running this script manually reveals no problems whatsoever and the script runs in a blink.
Thus, with the intermittency of this issue I guess that it is not related to the script itself, but rather the ability to execute this script, probably through some internal queuing mechanisms that I am not acquainted with.
Has anyone encountered such problem and knows a fix?
Alternatively, does anyone have a better knowledge on the internals of SCOM queuing/executing PowerShell-based tasks and can provide me with any clues as to where to look and how to diagnose it further?
To separate the issue from the UNIX/Linux discovery itself, I've created a completely separate PowerShell task running a simple
Write-Host "Hello World!" script and have invoked it via PowerShell (
The result is identical, ie. 90-95% timeout chance.