We applied the 4.6.6 patch/upgrade to our 4.6.2 setup (security vulnerabilities from McAfee).
After the upgrade, the McAfee epo 4.6.6. Server service won't start
If you try to login on the web GUI you get the error:
The Local Agent Handler service is not running.
If you try to manually start the above service (the event parser and apache server ARE running) you get:
Windows could not start the service, please check the event logs blah blah blah
Event Viewer shows:
The McAfee service terminated with server specific error:
The only posting I could see was back in Dec. 2012 and nobody responded, so I'm hoping there's new information since then. This seems to be a common problem with EPO in 4.6.x for some strange reason
Can we confirm exactly what won't start? The agent handler and the server service are the same thing, which is apache, but you said that apache is actually running.
Is it the Application Server service that won't start?
There are 3 services that show in in the MMC:
McAfee ePolicy Orchestrator 4.6.6 Application Server
McAfee ePolicy Orchestrator 4.6.6 Event Parser
McAfee ePolicy Orchestrator 4.6.6 Server
The top two are running, the bottom one does not.
Sorry, the apache server is what won't start (that equates to the bottom one)
Look in the ..\apache2\logs folder - there should be various error logs. Can you attach the most recent one here?
So in the meantime, I'd opened an SR with McAfee Gold support.
They couldn't recognize the error code and there were all sorts of java errors in the apache log and it mentioning extensions that weren't updated or something.
Fortunately these are VM and I had snapped them before doing the upgrade.
So Gold support suggested I restore the snaps and retry the upgrade again.
That seems to have resolved the issue
Sorry I didn't have the old logs for you.
Just glad I remembered to snap the VM's.
Okay, something very odd there then - extensions and java errors would point to the orion log, which in turn would point to the application server service rather than apache.
Anyway, glad it's OK now VM snapshots are a wonderful thing...
I ran into this a while back. In both my Prod and Dev environments.
After working with support the issue was traced to a problem where apache had an "unknown" issue with C:\Windows\SysWOW64\KernelBase.dll
The resolution was to backup the newly updated 4.6.6 directories and database of the instance having this issue. Completely uninstall and then re-install 4.6.6 clean and restore the backup.
I also needed to re-install most of the agent handlers due to the same issue on them, but once done, all is working correctly now on 4.6.6.
I just applied the same update to my test envrionment yesterday. Same thing the server service wouldn't start. I then contacted McAfee Platinum Support (yep we use McAfee heavily) and fortunately we had a snap of the VM and backup of the database. We then rolled back everthing and made sure all server tasks were disabled prior of the install. We also ran setup.exe DEBUGOUTPUT=2 so we could each each error as they appeared. Afterward, everything is working fine.
Thanks for the switch (debugoutput=2). I'll remember to use that from now on.
I know, in my case, no scheduled tasks were running (they don't run for like another 2 hours after the upgrade--in my case).
My tasks were not supposed to run during the update either, but somehow disabling them fixed the issue.