PDQ Deploy background service has extremely high memory usage
My Deploy background service was using 14.5 GB of memory yesterday. I restarted it to clear it, hoping it might be a one-time glitch. But this morning the service is already back to consuming 5.7 GB. Has anybody else seen this kind of thing?
For context, I recently moved our free version database from a shared server to my local workstation, as I'm now the only person who uses Deploy. My machine also has a Central Server client of PDQ Inventory. Whether these are related, I'm not sure.
A similar item is listed at https://support.pdq.com/hc/en-us/community/posts/211688527-High-Memory-Usage-and-Memory-Leak with some recent activity, but my issue is with everything sitting idle.
There's not much in the event log, but there was one entry that seemed a little odd:
Thread was being aborted.
System.Threading.ThreadAbortException
ServiceThread: AdminArsenal.PDQDeploy.RepositoryTestThread
GetHashCode: 765445
Now: 1/4/2018 2:59:28 PM
started: 1/4/2018 2:51:54 PM
HResult: 0x0030 (19)
at AdminArsenal.PDQDeploy.RepositorySettings.TestUnusedFiles()
at AdminArsenal.PDQDeploy.RepositoryTestThread.Execute()
at AdminArsenal.Service.ServiceThread.TimerElapsed(Object sender, ElapsedEventArgs e)
It seems like this thread was stuck. Even though I wasn't running the console the background service was testing the repository files. Maybe that's normal, but since the DB recently changed machines it may not be happy with how some/many of the file moved around.
Any ideas?
Thanks!
Comments
I accidentally answered my own question but waited a couple of days to ensure I'd found the issue.
When I installed the new PDQ Deploy instance on my computer I got a new-to-me prompt about enabling the Repository so that I could use it as a system variable. This seemed handy because almost all of our software is in one place and it would save me browsing around to it.
When I launched Deploy last week looking for an error or glitch to help me troubleshoot I noticed a new little alert in the corner: "148088 unused files in the Repository." This message, combined with the Event Log entry above, pointed out that the Repository setup isn't just a variable; it's also an active scanner as part of its optional cleanup routine. That scanner was hammering my system because of how many non-PDQ files we keep in our software network share.
I deleted my Repository entry and it reverted to the default (and empty) location on my local hard drive. The PDQ Deploy service is now happily parked at its normal ~30 MB of memory.
I meant to include a couple images should somebody run into this same thing in the future.
Hey Alan,
I have turned this into a support ticket, I'll be contacting you soon via email.
Any chance that this is still the case?
We have a lot of unused files in our PDQ repository also and often have high memory and cpu usage.