Whenever the MongoDB is activated immediately after Deadline Repository installation, a process called Local Security Authority Process (LSAP) peaks at around 90% CPU usage, making the Database computer nearly unusable. Workers and Monitors will rarely connect to it due to timeouts.
When the MongoDB Database Server process is terminated in Task Manager, the LSAP process decreases CPU usage immediately and the computer goes back to normal.
I’ve tried installing Deadline Repository versions 10.1.18.5, 10.1.17.4 and 10.1.16.9 -same results. On the first and third versions noted, I tried installing them with a manual download of the latest MongoDB version 4.2.17 -same results. There are no Workers or Monitors running on any of the farm computers. Windows 10 Pro fully updated.
Any suggestions will be greatly appreciated. Thanks in advance!
Please let me know if there is a fix, we are facing the same issue. Issue started randomly after vacation, now checking if a windows update will fix it and after that checking for new deadline version…
I just got this notification. Hopefully you’ve already solved it, but if not, I hope this helps.
After checking the MongoDB logs, it showed an IP of a computer on the network that was constantly trying to connect to the server and failing, due to either having an older certificate or bad credentials. That was driving the CPU usage up.
I removed the PC from the network and it solved the issue after a server restart.
I recommend you check the MongoDB logs to see if there’s a request being rejected and what it’s IP is to identify the computer. If not, I presume you could turn off all your PCs on the network then turn them on one by one, until you see a change in CPU usage.
Again, mainly check that all Deadline installs are running the same version as the server and that you’re using the correct certificates and credentials.
We end up shutting all machines down and restarting the server machine, and then it is all good.
It would be great If this issue was solved. It’s a big pain in the ass every time.
Thanks for reaching out on this thread, this issue has present itself in the past due to two possible causes.
One of the clients in your farm is spamming the database hence causing the DB machine CPU utilization high
LSAP is detected a malware by Windows defender and Windows is blocking it hence using a lot of resources
For the rouge client machine we need to troubleshoot like below:
Open an administrative command prompt on Windows machine which is hosting the Database and run below:
netstat -p tcp -b
If you are running DB on Linux then run below:
netstat -tlp | grep mongo
Share both the outputs. Also share the DB logs form [database]/mongo/data/logs/ share the latest from here.
Reach out to Microsoft on the fix. I researched on the web and seems like this link has the fix in it. I think you should open a ticket to Microsoft just to be sure.
We’re running Deadline 10.1.20.3 and ran into high CPU usage on the MongoD as well.
After increasing the vcpu on the VM we have running I was able to access the system, barely though… it was spiking at 95%+ on 6 cores and 16Gb of RAM.
After treading through this topic, I figured it might be many connections to the DB… so I ran the following:
Rhis showed me a huge list of connections… but one machine was popping out with 300+ connections! I logged on to this client, ran systemctl restart deadline10launcher.service and the usage of the CPU dropped instant! … Down to 3% usage. As it should be.
