I have a Media PC with 4 GB of RAM. I started to notice a Memory Utilization issue in 1.15 where the PMS would use up all the system’s available memory. I recently upgrade to version 1.16, and it looks like the issue still exists. Previous analysis of the logs seemed to indicate an issue with the EPG process. As a workaround, I have been downgrading to 1.14.1.5488 where I have not observed the problem.
One thing I have noticed is that in 1.16, it looks like the server crashed @ 06/26/19 19:26:51, as the PID updated and memory utilization returned to normal. New feature?
I ran the following script to produce the memory utilization logs.
Memory utilization got so high last night that neither the local console nor the remote console (via SSH) would respond. I had to reset with the physical power button.
I’ll be upgrading to 1.16.1.1291 this morning and continuing to monitor.
I believe you need to update to 1.16.2.1297 for the memory correction.
From what I’ve seen in the forum, there is substantial, if not complete, progress.
Was on vacation, so I had a cron job restarting the service at 12 PM (noon) every day while I was away. This automatic service restart allowed the PMS service to run without any issues while I was away.
Upon my return, I removed the cron job, and I upgraded to Version 1.16.3.1402 on July 28, and began to gather statistics again.
The service required a manual restart on the evening of July 30, as the system was all out of memory.
It appears that the service crashed and restarted on the morning of July 30 and the evening of July 31.
The service required a manual restart on the morning of August 2, as the system was out of memory again.
Modified the script to record Resident memory instead of Virtual memory.
Just before the service is restarted, the script fails occasionally to output data to the file, probably because the server is out of resources.
Let system run until 99% of system memory was utilized. I almost was not able to restart the service from the server command line due to lack of system resources (typing at the server console was slow and delayed).
It appears that the PMS service crashed and restarted on its own twice over the past 4 days. This took care of the need to manually restart the service in order to free up memory.