We had an issue in our production environment due to high memory utilization of Java, We have raised a SR and have done all the analysis regarding the issue but could not find the exact root cause.
Then due to some requirement we had to clear the cache from pegatempdir on server, Strangely from then, the nodes went down very less times.
Our query is can we delete the files from pegatempdir manually on weekly basis?
@MeghaR59 please can you specify the Pega version?
For earlier releases there were some known issues. General functionality is described here.
On the whole we do not recommend that you clear the pegatempdir .
Only in very special circumstances would we recommend truncating Pega Caches database tables: (after taking a full backup of PEGARULES and PEGADATA database schemas and stopping Application Server Node(s) / JVM(s) in the PRPC cluster) . This would be done for the PRPC Database cache tables to regenerate the entries and become in-sync. The last step would be to delete the contents of the PegaTempDir (directories “PRGenJava”, “PRGenClasses” and “PegaRULES_Extract_Marker.txt”) for each PRPC Node / Server JVM.
Although not needed in ordinary system operation, you can delete (re-initialize to empty) the static content directories
In a local area network, the temporary files directory can be on any disk, but each node in a multinode clustered Process Commander system must have a distinct, dedicated temporary files directory.
After clearing the directory, PRPC is subject the First Use Assembly of all rules. Therefore following a clear down your users may experience a slower response whilst the rules are extracted from the database and compiled for the first time.
For Pega 8 we recommend the tips listed here.
More info about the use of PegaTempDir is here.
Can you also provide the support Incident number so that we can track the investigation that was carried out?
@MarijeSchillern HI Marije, Thank you for your response.
From your feedback, we understood that truncating pegatempdir on weekly basis is not a suggestable approach. We would also go through the links that you have provided.
Application is currently on Pega v7.2.2. Upgrade to v8.6 is in progress and will be completed by July.
Please find the incident number INC-190005 for the issue on memory utilization.
@ShirishaB4143 thanks for the update. I believe our support department helped with that ticket and the final analysis before closure was that the ‘daassearch;’ activity was causing lots of Requestor Lock Exceptions.
This can cause a requestor to not get cleared from the System as the particular Thread is hung - you were asked to work with your application team to fix this.
I believe that for now this thread can be closed whilst we wait for your upgrade to go ahead.