1) Check if any scripts at OS level is scheduled to run.
2) Check Integration Service logs for that time period and check if any session with clean up script is running during that time.
This isn't a standard Informatica process.
Very likely a script is triggered to clean up the cache folder(s) and I can only imagine it has been developed in the past because of very limited disk space for the cache files combined with regular failure of jobs.
It could either be a job triggered on your operating system or a (post session) command in a workflow.
Is this job always running at exactly the same moment or does it start at various times?
Thanks for both replies.
This must be a process run by Informatica server. It runs consistently every 24 hours on each node in our grid with a difference of 2 - 3 seconds. The console log records show that it gets rescheduled for 24 hours after it finished, and the scheduled time corresponds to the time when Informatica services were restarted and finished initialization.
When services finished initialization after restarting I see record in the IS log:
"Persistent session cache file cleanup is scheduled to run on ..."
and it gets scheduled for 24 hours later.
I am positive that it is a process within Informatica server and I would like to know if there is any way to control it.
Thanks for the link but it is not what I am looking for, and I don't have a problem described by the article.
I wold like to know how Informatica PowerCenter server schedules job that cleans up persistent session cache files. The job runs every 24 hours and I am positive that it is scheduled by the server.
The process deletes cache files that are in $PMCacheDir and have default names starting with PM. If cache files have defined names they will not be deleted.
I get it, so you have persistent cache files without defined names?
Generally you want to use persistent cache files when you use the same data in multiple mappings/sessions or multiple times in one mapping/session and it takes a lot of time to cache the data (mostly lookup cache)
It is therefore essential to have a defined name for it because otherwise it is impossible to use the same cached data downstream.
Non-persistent cache files are built at runtime and automatically deleted when a session ends normally. This is not a scheduled process once every 24 hrs or so.
I'm not aware of an Informatica process which cleans up the $PMCacheDir every 24hrs or 48 hrs.
You either have persistent cache files with a defined name which aren't deleted after a session ends normally or you have non-persistent cache files which are deleted after the session ends normally.
In case of crashes of the Informatica processes it can happen non-persistent cache files remain in the $PMCacheDir which you want to clean up manually or by a script which runs weekly or monthly (assuming you occasionally have a process which crashes).
Thanks for the reply and explanation of using cache.
But again, it is not what I am asking for...
I see records in the IS logs that a process to clean up persistent cache files runs and it gets rescheduled to run 24 hours later. Here is an example:
01/05/2022 03:06:34.282 AM Persistent session cache file cleanup is scheduled to run on [Thu Jan 06 03:06:34 2022].
01/06/2022 03:06:34.283 AM Start cleaning up obsolete persistent session cache files.
01/06/2022 03:06:34.607 AM Finish cleaning up obsolete persistent session cache files.
01/06/2022 03:06:34.609 AM Persistent session cache file cleanup is scheduled to run on [Fri Jan 07 03:06:34 2022].
I don't see any property or parameter related to that process in the Admin console and I also don't see anything related to it in the basic startup scripts. So I am wondering if anybody in the community knows if and how we can manage that process.
2 of 2 people found this helpful
Just to follow up on the topic.
I have contacted Informatica Support and the answer was that there is no access to control the process.
If sessions are affected, caching needs to be managed as per the replies and links in the thread.