Recently there was a database indexing scheduled in our production environment during sunday.
Below are the items handled during indexing
Check all DB index that have fragmentation more than 30% to perform do re-indexing.
Once re-indexing completed, they will perform DBCC checkdb
After the above two activities are performed application server is not restarted. When users started to access the application on the next day there were slowness observed.
On validating the logs there are more number of PEGA0005 alerts triggered which was not logged before this activity. Even some of the out of the box SQL queries which was not timed out previously resulted in time out.
Thus we have restarted the application server after this issue is surfaced as a resolution and the application works fine.
Hence we have a query whether is it mandatory to perform a application restart whenever the above two activities are executed.
Request your response on this item.
***Moderator Edit-Vidyaranjan: Updated SR details***
***Edited by Moderator: Pallavi to update platform capability tags***
Thanks for your comments. The scenario mentioned in the link is different from what we face. Normally we don't restart server for adding new indexes but the scenario which we mentioned is a database indexing job.
DBA has initiated the two activities on sunday and on monday we see lot of queries are timed out, which was not faced before the job.
Also sometimes users were parallely accessing the application when database indexing job is being executed.
Upon reviewing the associated SR, the resolution was as follows:
Gave an explanation to the User that the application restart is not required when re-indexing process is done.
Initially the load on the DB is more because of the indexing process, and concurrently when a query is run ,the response time limit might have exceeded the threshold limit , hence the alerts were observed.
As per Pega suggestion restart is not required, but we are not sure why the application was down after the re-indexing. We have seen more number of Pega Engine queries getting timed out after re-indexing is completed. This is overcomed by restarting the application server. Hence for providing a resolution Pega Team has raised a spin off SR and currently validating the logs. SR No. "SR-C92380"
Posted: 3 years ago
Posted: 19 Feb 2019 11:42 EST
Marissa Rogers (MarissaRogers)
Senior Knowledge Management Specialist