Header Ads

  • Breaking News

    Query Administration - Part 2 - Tests to run after killing a PS Query

    Once a PS-Query is killed, I observed that if the configuration is improper, then it leads to certain issues. So, it is important to run perform certain validations at least once after configuring the environment. With the help of several resources, I have listed below some of the important tests. Specifically, while performing upgrades or while creating new environments it is absolutely necessary to perform these tests.
    1. Once a query is killed from the query administrator page, verify that it has got killed within approximately 1 minute.
    2. If a timeout value is specified, ensure that a query running beyond the timeout period gets auto-killed within 1 minute of its timeout time. 
    3. When time out value is set to 0, the query should never get timed out.
    4. In PeopleTools > Utilities > Administration > Query Administration > Settings verify that a value for global timeout is set.
    5. Verify that the query gets timed out as per the value set in the permission list. If a user has access to permission lists with different timeout values then the query gets timed out as per the greatest value set in those permission lists.
    6. Verify that the query stats GUI, reflects the number of times a query is killed.


    7. In PeopleTools > Utilities > Administration > Query Administration > Executing when logging is disabled, ensure that a row is still written to PSQRYTRANS record.
    8. Make sure that rows are not written to the PSQRYTRANS record when queries are ran from something other than query manager and query viewer.
    9. For every query executed verify that a row is written to PSQRYEXECLOG record.
    10. Ensure that EXECDTTM field in PSQRYEXECLOG record reflects correct time. 
    11. The KILLEDREASON field in PSQRYEXECLOG reflects the appropriate reason, 'K' or 'T'.
    12. Make sure that the query monitor will not kill a process that is running something else – ie the process id in the row is a process id for something other than PSAPPSRV or PSQRYSRV.  Also, make sure that if the process id points to a PSAPPSRV or PSQRYSRV that is running a different operation it doesn’t kill it. Details on these services are available in my previous post - Information on PSAPPSRV, PSQCKSRV, and PSQRYSRV, PSQCKSRV processes and their impact on PS Queries.
    13. Ensure that the query monitor cleans up orphaned rows.
    14. Ensure that one domains query monitor does not affect queries in another domain.

    No comments

    Please refrain for marketing messages and unnecessary back links.

    Post Top Ad

    Post Bottom Ad