InfoSphere Optim

 View Only
Expand all | Collapse all

Archive Indexes Not Used by Attunity after Cutover to v11.7 z/OS

  • 1.  Archive Indexes Not Used by Attunity after Cutover to v11.7 z/OS

    Posted Wed December 14, 2022 11:53 AM
    We recently cut over from Optim z/OS v11.3 to v11.7 for both Data Growth and TDM. We migrated Optim objects by making a copy of the v11.3 Directory DB and its contents. We then used the installation upgrade panels to upgrade all the Optim objects in the new, copied Directory DB in DB2 and it becomes the Directory tied to v11.7. Same process we followed when we went from v7.2 to v11.3.

    After cutover to v11.7 and running verification tests, we hit an issue using Optim Connect against our archive collections as it appears Attunity is not really using the index files tied to the archive files in the Collection being queried.

    The collection is being queried via a Cobol program invoking z/OS version of ODBC and going against the Binding definition in Attunity. The SQL is a simple keyed query against a single table containing a dense index. There are about 80 archive files in the collection, so it's decent size.

    We executed the exact same baseline test with same queries pre-cutover in v11.3 and the job ran in less than 5 minutes. We ran this test after cutover and the job ran for 4 hours before I killed it. Looking at the JESLOG and JESMSGS showed it opened all the VSAM index files, but then proceeded to open and allocate *every* file in the collection (these are virtual tape, so it's not quick) instead of just allocating the archive files that match the SQL criteria. I killed the job after sensing it was doing scans on the archive files as opposed to leveraging the index files.

    Wondering if any other users have experienced this issue with Optim Connect ignoring archive index results after a migration to a new version of Optim? And if so, how did you resolve it?

    Appreciate the help, thank you.

    ------------------------------
    Keith Tidball
    Progressive Insurance
    ------------------------------

    #InfoSphereOptim
    #Optim


  • 2.  RE: Archive Indexes Not Used by Attunity after Cutover to v11.7 z/OS

    Posted Mon January 09, 2023 05:46 PM
    Edited by System Fri January 20, 2023 04:13 PM
    Some updates on this issue: 

    1. I was able to isolate the root cause as Optim wanting *all* of the archive files in a collection to be freshly imported into v11.7. Importing just some of the files won't work and Attunity will still run scans on all the archive files in the collection. 

    2. Optim Support has confirmed this is a bug with the v11.7 software and APAR PH51686 has been opened. Collections created prior to v11.7 which have archive files with indexes are missing some new metadata values created as part of v11.7.

    3. As mentioned in bullet #1 the workaround is to manually re-import all the archive files contained in the Collection freshly into Optim. These must be done one at a time listing out each file and its associated index file name. There is no bulk import tool nor any method of using wildcard values (*, %) to pick up many files at once. This is a very slow & tedious process which I had to go through to get our Archive Collections into a usable state. Once the APAR is available, this workaround should be moot. 

    At least now this is documented in the event any others Optim users run across the same issue.

    ------------------------------
    Keith Tidball
    Progressive Insurance
    ------------------------------

    Original Message