1. Community Tip: Please Give Thanks to Those Sharing Their Knowledge.
    If you receive helpful answer on this forum, please show thanks to the poster by clicking "LIKE" link for the answer that you found helpful.
  2. Community Tip: Forum Rules (PLEASE CLICK HERE TO READ BEFORE POSTING)
    Click the link above to access ADSM.ORG Acceptable Use Policy and forum rules which should be observed when using this website. Violators may be banned from this website. This notice will disappear after you have made at least 3 posts.

Same TDPO_NODE name shared with many servers?

Discussion in 'Oracle' started by RecoveryOne, May 16, 2017.

  1. RecoveryOne

    RecoveryOne ADSM.ORG Member

    Joined:
    Mar 15, 2017
    Messages:
    34
    Likes Received:
    4
    Hello Everyone,

    Long time lurker, first time posting here. Have a question regarding TDPO configuration. We have four Oracle boxes each running their own independent databases. No RAC involved. Currently they are all set up to use the same TDPO_NODE name. My understanding the reason for this was our DBA wanted to easily restore NODE_A to NODE_* with ease. Reading the doc at: https://www.ibm.com/support/knowledgecenter/en/SSGSG7_7.1.0/com.ibm.itsm.db.orc.doc/r_dporc_cfg_options.html#r_dporc_cfg_options
    “To restore data from one Oracle server to another Oracle server with Data Protection for Oracle, be aware of the following tdpo_node considerations… The value of the tdpo_node option in the tdpo.opt file on the target Oracle server, must equal the value of the tdpo_node option in the tdpo.opt file on the source Oracle server.” Basically, everything was kept the same between all four nodes.

    The concern I have is the constant attribute changes in the actlog, I see “ANR1639I Attributes changed for node DB_ORA: TCP Name from node_a to node_b…” or vice versa. Especially around the time the full backups kick off for the various nodes. Now I’m not seeing any evidence of the full or archive’s failing; however, wondering if the attribute switching could cause trouble? There are cases daily where one node is doing a full backup, and another comes in to send its archives.

    Would it be best to break out each db into their own name, and then use asnodename for those (hopefully) rare cases of when we want to restore to a different node? Or just reference a different tdpo.opt file all together for that task?

    Thanks everyone.
     
  2.  
  3. DanGiles

    DanGiles ADSM.ORG Senior Member

    Joined:
    Oct 25, 2002
    Messages:
    611
    Likes Received:
    15
    Occupation:
    Sr. Storage Admin
    Location:
    Toronto, Ont. Canada
    You REALLY don't want to do that! I mean, you could, but then you would want to have each server use a different filespace name. If you see errors in the actlog, how would you know which server is throwing out those errors? Also, since each session is under the same node name, you would need to have a very high maxnummp - high enough to let all of them back up at the same time.
    I don't think the ASNODE will work with Oracle. (If I'm wrong, someone will correct me!)
    I use different node names: servername_ORACLE. If the DBA wants to restore the database from nodeA to nodeB, I create a separate TDPO and dsm.opt (and of course a separate dsm.sys stanza). The DBA will point to this alternate TDPO to restore from nodeB.
     
    marclant likes this.
  4. RecoveryOne

    RecoveryOne ADSM.ORG Member

    Joined:
    Mar 15, 2017
    Messages:
    34
    Likes Received:
    4
    As much as I hate to say it, maxnummp really isn't much of an issue. Each box can use 4 threads(sessions) total and the node is set to 20. As to errors in actlog, yes it currently is a pain (However our DBA is really on point saying 'hey this didn't work'). And then that involves tracking down the session number to see which box it is... Also, when the Oracle backup is running its pretty clear which one is sending data. And yeah, they all use the same FS name in the tdpo.opt. As to asnodename found this doc: http://www-01.ibm.com/support/docview.wss?uid=swg21260228 which leads me to believe one could, even in non-RAC configurations.

    This is actually one of my last 'physical' nodes left on our old 6.3 server and from my understanding it has been setup this way for about 10 years. To make matters even more fun its still using a 5.5 tdpo client with a 6.3 baclient, so trying to take the time to update everything and set it up correctly on our 7.1.7. End of the day, want it best we can with out the DBA paging me at 4am saying he needs to run a restore of node_a to node_b!

    Thanks again!
     

Share This Page