Mirroring DR Test Failover/Failback

  • All,

    As part of our DR testing.....we have to failover mirroring databases and after confirming everything is working properly on the new production side, then we need to failback to original production.

    What are the steps involved in mirroring failover/failback.

    Please share your experience. Let me know if you have any questions or need anything additional.

    Thanks,

    SueTons.

    Regards,
    SQLisAwe5oMe.

  • Make sure the mirroring is in "High Safety" - Synchronous mode, usually it is just in seconds the data is transferred from and to mirror so there is almost no data loss, just go to the properties of the mirrored primary database on the Mirroring section you have to click on manual failover which then mirrored database on the other machine becomes the new primary and once your production is ready and synchronized with the new primary you can do the perform same vice versa on the new now primary to fail over back to old PRD mirrored DB which now becomes primary again, make sure the mirrored server is same performance as your primary if the burden on the database is too high it could bottleneck.

  • I also have two servers set up in a mirror.

    I would like that server1 is always principle unless it goes down. Is it possible to configure automaic failback to server1 when that one is ready to serve a principal again?

    /ole

  • Superdoc (10/25/2012)


    Make sure the mirroring is in "High Safety" - Synchronous mode, usually it is just in seconds the data is transferred from and to mirror so there is almost no data loss, just go to the properties of the mirrored primary database on the Mirroring section you have to click on manual failover which then mirrored database on the other machine becomes the new primary and once your production is ready and synchronized with the new primary you can do the perform same vice versa on the new now primary to fail over back to old PRD mirrored DB which now becomes primary again, make sure the mirrored server is same performance as your primary if the burden on the database is too high it could bottleneck.

    The mirroring already in synchronous mode.

    I was thinking more of the steps as below.

    -Disable all the jobs

    -Failover one database at a time

    -Now, the DR(mirror) becomes the new principal

    -Application team will test to see if everything is working properly on the new Principal

    -Once confirmed, I will fail back to the original production(principal)

    -Then Enable all the jobs

    What do you guys think about these steps? Let me know if I am missing anything.

    Thanks,

    SueTons.

    Regards,
    SQLisAwe5oMe.

  • Looks to be fine to me.

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply