metadb question

From: Yura Pismerov (ypismerov_at_tucows.com)
Date: 07/03/03

  • Next message: Steve Howie: "concatenating and mirroring in a Sun Cluster Disk Set" 
    Date: Wed, 02 Jul 2003 21:31:55 -0400
To: sunmanagers@sunmanagers.org

    I've made a mistake and initially create just on metadb replica on a
    metadb slice (on each mirrored disk).
    man metadb says "All replicas that are located on the same slice must
    be created at the same time." So I figured I can not just add one more
    replica to the same slice.
    So the question is, if I can just delete the metadb from one of the
    submirrors and re-create it with more replicas ? Can it be done on the
    fly ? Will the system survive with just one replica for the time I
    re-create them ? Theoretically it should because it is the point of
    having redundant metadb, but I want to make sure.
    Second question. If I do it (re-creating the metadb with more slices)
    successfully on one of the submirrors, I guess I can repeat it on the
    second one, right ? In this case, do I have to do anything else in
    between the procedures so the newly created replicas would get in sync
    with the existing (old one) before I destroy the old one ? Or they get
    in sync instantly upon the creation ?

    Just to clarify, here is what I currently have:

    # metadb
            flags first blk block count
         a m p luo 16 8192
    /dev/dsk/c1t0d0s6
         a p luo 16 8192
    /dev/dsk/c1t1d0s6

    I want to destroy all the replicas on the 2nd partition (2nd disk) using
    metadb -d c1t1d0s6.
    Then I want to re-create 2 or 3 replicas on that slice (metadb -a -c2
    c1t1d0s6).
    Would it work without a problem ? If yes, can I just do the same for the
    first drive or I need to do something else before that ?

    TIA.
    _______________________________________________
    sunmanagers mailing list
    sunmanagers@sunmanagers.org
    http://www.sunmanagers.org/mailman/listinfo/sunmanagers

  • Next message: Steve Howie: "concatenating and mirroring in a Sun Cluster Disk Set"

    Relevant Pages

    • Re: Trying to make metadevices (soft partitions) available after reinstall
      ... >> problem is that I did not use a dedicated slice for the replica when I ... > there before attempting to recreate the metadevice on the same slice. ... metadb: saustall: there are no existing databases ...
      (comp.unix.solaris)
    • SUMMARY: metadb question
      ... Also the replicas will get in sync ... > metadb slice. ...
      (SunManagers)
    • Disksuite metadbs and HA 1.3 question
      ... Configuration: E4000 HA pair running Solaris 2.6, Disksuite 4.1, ... had a requirement for 3 arrays to distribute the metadbs. ... I'd like to delete the metadb replicas on the SSA114 but don't ... o - replica active prior to last mddb configuration change ...
      (SunManagers)
    • SUMMARY: metadb output question
      ... will get the extra flags on the metadb output showing up as expected. ... the system is not performing in a degraded state until you reboot. ... o - replica active prior to last mddb configuration change ...
      (SunManagers)
    • Re: Trying to make metadevices (soft partitions) available after reinstall
      ... >> c0d0s6 before metainit or is it possible that the db overwrites the data I ... Can metarecover find the soft parts even if the metadb ... > can't create metadevices without a replica to store the information in. ...
      (comp.unix.solaris)