MJStorage

updated

12 years ago

M

MJStorage

1 Rookie

•

11 Posts

0

3198

July 2nd, 2014 14:00

LUN ID inconsistent

We have 5 node ESX cluster and have presented VMAX LUNs to them.

Out of 5 hosts, one node is having different LUN ID, but all others having same LUN ID

How to make LUN ID consistent on that particular node?

  • PSR2

    21 Posts

    1983

    0

    Posted July 2nd, 2014 15:00

    Here is plan you need to follow:

    Identify the ESX server which we need to start with first
    Migrate all VMs from the first ESX host;
    Put ESX host in maintenance mode;
    Remove the ESX host’s two WWNs from existing InitiatorGroup in storage level;
    Restart the ESX host in maintenance mode and confirm that NO STORAGE IS VISIBLE at the host;
    Create two new InitiatorGroups,with each the WWN of the ESX host being woked on and Enable "CONSISTENT_LUN" option on new IGs;
    Add the same SG to new masking view.
    Create a new PG with the two FA ports to which the abopve ESX host WWNs are zoned to and logged in.
    Create a cascaded IG and add the above newly created IGs with the single WWN of ESX host and enable "CONSISTENT_LUN" optionon new cascaded IG;
    Create a new masking view, with the newly created SG, Cascaded InitiatorGroup and the newly created PG with the two FA ports;
    Restart the ESX host to detect storage from the new masking view;
    Ensure that luns have same lun id from both the paths on this ESX host;
    Ensure that the number of devices/LUNs presented to this host is same as that of the other host nodes in the cluster;
    Disable HA and DRS of the cluster for which the ESX host was part of;
    Take the host out of maintenance mode;
    Restart the host (is this necessary?)
    Manually perform host vMotion for 1-2 VMs to the newly fixed ESX host;
    Manually perform storage vMotion for 1 small VM from one LUN to another LUN;
    Enable HA and DRS for that ESX cluster and make sure that VMs are balanced among each ESX hosts.
  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    1983

    0

    Posted July 2nd, 2014 14:00

    did you use consistent_lun parameter ?

  • M_Salem

    213 Posts

    1983

    0

    Posted July 2nd, 2014 14:00

    Hello mlj1,

    Welcome to EMC Community.

    Kindly find this KB article from EMC. There is an example there for the same issue you have.

    https://support.emc.com/kb/14934

    Hope that helps

    Mohammed Salem

  • PSR2

    21 Posts

    1983

    0

    Posted July 2nd, 2014 14:00

    Check and Enable the consistent_lun option on initiator group level in VMAX array...Also, it is recommended to create cascaded IG within that add each ESX hosts IG which is created with consistent_lun option enabled...

    symaccess -sid <> show < IG Name> -type initiator -detail

    You can't enable that option on the fly as it will not affect to existing LUNs and but to new LUNs...

  • rawstorage

    4 Apprentice

    •

    423 Posts

    1983

    2

    Posted July 2nd, 2014 23:00

    Please refer to the following KB article I wrote for this, very similar to what PRS Reddy has described.

    https://support.emc.com/kb/184306  -How do I correct consistent lun violation?

  • AranH1

    2163 Posts

    1983

    0

    Posted July 3rd, 2014 09:00

    Nice write-up Paul!

  • rawstorage

    4 Apprentice

    •

    423 Posts

    1983

    0

    Posted July 8th, 2014 05:00

    Thanks Aran, Once I have to draft an email more than once on a subject it's usually worthy of a KB