UNSOLVED

JoeO2

updated

14 years ago

J

JoeO2

34 Posts

0

775

February 27th, 2013 07:00

Esx 4.0 update 2 host initiator mode keeps changing to 1 in VNX5500

We are about to migrate our VMWare Esx 4.0 update 2 hosts to EMC VNX5500 Storage array from Netapp.The hosts are connected to both arrays via Iscsi.I noticed that some of the  Esx host initiators have a failover mode of 1 in the VNX instead of 4. I've changed them to 4 several times but they keep changing back to 1. This also appears to be causing the luns connected to these hosts to be trespassing to one sp resulting in only 1 sp owning all the luns (about 26). We set the Esx hosts failover path policy to FIxed path. EMC support recommended rebooting all the ESX hosts as per primus 291503.I am wondering if anyone else has encountered this since it will be very disruptive for us to reboot our 16 ESX hosts that host over 200 guests.

  • GearoidG

    251 Posts

    369

    0

    Posted February 27th, 2013 07:00

    Suspect its emc262738

    Do you have a HLU 0 defined and have you rescanned after it was defined (you may need to rescan several times for it to work)

    Regards

    Gearoid

  • JoeO2

    34 Posts

    369

    0

    Posted February 28th, 2013 06:00

    Geroid,thank you for your response.What is HLU 0 ?

  • RRR

    6 Operator

    •

    5739 Posts

    369

    0

    Posted February 28th, 2013 06:00

    A LUN is a Logical Unit Number, an address. In the Clariion and VNX world an HLU is a Host Lun Unit (number). Inside the midrange arrays virtual disks are called LUNs and to name them LUNs to the hosts could cause confusion, so they're called HLUs: Host Logical Unit (number)s.

    HLU 0 is a non existing LUN which will be displayed to your host when no real HLU 0 is actually there for your host.

    When you did add HLU 0 to a LUN in a storage group, HLU 0 represents an actual LUN on which you can store data.

  • JoeO2

    34 Posts

    369

    0

    Posted February 28th, 2013 07:00

    Hi RRR, Thank you for your response. I know what a lun and lun id are and the difference between the storage array and host perspectives. I just wanted to ensure Gearoid was not referring to something else.We created a 1 TB lun with lun id 0 in the vnx,then created a datastore in VMware esx 4.0 on that lun with lun id 0 to use up all the available space for data. We did a rescan and all our 14 VMware Esx 4.0 hosts can access that lun.Are we supposed to change the lun id of this datastore and create a small lun of 1.6GB with lun id of 0 according to the article or does it not matter that we have data on lun id 0?

  • GearoidG

    251 Posts

    369

    0

    Posted February 28th, 2013 08:00

    You are correct it does not matter if you store data on it and 1tb is fine

    However remember you should rescan several times as otherwise the hlu will not properly take

    See Vmware KB http://kb.vmware.com/kb/2018802

    Gearoid