LUN Tresspass not working...

if you create a brand new LUN, can you trespass it back and forth between SPs ?

Check your hosts for the preferred path - if the hosts believe that the LUNs belong on SPA, when you trespass the LUN to SPB, it will trespass the LUN back.

Each LUN has a Default SP owner and a Current SP Owner (not for Pool LUNs) and ESX will look at the Default SP owner when it first discovers the LUNs - if SPB was not connected when you set up the hosts they may be confused as to which SP the LUNs belong to.

For Best Practices for setting up iSCSI, please see KB 71615 (master) and KB 41172 for best practice when using multiple SP ports on the each SP.

glen

Thanks alot for responding to the query.

The mulipath being used is VMware NMP, and i think (i need to cross check) the multipath is MRU. So i think the MRU will might be causing this. Also i should check what failover mode is selected for the LUN at clariion.

If i change the failover at clariion or Multipath at the server ESX will it need a downtime or i/o impact.

Can you please suggest on this.

Thanks in advance...

No... We had an unused LUN created at SPB, which was not showing tresspassed (as it was not allocated)

I assigned to ESX and created Datastore (change multipath to Fixed), when i went to Unisphere this LUN also tresspassed to SPA, and is not getting tresspassed back to original SPB.

I checked SPA A2 and SPB B3 are configured as iscsi, and both are seen at VMWare iscsi target list, but as soon as any lun is assigned or gets used it goes to SPA (no luns are staying at SPB)

Moreover when i am trying to create a new LUN using Unisphere, the option to select SP is greyed out and i just see SPA is by default selected choice not sure, what exactly can be done....

I see the Failover mode at the Clariion is set to 1... Can this be issue.

Please suggest....

You may find this useful, it shows all (native) supported configurations.

Bold = Recommended

Blue = VAAI supported

ALUA with native failovermode is supported with Flare 28.704 or higher; therefore you need to have Powerpath on CX3 if you want to utilize ALUA.

Chad wrote a nice article regarding this: http://virtualgeek.typepad.com/virtual_geek/2009/09/a-couple-important-alua-and-srm-notes.html

Mathias

AX4

ESX 3.5 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 5.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 5.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

CX1+CX2;

ESX 3.5 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.0 : not supported

ESX 4.1 : not supported

ESX 5.0 : not supported

ESX 5.1 : not supported

CX3:

ESX 3.5 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1

ESX 5.0 : not supported

ESX 5.1 : not supported

CX4:

ESX 3.5 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

ESX 4.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed” ; Array Failovermode = “4” (Fixed AP only with EMC VSI installed on ESX 4.0!)

ESX 4.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed with Array preferred” ; Array Failovermode = “4” | Host Failovermode = “Round Robin” ; Array Failovermode = “4”

ESX 5.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed with Array preferred” ; Array Failovermode = “4” | Host Multipath policy = “Round Robin” ; Array Failovermode = “4”

ESX 5.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Round Robin” ; Array Failovermode = “4” | Host Multipath policy = “Fixed with Array preferred” ; Array Failovermode = “4”

VNX:

ESX 3.5 : not supported

ESX 4.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed” ; Array Failovermode = “4” (Fixed AP only with EMC VSI installed on ESX 4.0!) | Host Multipath policy = “Round Robin” ; Array Failovermode = “4”

ESX 4.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed with Array preferred” ; Array Failovermode = “4” | Host Multipath policy = “Round Robin” ; Array Failovermode = “4”

ESX 5.0 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Fixed with Array preferred” ; Array Failovermode = “4” | Host Multipath policy = “Round Robin” ; Array Failovermode = “4”

ESX 5.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1” | Host Multipath policy = “Round Robin” ; Array Failovermode = “4”

glen

Thanks Glen, quite lot of handy information, i have already bookmarked it. However the way we have configured is

Clariion CX4 120 using iscsi:

ESX 4.1 : Host Multipath policy = “MRU” ; Array Failovermode = “1”

NMP of VMWare.

So isn't it fine? If yes we are facing these issues even though. Please suggest if there is something i should look at...

Thanks for your help....

Hi Dynamox,

Below is the output i get when i try analyzing SP collect, can this be a reason.

But the LUN's are tresspassed at SPA not at  SPB, neither SPB is being selected while creating new lun.

did you try my test of creating a brand new lun and trepassing it between SPs ..trying to figure out if it's SP related or host triggered trespass.

Yes i did without success. First while creating a new  LUN, i am not getting option to select SPB as owner. It is showing SPA as default owner selected and greyed option. Luckily i had some new luns created earlier by someone which were not allocated to any host, and were somehow created on SPB, when i allocate to host it gets tresspassed to SPA immdeately  (does that means it is host initiated tresspass?) and when i tresspass it back, it doesn't goes back...

Yes i did collected SP collects from SPB, system is not showing any faults except the critical issues about host being registered and logged in only at SPA...

Just to highlight again we have failover mode as 1, VMWare 4.1 with NMP, and MRU as multipath. CX4 120 with flare 30, iscsi.... with two iscsi targets on each SP

Failover Mode 1 – Passive Not Ready Mode

In this mode of operation the non-owning SP will report that all non-owned LUNs exist and are available for access.  Any I/O request that is made to the non-owning SP will be rejected. A Test Unit Ready (TUR) command sent to the non-owning SP will return with a status of device not ready. This mode is similar to Failover Mode 0 with Auto-Trespass disabled.

i think you are missing the point of what i am saying. If you create a brand new LUN, do not present it to any hosts ..can you trespass it to SPB ?  If you can, then SPB is not the culprit and you need to look at your zoning, registration, multipathing.

Yeah you are correct, there were some luns not alloctated to any host which were created and were on SPB, but as soon as i mapped those luns (added to storage group) to the hosts, i went under tresspassed luns and saw those luns as tresspassed to SPA.... No clue....

if a brand new LUN that is not presented to any system cannot be trespassed to SPB ..don't think it has anything to do with your multipathing settings.

if you can't select SP for a new LUN or trespass an existing LUN to SPB that should be an indication something is wrong with that SP. Have you engaged EMC support ? Can you collect SPCollects from SPB ?