HPE Storage Users Group

A Storage Administrator Community




Post new topic Reply to topic  [ 4 posts ] 
Author Message
 Post subject: Windows 2012 R2 MPIO Load Balancing Policies
PostPosted: Sun Nov 26, 2017 2:30 pm 

Joined: Sat Nov 25, 2017 1:04 pm
Posts: 3
Issues with the Microsoft DSM and 3par peer persistence 3.2.2 MU4

Two node hyper V cluster 2012R2

One node reporting all volumes as "08 Paths, Round Robin with Subset, Implicit Only"
Hovewer the other node is reporting some volumes as "08 Paths, Round Robin, Symmetric Access"
If i manually override this with mpclaim -l -d command I get the following LB policy:
"08 Paths, Round Robin with Subset, Symmetric Access"

so whats up with this, and what is the difference between:
"08 Paths, Round Robin with Subset, Symmetric Access"
"08 Paths, Round Robin with Subset, Implicit Only"


Top
 Profile  
Reply with quote  
 Post subject: Re: Windows 2012 R2 MPIO Load Balancing Policies
PostPosted: Mon Nov 27, 2017 12:47 pm 

Joined: Tue Nov 18, 2014 5:13 am
Posts: 17
hello,

We had a similar MPIO setting discrepancy case in some ouf our hyper-V clusters , and it turns out Peer Persistence has an addditionnal prerequisite : on both Arrays, the vlun for the replicated volumes MUST HAVE an identical LUNid !
I have been putting lunid in auto for years in my commands createvlun MY_CSV auto set:MY_HOSTSET, thinkingthe lunid being quite an useless parameter in modern OSes...


To fix that, Since we can't perform a removevlun on the passive side as long as the CSV is in the cluster (the hyper-v Cluster service puts a SCSI reservation on every path of the CSV), I ended up creating a new VV for each faulty CSV, the new ones having the correct configuraiton for PP. the I moved my VMs online with Hyper-V Storage Migration, removing the CSV from the cluster once its is empty, and so on to the next one ( I had 28 dubious CSVs)


here is a bad one, with LUNId 2 on the active and LUNid 3 on the passive, right before I destroyed it :

Code:
8440dsimg cli% showvlun -v CSV-SAS-PROD-D-01
[...]
VLUN Templates
Lun VVName            HostName          -Host_WWN/iSCSI_Name- Port     Type
[b]  2 [/b]CSV-SAS-PROD-D-01 set:Hset_CLUHYSAS ----------------       --- host set
---------------------------------------------------------------------------
  1 total                                                                 
8440dsimg cli%

8440batc cli% showvlun -v CSV-SAS-PROD-D-01.r
[...]
VLUN Templates
Lun VVName              HostName          -Host_WWN/iSCSI_Name- Port     Type
 [b] 3[/b] CSV-SAS-PROD-D-01.r set:Hset_CLUHYSAS ----------------       --- host set
-----------------------------------------------------------------------------

and checking the cluster doesn't lock it, becore I removevv it :

8440batc cli% showrsv CSV-SAS-PROD-D-01.r

no reservations found
8440batc cli%


Top
 Profile  
Reply with quote  
 Post subject: Re: Windows 2012 R2 MPIO Load Balancing Policies
PostPosted: Tue Nov 28, 2017 7:24 am 

Joined: Sat Nov 25, 2017 1:04 pm
Posts: 3
Thanks for reply,

However this is not the case here, they have the same lunid


Top
 Profile  
Reply with quote  
 Post subject: Re: Windows 2012 R2 MPIO Load Balancing Policies
PostPosted: Wed Apr 11, 2018 6:19 am 

Joined: Wed Oct 22, 2014 10:22 am
Posts: 7
Peer Persistence requires the source and target VVs to use an identical *WWN*, not LUN ID.

The source and target VVs must be exported to the same host from both arrays. Using ALUA, the paths to the source volumes are presented as "active", while the paths to the target volumes are presented as "standby".

If source and target VV WWNs are not identical, you can use setvv –wwn <WWN> <VVName> to change (the target) VV to be the same as the source VV.

NOTE: Changing VV WWNs is disruptive to I/O and requires the volume to be unexported from the host before changing the VV’s WWN.

When changing VV WWNs the following should be kept in mind:

    1. Operating Systems that use the VV WWN as identification may need additional steps to correctly recognise the LUN, once its WWN has changed. For example, VMware uses the NAA..<diskwwid> for multipathing. If the WWID is changed VMware may see the updated VV WWN as a separate volume or snapshot. In this instance the inventory may have to be reimported for the correct operation of cluster DRS.
    2. Volumes that are part of a Remote Copy Group cannot have their WWNs updated whilst part of a group. The group must be stopped, volumes removed from the group, have their WWNs updated and then added back into the group, before restarting the group.
    3. For Peer Persistence to function, the (Remote Copy) source and target volumes must have the same WWN. Target volumes created automatically during the Remote Copy Group setup process share the same WWN by default. Volumes created manually during Remote Copy Group setup do not and will need to be updated to match.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 4 posts ] 


Who is online

Users browsing this forum: Google [Bot] and 19 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group | DVGFX2 by: Matt