STATUS_CLUSTER_CSV_AUTO_PAUSE_ERROR not fixed with KB2848344

we still receiving following errors when dpm 2012 sp1 tries backup vms on cluster nodes.

cluster shared volume 'volume6' ('cluster disk 6') no longer available on node because of 'status_cluster_csv_auto_pause_error(c0130021)'. i/o temporarily queued until path volume re-established.

software snapshot creation on cluster shared volume(s) ('\\?\volume{xxxxxxxxxxxxxxxxxxxxxxxxx}\') snapshot set id '{xxxxxxxxxxxxxxxxxxxxxxx}' failed error 'hrerror(0x0000139f)(5023)'. please check state of csv resources , system events of resource owner nodes.

we running following configuration.

dpm 2012 sp1 following update
-kb2822782 (update rollup 2 sp1)

windows server 2012 patched following hotfixes added.
- kb2838669
- kb2848344

ibm ds3524 latest firmware. ibm x3650 server have latest firmware , drivers.

using dual port ibm (emulex) 8gb fiber channel hbas latest firmware , drivers.

ibm has confirmed configuration good.

i have confirmed odx drivers show 3 running following
get-itemproperty hklm:\system\currentcontrolset\services\fsdepends -name "supportedfeatures"

 
have tried disabling odx option on each cluster node running following , evening rebooting nodes.
set-itemproperty hklm:\system\currentcontrolset\control\filesystem -name "filtersupportedfeaturesmode" -value 1

 
have tired reenabling , trying again.
set-itemproperty hklm:\system\currentcontrolset\control\filesystem -name "filtersupportedfeaturesmode" -value 0

any appreciated having been trying resolve these issues 3 weeks , nothing seems fixing problems.

seeing event 5120 error code of status_cluster_csv_auto_pause_error may expected , can safely ignored in situations.  means clustering knew of software snapshot, software snapshot deleted.  clustering resynchronizing it’s state on view of snapshots.

one scenario dpm may delete software snapshot, , volsnap marks software snapshot deletion… volsnap doesn’t delete snapshot right away.  clustering detect software snapshot still exists , needs handle appropriately…  volsnap later goes , deletes software snapshot (which marked deletion).  clustering notices software snapshot knew of went away , needs resynchronize it’s view of snapshots.

think of clustering getting surprised un-notified software snapshot deletion, , telling various components within clustering (dcm, csvfs, etc…) need resync views of snapshots.

so in general, should worried if see lots of 5120’s error code of status_cluster_csv_auto_pause_error.  sign clustering in need of resync’ing it’s state snapshots.

we in process of publishing official guidance on issue...

Windows Server  >  High Availability (Clustering)