Good morning, all -

Two node SQL 2008 cluster attaching to SAN storage. Dual path software in
place on both nodes. On a failover attempt (planned or unplanned), the
cluster controller node will not allow clustered disks (including quorum
resource) to failover unless a complete reboot is performed. For some reason,
not releasing the disk resource.

Anyone have any insight with this one before?

Thanks much

I have seen this in early versions of x64 Emulex drivers for Windows 2003.
Make sure you have latest drivers, etc for the HBAs or iSCSI NICs. Do not
team iSCSI NICs. Make sure the SAN has multi-hosting set to on so it
supports clustering.

Finally, do a cluster validation test. That should tell you exactly what is
not working and maybe why.

--
Geoff N. Hiten
Principal SQL Infrastructure Consultant
Microsoft SQL Server MVP


"GaTech_Geek" <GaTechGeek (AT) discussions (DOT) microsoft.com> wrote

that sounds like a SCSI PR3 reservation/release issue here. That are mostly
related to HBA and or MPIO DSM drivers.
As Geoff posted it is important to check if you are running a supported
version of firmware, drivers, framework... Most is mostly regulated by the
target storage vendor.

If all that is already up-to-date how does looks your validation report ?
The storage section wpould be the interesting part here.

PS : Has this ever worked on that SAN or this is this the 1st W2K8 cluster
on it ?

Regards
Ramazan


"Geoff N. Hiten" <SQLCraftsman (AT) gmail (DOT) com> wrote