Re: Clustering Windows 2008

Thanks Overstear for the reply...
I found out our EVA supports Windows 2008 but in the OS type on the EVA I
had Microsoft Windows as the host type and I did not see the Microsoft
Windows LH in the list. LH stands for LongHorn and there is an EVA
controller code update that is supposed to change that to Microsoft Windows
2008 (I have not applied that yet). Anyway, once I selected "LH" for OS
type, that error went away.

The first error was resolved by using the GUI to show hidden devices and
disabling that Teredo device. For some reason doing it by command line
would not work for us, but GUI did.

Now the cluster is up and I am just trying to figure it out....
Thanks again.
Jeff





"OVERSTEAR" <OVERSTEAR@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:42DE06FF-EF33-45D0-A363-C3AC42A167CA@xxxxxxxxxxxxxxxx
I can't help you with the first error but I can explain the second error.
Windows 2008 uses Persistent SCSI reservations. The SCSI Reserve command
and
the SCSI Persistent Reserve command are specified by the SCSI standards.
Servers can use these commands to prevent HBA ports in other servers from
accessing the LUN.
This prevents accidental data corruption that is caused when a server
overwrites data on another server. This is why you can add your disks to
both
nodes before the cluster is configured. The Reserve and Persistent
Reserve
commands are often used by clustering software to control access to SAN
Volume Controller virtual disks (VDisks).
If a server is not shut down or removed from the server cluster in a
controlled way, the server reserves and persistent reserves are
maintained.
This prevents other servers from accessing data that is no longer in use
by
the server that holds the reservation. In this situation, you might want
to
release the reservation and allow a new server to access the VDisk. If you
ever had a node in a cluster hang and not failover to other nodes in the
cluster this will help by letting a different node force access to the
disk
and allow failover.
Now for the bad news your SAN must support SCSI reservations for Windows
2008 Failover Clustering. There are no exceptions. Have a new EMC array
that
supports SCSI reservations and have built 3 Windows 2008 failover
clusters.
Let me say W2008 Failover Clustering is leaps and bounds better than MSCS.
In my co-location we rent SAN disk from the provider and they don't
support
persistent reservation and I am stuck using MSCS on window 2003. In case
you
care my new EMC array will be installed in our co-location on Friday and I
will be migrating to Failover clustering A.S.A.P.


"JJP" wrote:

Hi, we are trying to make a simple file share cluster with a couple of
Windows 2008 64-bit servers. The Step By Step Guide from MS is not
nearly
as good and detailed for 2008 as it was for 2003.

I presented 2 LUN's to both servers (which seems incorrect to begin with,
but that is what the documents says) and both servers see the LUN's, MPIO
is
installed. So....

I run the Validate a cluster tool.

First error it gives:
"Found duplicate IP address, blah blah blah...." This is the Teredo
Tunneling Pseudo Interface. I found a previous post and ran the command
"netsh interface teredo set state disabled", the server said "Ok"
(whatever
THAT means) and it still errors on the validation. I rebooted the
potential
nodes and it still errors on the validation.

Second error it gives:
"Validate storage supports SCSI-3 Persistent Reservation commands"
"Failed to read drive layout of Cluster Disk 0 from node NodeNameB,
status
21"
"Cluster Disk 0 does not support Persistent Reservation"
"Cluster Disk 1 does not support Persisent Reservation"
So does that have to do with the fact that I presented both LUN's to both
potential nodes without cluster installed? Usually I would not do that
but
the step by step guide says 2 LUN's must be presented to the nodes. Then
again it is trying to test the LUN's from both nodes so I don't know.

This is frustrating, especially because even if we get past these errors,
all the other tests that didnt run because of these first errors will
probably just create more errors.

Thanks in advance for any help.





.

Relevant Pages

  • Re: File Server on R2 with Clustering
    ... Windows 2003 enterprise is the O/S. ... Install Cluster service on both servers. ... Use the File Server Migration Tool to migrate the data from the standalone ...
    (microsoft.public.windows.server.clustering)
  • RE: DHCP Cluster
    ... the server cluster function is integrated in Windows ... Microsoft CSS Online Newsgroup Support ...
    (microsoft.public.windows.server.sbs)
  • Re: Looking for Server Reduction
    ... Another 2 Servers, Windows 2003 Std. ... 64Bit Running IIS. ... Example (Windows 2003 Cluster) ... MVP - Windows Server - Clustering ...
    (microsoft.public.windows.server.clustering)
  • Re: How to make 2 servers fail-over?
    ... NT4 had an "Enterprise Edition" which has clustering capability. ... 2000 has "Advanced Server" or Datacenter Edition for clustering, and Windows ... MS cluster is considered "High Availability" and not "Fault Tolerance. ...
    (microsoft.public.windows.server.clustering)
  • LSASS.exe slowly increasing in memory and VM usage
    ... I currently have a 2 node cluster with Windows 2000 Advance Server ... Description: 18264: Database backed up: Database: BEDB, ...
    (microsoft.public.win2000.advanced_server)