Failover not starting SQL Service

[John]

I have a SQL 2005 cluster, 2 node, running on 2003 Server, all patches/
sp's up to date. It was working fine until yesterday. We installed
Reporting services a couple of days ago and installed iscsi initiator
so we could use some storage from NAS yesterday. The failover ot B-
node went fine but took longer than usual, maybe 4 minutes vice about
1. When we tried to fail back it failed on the SQL Server service 3
times, then i manually tried a bring online and it worked. So last
night I failed over to B-node and same deal failed 3 times, I manually
had to bring online. Going back to A-Node I had to actually start the
SQL service in MMC console to get it to come online. Where can I start
troubleshooting? The log in SQL server isn't giving me much, Event
logs has a generic error thats KB articles say is related to TEMPDB
not having space, but there is plenty. Any ideas are much appreciated.

John

[Geoff N. Hiten]

First, you shouldn't be running Reporting Services on a production cluster.

Did you bring the iSCSI resources up on all nodes, promote them to clustered
resources, and test their failover before using them?

It may be that the iSCSI devices are not visible on all nodes or that the
drive letters are in conflict with clustered disk resources.

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

"John" <jcorb...@gmail.com> wrote in message
news:b5640838-3715-4600...@j19g2000yqk.googlegroups.com...

[John]

I am removing the Reporting Services form the cluster - read some
articles and now understand. Dev team install doc's said to put it on
the prod DB....as you can see I am not a DBA by trade.

I did create the iSCSI LUN, created it as a resource on Node A then
went to B, setup the iscsi info in the initiator, and when it failed
over the first time I could see the drive. There is no data on that
drive yet - just there for testing/config right now.

Is the ERRLOG.x the only error log file SQL writes?? I looked at it
last night and it didn't reveal a whole lot....

On Jan 13, 9:19 am, "Geoff N. Hiten" <SQLCrafts...@gmail.com> wrote:
> First, you shouldn't be running Reporting Services on a production cluster.
>
> Did you bring the iSCSI resources up on all nodes, promote them to clustered
> resources, and test their failover before using them?
>
> It may be that the iSCSI devices are not visible on all nodes or that the
> drive letters are in conflict with clustered disk resources.
>
> --
> Geoff N. Hiten
> Principal SQL Infrastructure Consultant
> Microsoft SQL Server MVP
>

> "John" <jcorbin...@gmail.com> wrote in message

>
> news:b5640838-3715-4600...@j19g2000yqk.googlegroups.com...
>
>
>
> >I have a SQL 2005 cluster, 2 node, running on 2003 Server, all patches/
> > sp's up to date. It was working fine until yesterday. We installed
> > Reporting services a couple of days ago and installed iscsi initiator
> > so we could use some storage from NAS yesterday. The failover ot B-
> > node went fine but took longer than usual, maybe 4 minutes vice about
> > 1. When we tried to fail back it failed on the SQL Server service 3
> > times, then i manually tried a bring online and it worked. So last
> > night I failed over to B-node and same deal failed 3 times, I manually
> > had to bring online. Going back to A-Node I had to actually start the
> > SQL service in MMC console to get it to come online. Where can I start
> > troubleshooting? The log in SQL server isn't giving me much, Event
> > logs has a generic error thats KB articles say is related to TEMPDB
> > not having space, but there is plenty. Any ideas are much appreciated.
>

> > John- Hide quoted text -
>
> - Show quoted text -

[Geoff N. Hiten]

SQL Server also writes to the system application log. That may be more
informative, especially in cases where it fails to start.

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

"John" <jcorb...@gmail.com> wrote in message

news:38ecb977-abf3-4728...@p8g2000yqb.googlegroups.com...