Database marked as suspect when bringing up a clustered server

We have a 2 node clustered server (active/passive failover) that when it
fails over (forced or otherwise) or if we just reboot the server, when it
comes up, some of the user databases come up marked as suspect. Which
databases get marked varies each time. If I execute an sp_resetstatus on the
database it tells me that nothing was changed because the flags are fine. I
then run a restore database with recovery it tells me that it completed
successfully and processed 0 pages in 0 sec.
I then go into the cluster administrator, take sql server offline and then
bring it back up and the databases will be ok. We're looking for some sort of
explanation of this behavior.The SQL Server error log shows: Error 9004, Severity: 21, State 10
An error occurred when processing the log for database .....
As far as stopping and restarting SQL Server, I'm doing it from the Cluster
Administrator, go to the Active Resources folder and take SQL Server offline
and
then bring it back online. And then the database(s) will be available.
"AnthonyThomas" wrote:
> What does the SQL Server error logs say when the system is coming online and the databases are marked suspect?
> I'd check the SQL Server cluster resource dependencies. When you say that after you've repaired everything and recycle SQL Server, how are you doing that? Are you just restarting services, taking cluster resources offline, taking the entire resource group offline, or using SQL EM to stop and restart the services?
> Sincerely,
>
> Anthony Thomas
>
> --
> "tim" <tim@.discussions.microsoft.com> wrote in message news:6B363DFC-293D-46F4-9F5E-CAAB0A669B33@.microsoft.com...
> We have a 2 node clustered server (active/passive failover) that when it
> fails over (forced or otherwise) or if we just reboot the server, when it
> comes up, some of the user databases come up marked as suspect. Which
> databases get marked varies each time. If I execute an sp_resetstatus on the
> database it tells me that nothing was changed because the flags are fine. I
> then run a restore database with recovery it tells me that it completed
> successfully and processed 0 pages in 0 sec.
> I then go into the cluster administrator, take sql server offline and then
> bring it back up and the databases will be ok. We're looking for some sort of
> explanation of this behavior