> A few days a fail over occurred due to a network outage. After this
> the event log filled with Error 823 messages. After reading books on
> line I got worried since it talks about disk problems. I kind of
> figured out what is going on but have not found a solution yet.
>
> Just before the outage a full text index was rebuilding. During the
> outage instance-1 moved from node-2 to node-1. After this instance-1
> on node-1 is reporting failed logins from the system account from
> node-2, and the windows event log reports Error 823 events from
> sqlserver-instance-1 on host virtualserver-instace-1. Both stop when I
> stop the microsoft search service on node-2. My conclusion is that the
> search service on node-2 is still trying to update the the full text
> index on instance-1. Since the disk with the full text index is now
> owned by node-1, node-2 trying to reach this disk will end in a
> serious i/o error, hence the 823 Errors.
>
> Restarting the full text on node-2 did not resolve the issue. I have
> searched the Microsoft search registry on node-2, but can not find any
> regkey that still points toward instance-1. How do i make the
> Microsoft search service on node-2 get over its close relation with
> instance-1 and forget it ever existed?
>
> A second question: node-2 is writing events to the eventlog witch are
> indistinguishable from sql-instance-1 events from node-1 since it uses
> the same virtual host name and service name. How can I find out on
> which node an event originated?
>

Simon Sabin wrote:
> Hello Hans,
>
> Can I suggest cross posting to the FullText NG. I am sure Hilary can
> answer this.
>
> Simon Sabin
> SQL Server MVP
> http://sqlblogcasts.com/blogs/simons
>
>
>> A few days a fail over occurred due to a network outage. After this
>> the event log filled with Error 823 messages. After reading books on
>> line I got worried since it talks about disk problems. I kind of
>> figured out what is going on but have not found a solution yet.
>>
>> Just before the outage a full text index was rebuilding. During the
>> outage instance-1 moved from node-2 to node-1. After this instance-1
>> on node-1 is reporting failed logins from the system account from
>> node-2, and the windows event log reports Error 823 events from
>> sqlserver-instance-1 on host virtualserver-instace-1. Both stop when I
>> stop the microsoft search service on node-2. My conclusion is that the
>> search service on node-2 is still trying to update the the full text
>> index on instance-1. Since the disk with the full text index is now
>> owned by node-1, node-2 trying to reach this disk will end in a
>> serious i/o error, hence the 823 Errors.
>>
>> Restarting the full text on node-2 did not resolve the issue. I have
>> searched the Microsoft search registry on node-2, but can not find any
>> regkey that still points toward instance-1. How do i make the
>> Microsoft search service on node-2 get over its close relation with
>> instance-1 and forget it ever existed?
>>
>> A second question: node-2 is writing events to the eventlog witch are
>> indistinguishable from sql-instance-1 events from node-1 since it uses
>> the same virtual host name and service name. How can I find out on
>> which node an event originated?
>>
>