1

We have three dedicated SQL 2008 64 bit patched to SP4 (10.0.6241.0) in production. They run Windows 2008 R2 (SP1 and patched) under ESXI 6, and have 32GB ram (24gb max for sqlserver) and 8 vcpu. Each of these servers hosts 800-1200 small (less than 2gb) databases, though there are a few outliers of 3-100GB.

Periodically (7 dumps on one server since late June) we get a stack dump in the error log.

Error: 30087, Severity: 17, State: 5.
Data coming back to the SQL Server process from the filter daemon host is corrupted. 
This may be caused by a bad filter.  
The batch for the indexing operation will automatically be retried using a smaller batch size.

While the stack dump is being created, I'll get about 20 errors from our web servers that connect to the SQL dumping host

Database connection errors: Array
(
    [0] => Array
        (
            [0] => 08001
            [SQLSTATE] => 08001
            [1] => 258
            [code] => 258
            [2] => [Microsoft][ODBC Driver 11 for SQL Server]TCP Provider: Timeout error [258]. 
            [message] => [Microsoft][ODBC Driver 11 for SQL Server]TCP Provider: Timeout error [258]. 
        )

    [1] => Array
        (
            [0] => HYT00
            [SQLSTATE] => HYT00
            [1] => 0
            [code] => 0
            [2] => [Microsoft][ODBC Driver 11 for SQL Server]Login timeout expired
            [message] => [Microsoft][ODBC Driver 11 for SQL Server]Login timeout expired
        )

    [2] => Array
        (
            [0] => 08001
            [SQLSTATE] => 08001
            [1] => 258
            [code] => 258
            [2] => [Microsoft][ODBC Driver 11 for SQL Server]Unable to complete login process due to delay in prelogin response
            [message] => [Microsoft][ODBC Driver 11 for SQL Server]Unable to complete login process due to delay in prelogin response
        )

)

I can't find much on this error 30087 on the web, except for some problems that were resolved in SP1.

This blog post was the closest thing to a suspect, but I can't find that same memory pressure error that the author located in his dump file. In my SQLDump0071.txt file, the only text I can decipher confirms what processor I use, thanks dump file.

[I.n.t.e.l.(.R.).] 
[ .X.e.o.n.(.R.).] 
[ .C.P.U. .E.7.-.] 
[ .4.8.5.0. . .@.] 
[ .2...0.0.G.H.z.] 

I'm pretty sure full text search is to blame, but not sure how to diagnose further. We use full text search on 5 columns in an indexed view, and 1 column in an Activity table for each database. We set them to AUTO repopulate.

  • We daily drop the indexed view on a few dozen database, update base tables, and re-create the view. This prompts FTS to re-index 3k-25k rows per database. If that volume seems alarming to others, we could investigate a more targeted way to update data in the base tables.

  • The Activity table is filled by specific user actions, that are ongoing

I don't have anything written in a SQLFT0096100NNN.LOG file at the time of failure.

Next steps?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.