FullText catalog status is shutdown?

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

John - 26 Apr 2005 03:29 GMT

Hi,

I have a test development database with a single table, which contains over
2 million rows. To check performance of fulltext vs index query, I enabled
FullText on the database. The fulltext index runs for few hours, after that
event viewer throws lots of (around 500) error messages (all messages are
same) as follow:

Event ID: 3011, Source: Microsoft Search, Type : Error, Category: Gatherer
The transaction cannot be updated in the project <SQLServer SQL0001400006>
queue. File: d:\FTData\SQL0001400006\SQL0001400006.Crwl0.gthr. Error:
8007054e - Unable to complete the requested operation because of either a
catastrophic media failure or a data structure corruption on the disk. .

I have changed full-text file location to another SCSI harddisk and re-index
it, but it gives same error message. I do not believe that it's harddisk
failure as both harddisks are great working condition and never had any
issues. The database has only one table for performance testing. Next day,
it shows FullText catalog as follow:
Status: Shutdown
Item Count 1241600
Size: 66MB
Unique Key Count: 213600.

I also tried to apply same fulltext settings on Northwind, and it works
without any issues. Northwind database full-text works.

I am using Microsoft SQL Server 2000 - 8.00.760 (Intel X86) Developer
Edition on Windows NT 5.2 (Build 3790: ) and Windows 2003 Server.

Does anyone know the error log issues? Does it require any configuration?

Any help is appreciated. Thanks in advance.

Regards,

Mital

Reply to this Message

John Kane - 26 Apr 2005 04:24 GMT

John,
Most likely there is some level of internal data corruption in the FT
Catalog files, i.e., "...data structure corruption...".
You should review the server's Application event log for all MssCi source
events and one or more of them will contain a stack trace that can further
identify the cause of the data structure corruption.

Meanwhile, I'd recommend that you, drop and re-create the FT Catalog and
re-run a Full Population on the same disk drive, if you have no other i/o
errors in the system event log. Also, can you confirm that on your SCSI
harddisk drive that you have at least 15% free disk space? As a lack of free
disk space as well as low memory can sometimes be the source of these
problems. If the corruptions persists, I'd recommend that you drop and
re-create the FT Catalog and re-run a Full Population on another disk drive
and if this fails as well, then open a support case with Microsoft PSS SQL
Server support on this issue.

Regards,
John

Signature

SQL Full Text Search Blog
http://spaces.msn.com/members/jtkane/

> Hi,
>
[quoted text clipped - 12 lines]
> I have changed full-text file location to another SCSI harddisk and re-index
> it, but it gives same error message. I do not believe that it's harddisk

> failure as both harddisks are great working condition and never had any
> issues. The database has only one table for performance testing. Next day,
[quoted text clipped - 17 lines]
>
> Mital

Reply to this Message

John - 26 Apr 2005 06:55 GMT

Hi John,

Thanks for your feedback. There are 5 Information events for MssCi source as
follow:

(Two of them): Master merge was started on
d:\ftdata\sql0001400006\build\indexer\cifiles because more than 500000
documents have changed since the last master merge.
(Three of them): Master merge has completed on
d:\ftdata\sql0001400006\build\indexer\cifiles.

But it deos not looks like any data corruption event log under MssCi source.
Does it write any other source for data corruption?

I have done drop and re-create FT catalog few times with another SCSI
harddisk, but same errors. The development server has around 35% free
harddisk space in both harddisk(10 GB each) and 1.0GB of RAM.

What does Shutdown status means exactly?

Thanks again for your quick response.

Kind Regards.

> John,
> Most likely there is some level of internal data corruption in the FT
[quoted text clipped - 17 lines]
> Regards,
> John

Reply to this Message

John Kane - 26 Apr 2005 21:15 GMT

You're welcome, John
All of those MssCi source events are normal & do not provide the details on
the possible data corruption. You can also review the binary Gather log
files under the folder \FTDATA\SQLServer\GatherLogs\ where you have SQL
Server installed.
You can use the gthrlog.vbs app to read these files. Below is an Exchange KB
article that documents the use of the vbs file as this can be used for SQL
Server / MSSearch as well as exchange, just change the path info:

XADM: Using the Gthrlog.vbs Utility to View Gather Logs
http://support.microsoft.com/?kbid=270058

C:\Program Files\Common Files\Microsoft Shared\MSSearch\Bin
or
c:\Program Files\Common Files\System\MSSearch\Bin

-- another example of use:
cd\
cscript //nologo gthrlog.vbs
drive_letter:<full_Path>\FTDATA\SQLServer\GatherLogs\ *.gthr file

Hopefully, with the above info, and if a gthr file has the errors that
occurred, you can post them here for further analysis!
Shutdown is a PopulateStatus value that indicates that the Full Population
has been shut down.

Thanks,
John

Signature

SQL Full Text Search Blog
http://spaces.msn.com/members/jtkane/

> Hi John,
>
[quoted text clipped - 41 lines]
> > Regards,
> > John

Reply to this Message