arev indexing with UD4.6 (General)
At 03 NOV 2010 10:37:38AM Craig Tildesley wrote:
We have an arev session doing the indexing by calling index.flush for every index in the system one at a time. Periodically (could be hours, days or weeks) it will fall over with a "cannot read record 0 in !file". The file is rarely the same one and there seems no reason for it. The UD server is logging error 30000's in the event log.
Has anyone ever seen this kind of error?
At 03 NOV 2010 03:22PM Jared Bratu wrote:
In the event log details, after the generic windows text there should be a message specific to the error, what does it say? One of the first values in the error message should correspond to an FS error. Having the error message or FS error will help diagnose the cause of the message, please post it to the discussion board.
At 04 NOV 2010 05:17AM Craig Tildesley wrote:
LHReadRecord, Error Code: 1009, File (2), Group 0, , Record ID: 0
At 04 NOV 2010 08:30AM Jared Bratu wrote:
The LinearHash service is reporting a problem opening a .LK/.OV file. If you open the Programmers Reference Manual help file from OpenInsight and lookup FS1009 it will correspond to Error Code 1009.
While your application is actively being used check the "Shared Folders\Open File" list on the server. If you see any .LK or .OV files in the list that means the REVPARAM file isn't directing the clients to use the LinearHash service for file access.
It could be a process on the server itself that is locking the file. Check any backup processes or antivirus programs. Using a tool like the Sysinternals Handle utility may help locate the process.
Finally, it may also be a problem with the file, directory permissions, etc. Verify the LinearHash service is running as the SYSTEM user and that user has Full Access to all files.