Hi all,
I am having one problem about the processing cube in SQL 2005 AS.
when cube runs i get this error in log file:
File system error: Error opening file; \Program Files (x86)\Microsoft SQL Server\90\Shared\ASConfig\msmdredir.ini is not a disk file or file is not accessible.
How can i solve this!!
thanks
Does any one have experienced it before with SSAS! how can i resolve lock issues on this file!!
thanks
|||This file contains the port numbers used by the SSAS instances on the server. I've never heard of a "locking" issue with it so I'll make a couple general suggestions.
First, try cycling (bouncing) the server to see if that clears the issue. If that doesn't work, try temporarily disabling the SQL Browser service. If this second action clears the problem, I wouldn't recommend leaving that service disabled, but at least you'll have a starting point for tech support.
Bryan Smith
|||Yes, it's real... I've logged it with PSS and have had it now confirmed as a bug introduced in SP2.
While it revolves around named instances, scheduling the processing to happen at different times per instance may not work for you.
From the testing I did prior to logging the call, I found two issues that were introduced with SP2. Together they make the issue very very painful.
1) When you process a database when SP2 is installed, the AS Instance requires Read/write access to the msdmredir.ini file. Prior to SP2 it didn't. I could even make the file read-only, and the processing would take place successfully.
2) the ini file gets updated approximately every 30 sec PER AS INSTANCE INSTALLED. You can see this by checking the timestamp of the file. So if there are two named instances running, it will update 4 times a min (but not every 15 sec, as the instances have different start times etc). Stopping one or more of the instances will reduce the frequency of the update.
Note that the behaviour described in point 2 above did not start straight away, but only after the first time I tried to synchronise a database from one instance to the next. My thought was that I could do all the processing in one instance, then sync it out to the other instances, and avoid the locking issue. But once the behaviour starts, it doesn't go away thereafter.
So, it means that even if you're only processing from one instance, this issue may get you!! If you're lucky and your ini file isn't continuously updating, scheduling each instance to a different time may work.
Regards,
Andrew
