What does Failed on rebuilding the indexes - Refresh failed for PhysicalPartition Mean

Guys,

Every weekend I attempt to reindex failes for the aforementioned reason, Refresh failed for PhysicalPartition '1'.  What does this mean, and how can I remediate this...

Dav This is screen with error message
dfr031260Asked:
Who is Participating?
 
Marten RuneConnect With a Mentor SQL Expert/Infrastructure ArchitectCommented:
I suppose a deadlock could occur. You could enable trace flag 1204 and trace flag 1222.

//Marten
0
 
Marten RuneSQL Expert/Infrastructure ArchitectCommented:
does the databases: AccessManager, Events, Trades all exist and none in offline or restoring state?
Do indexes reside on its own disk? Is this disk healthy or full?

//Marten
0
 
dfr031260Author Commented:
Yes to all of your questions...
0
The 14th Annual Expert Award Winners

The results are in! Meet the top members of our 2017 Expert Awards. Congratulations to all who qualified!

 
Marten RuneSQL Expert/Infrastructure ArchitectCommented:
Again: Is this disk .... or full
Is the disk containing indexes full. Can this be the cause?

What does the errorlog say?

//Marten
0
 
dfr031260Author Commented:
Indexes are on the same file filegroup and device as the tables.  There is plenty of space available.

There are no errors in the SQL Server log or the event log

Now there was another job (we are testing LiteSpeed for backups, so we are running both a native backup job and a LiteSpeed backup job), the LiteSpeed backup job is also doing a reindex.  Could it be that two reindexes running at the same time can throw this error?
0
 
dfr031260Author Commented:
We are turning off the Native backup and just going with the LiteSpeed and see if that clears up the problem.  I am going to accept you answer about a deadlock.
0
 
Mark WillsTopic AdvisorCommented:
>> Could it be that two reindexes running at the same time can throw this error?

Yes, that can create problems. If you are doing rebuild then there are various exclusive locks that need to be in place. Even with Enterprise and the ONLINE option there is still a possibility.

anyway, sorry to arrive late, but thought I would chime in...

would still use native backup, just dont do the reindexing as part of the backup. You could create seperate plans / jobs for reindexing and schedule so there is no conflict.
0
 
dfr031260Author Commented:
Totally appreciate the additional idea and information Mark...kudos!

Dav
0
All Courses

From novice to tech pro — start learning today.