ArcSDE Full Compress - Results in Corrupted Feature Classes

513
1
11-10-2010 05:01 PM
BillFish
New Contributor
We are running ArcGIS 9.3.1.  We perform an ArcSDE compress on a 3-4 week basis. 

We satisfy all the conditions to achieve a Full Compress. 

We have encountered corrupted (versioned) feature classes as a result of a full compress on a

number of occasions.  The Business, F and S tables become out of sync, and hence the fc cannot

be manipulated, viewed, deleted, etc.  I have managed to reconstruct this relationship using 

DBMS tools, but this kind of practice is not sustainable and is big waste of time. 

A full compress was meant to be a good thing.  The catch 22 for us is our organisation edits

heavily in arcSDE with a number of editors and if we put off a 'full' compress for too long the

performance of the database slows terribily. (ie hundreds of thousands of records in the delta

tables, I understand is bad practice).   It seems the pruning done behind the scenes is not going

according to plan! Has anyone encountered the same issue after a full compress

is achieved?  We are close to upgrading to ArcGIS 10.
0 Kudos
1 Reply
JamesSimard1
New Contributor III
I have come across the same issue. If anyone has any insight we would be grateful.

James

We are running ArcGIS 9.3.1.  We perform an ArcSDE compress on a 3-4 week basis. 

We satisfy all the conditions to achieve a Full Compress. 

We have encountered corrupted (versioned) feature classes as a result of a full compress on a

number of occasions.  The Business, F and S tables become out of sync, and hence the fc cannot

be manipulated, viewed, deleted, etc.  I have managed to reconstruct this relationship using 

DBMS tools, but this kind of practice is not sustainable and is big waste of time. 

A full compress was meant to be a good thing.  The catch 22 for us is our organisation edits

heavily in arcSDE with a number of editors and if we put off a 'full' compress for too long the

performance of the database slows terribily. (ie hundreds of thousands of records in the delta

tables, I understand is bad practice).   It seems the pruning done behind the scenes is not going

according to plan! Has anyone encountered the same issue after a full compress

is achieved?  We are close to upgrading to ArcGIS 10.
0 Kudos