Unable to reconcile: the target version is currently being reconciled against

2495
1
05-13-2016 01:09 PM
Highlighted
Occasional Contributor

Configuration:

Oracle 12c (12.1.0.2 Oct 2015 PSU)

SDE 10.3.1 + Patches

ArcGIS Desktop 10.3.1

Background information:

-Feature classes are registered as versioned, but not with the move to base option.

-Concurrent users edit against default version, there are no explicit versions in the database

-Replicas exist for one-way replication to child database

-Editors save changes every 10-15 minutes, but will have the same editing session open throughout the day

-Editors could be concurrently editing the same feature class, though less common

-Attribute Assistant is used

Issue:

Since the upgrade, several users have reported receiving the following error message. 44648 - Unable to reconcile: the target version is currently being reconciled against  and have lost their edits as a result. It is understood that saving a large volume of edits at the same time to the default version could produce this error message.

Concerns:

This error message is more common since we have upgraded from SDE 10.2.2 to 10.3.1/Oracle11g to Oracle12c. Performance certainly has degraded with the upgrade despite the increased database memory allocation and could account for the increased chance of a long reconcile. However are there other factors that should be considered as to the cause of the error message?

0 Kudos
1 Reply
Highlighted
Occasional Contributor III

It is difficult to say what is happening at your end during edit. You can log a support call to analyze the issue better.

Few suggestion you can try:

1. Save edits and come out of ArcMAP or any other client accessing and editing data.

2. Make sure no orphan connection  or orphan locks

3. Compress the geodatabase to state_id 0

4. start edit again

Please note that suggested version editing workflow is to create child version of default as QC/QA version and from  QA version create editor versions to edit. This will help you avoid any data loss as you mentioned and also have better management of data quality. This is just a suggestion.

0 Kudos