AnsweredAssumed Answered

SDE Compression and named versions

Question asked by cdd109 on Sep 2, 2015

There is much discussion about compression and I am hoping someone can clear up an issue I am having. We are using versioned editing. Each user has a version that they have created under their own connections. We were having an issue with compression ending states being high - around 40. Many of the versions were old and some were stale. So I rec/posted and deleted the versions. end state was 1. Then I asked everyone to create a new version for themselves.

 

The problem with this is that not everyone used the same name in creating their new version so many of the datasources in their maps became broken. It was a hassle, but the performance gains of the near-full compression were notable and worth the trouble. However, now that users have recreated their versions my state count is going up again after compression. Even if I reconcile the versions first, I can't get below 10 ending states - we have 10 versions.

 

I would like to script the deletion and recreation of the versions as part of an automated workflow, but since the users own their versions, I would have to have a connection as each user to be able to do that. Is that correct? I have the same issue with rebuilding indexes and analyzing datasets...I need to do it under the data owner connection - is that correct?

 

Should I just force all versions to be owned by DBO? Am I missing something that will allow me to achieve what I want without users having to constantly recreate their own versions? Is it normal that each version will account for at least 1 end state after compression even if it is first reconciled?

Outcomes