Basic Centera Collections Question

 Centera does a single instance at the file level. This is why the number of files is different from what you were expecting


Check your collections folder to ensure that you only have directories in there and no files. As long as the files are gone from the collections then they have been collected and placed on the Centera

There aren't any files in the collection folder.  So you are saying this is down to deduplication?

Yes

On average we see a 16% difference between whats in our Centera and what EV reports as being in the archives 

That worries me a little more then because ours is approx 60% in just 10000 files.  Would you happen to know what determines the duplicates?

I'm going to configure FSA reporting today too..  does that provide any more helpful information?

 The duplicates on a Centera are devised by an algarythim on the centera. When a file is stored on Centera the systems creates an imprint of the item. If any other item is stored and matches the imprint it then does not save the file but returns the first imprint details back to EV In EV this is stored as the SavesetID

Are you using EV8? If yes then most of the SIS will happen on the EV8 side because of the Fingerprint database. This is new to EV8 and provides a high level of SIS. If you are using EV8 then the FSA reporting will provide alot of info on the space savings. If you are not using EV8 then the reporting tool may not provide the info you require because the SIS is happening on the Centera and not on EV and the report only reports on what EV does not what is Single Instanced on the Centera

 Bear in mund i'm doing my storage on 30TB of mixed files for many divisions including engineeing, finance, HR and normal office admins. In my environment I would not expect much Single Instance seeing as it is a highly managed environment

 Have you run the usage report to see what EV see's as archived?

Thanks for the info.  The figure I quoted earlier was from the usage report.  I've just totalled up the TotalCount column in the Collections table in the database and reach the magic number.  Should this pretty much confirm that all is well?

Are you running v8.  Just curious as to whether your 16% was based on v7 or v8 or a bit of both.

 I'm still on EV7.5 SP3. Cant upgrade due to some other issues I'm working with Symantec at the moment but i do plan on moving there once my issues are ironed out

I've been looking through the database tables and looking at the archiving reports for the files in question and I cannot come to any conclusion how the number of files on Centera is so low.  The database tables all look how I expect including the number of clips and number of savesets.  The analysis I've done on the reports suggests to me that the level of duplication on this share wouldn't result in anything like a 60% descrease in the number of files so I really don't know what the Centera does under the hood to calculate this number of files.  The fact that Centera is reporting the 11GB that EV is showing exactly is the only re-assurance I can find that all is well.  I'm going to have to trust that it is just doing something I don't fully understand..