umut_var
2 Bronze

Deduplication ratio

Hi, Networker shows the deduplication ratio through the "save sets detail report page", But I can not query this using mminfo. I tried "mminfo -avot -r ssid,attrs"  and "mminfo -avq ssid=123.. -S" but deduplication ratio did not show up. Do you know how to query deduplication ratio through command line. Thanks

0 Kudos
5 Replies
bingo.1
4 Ruthenium

Re: Deduplication ratio

0 Kudos
umut_var
2 Bronze

Re: Deduplication ratio

thank you, we dont have ddboost in the system, I think that is why I dont see dedupe statistics

0 Kudos
bingo.1
4 Ruthenium

Re: Deduplication ratio

wrong. The mminfo report will also not show the dedupe ratio for a ddboost save set.

If you use an AFTD as backup device and later use mminfo -q ssid=# -S, the line *ss data domain dedup statistics: "v1: ...." will not show up at all. For curiosity, you might get that information querying the DD for each save set but this is a longer story. In the end the reported data is the same - no matter if you look at an AFTD or a DDBoost volume.

 

0 Kudos
barry_beckers
3 Silver

Re: Deduplication ratio

Might I ask what you are also looking for in that output? To see how well individual backups are deduping yes or no? And then acting upon it if dedupe is not that good?

However this will not give you insights on how much data a client in total is occupying on a DD, that is something we try to achieve using some DPA reports about NW clients, that are to be enriched with data collected from the DD systems through DPA, however results are not yet that positive. Also to be combined with current Dell DPS licensing method that is using front end protected capacity (largest backup in a 60 day timeframe) to determine license costs.

It is already long overdue that Dell makes live easier so to be able to report about actual disk consumption of NW clients so that billing might better reflect this. Clients that have bad dedupe (either high change rate, encryption, compression) should have their backups cost more than clients that dedupe very well. Now one is still mainly stuck with retention time as a cost factor, whereas ideally you would use different billing method as deduplication is a whole different beast altogether, not really fit for using methods of old.

one can imagine for example that if data dedupes very well, that it might not even matter to keep longer retentions as underlying costs with dedupe would not increase as much compared to making backups to tape (or non-dedupe disk based solutions).

0 Kudos
umut_var
2 Bronze

Re: Deduplication ratio

thanks you are absolutely correct

0 Kudos