Skip to main content

Storage Node with Deduplication data growth

Thread needs solution

I have a question re how the storage node retains data. Current backup plans run backups weekly and I chose to retain only 2 backups. All are Windows 7/10 x64 PCs.

Right now the storage node is still chewing through space (3.48TB of 3.82TB used) after each machine has backed up at least once and most have backed up twice. Some of the machines are showing 3 backups with the oldest having a trashcan icon when click on "show backups" in the list of machines backed up to the storage node.

So the storage node is growing and likely to reach its max disk space but I'm not sure if this is because of normal data growth or because the storage node is keeping more than 2 backups in a sort of "recycle bin" and that is eating up data.

Also, how does one get statistics from the storage node of what each machine is using for data? The reported "size" of each machine in the storage location view seems rather small. Is there a web service or anything I can query on the storage node? The SQL DB for the AMS doesn't seem to hold any data about the storage node.

thanks

Pete

0 Users found this helpful
frestogaslorastaswastavewroviwroclolacorashibushurutraciwrubrishabenichikucrijorejenufrilomuwrigaslowrikejawrachosleratiswurelaseriprouobrunoviswosuthitribrepakotritopislivadrauibretisetewrapenuwrapi
Posts: 22
Comments: 3800

Hi Pete,

>> I'm not sure if this is because of normal data growth or because the storage node is keeping more than 2 backups in a sort of "recycle bin" and that is eating up data.

To physicall reclaim the space from backups which were removed by retention rules there is a location "compacting" task regularly performed (default=on Sundays at 3:00AM), but from your description it looks like it makes sense to try running it manually via acrocmd tool. See below sample (assuming the tool is run on ASN locally):

acrocmd compact vault --name=[location name] --host=127.0.0.1 --credentials=Administrator,[password] --service=asn

If it doesn't help and you need to investigate the issue further, then please contact our support team for assistance.

>> Also, how does one get statistics from the storage node of what each machine is using for data? The reported "size" of each machine in the storage location view seems rather small. 

The size shown should be for unique data which belongs to particular machine, however it may not show these values correctly (it's a known issue related to per-archive statistics). Also the common (deduplicated) data size cannot be calculated, since it's re-used by many backups. Therefore to estimate the effectiveness of deduplication you need to look at the entire location statistics which was added in the latest update for Acronis Backup 12.5 Update 1 (build #7970). See Backups->Locations->check Details for the location after selecting it in the grid:

dedupstats.PNG

Thank you.

frestogaslorastaswastavewroviwroclolacorashibushurutraciwrubrishabenichikucrijorejenufrilomuwrigaslowrikejawrachosleratiswurelaseriprouobrunoviswosuthitribrepakotritopislivadrauibretisetewrapenuwrapi
Posts: 22
Comments: 3800

In the next update (Update 2 coming later this year) we will also add per-archive information to show the original data size which was backed up in each backup - the "Size" column will be replaced with "Original data size" one. Still the effectiveness of deduplication can be checked only from entire location statistics, rather than per-archive one.

Vasily, thanks for the great explanation in both posts. They answered my questions.

I will update my ABA version as I was not on build 7970.

I also looked over the file system on the storage node and saw how each machine ID existed there as a directory name and the size of each dir matched what the web console reported for the size of unique data on each machine. Still, it would be great if there were some way I could programmatically query either the SQL DB or the storage node to get this info you're going to add in update 2.

I used to have a number of great automated reports that emailed myself every day with summary info on various things like this on 11.5 (without storage node or managed vault). It'd be great to do that again. It's tough to maintain the 200+ machines we have scrolling through only the GUI.

Thanks as always for your great help. good job Vasily!

Pete

frestogaslorastaswastavewroviwroclolacorashibushurutraciwrubrishabenichikucrijorejenufrilomuwrigaslowrikejawrachosleratiswurelaseriprouobrunoviswosuthitribrepakotritopislivadrauibretisetewrapenuwrapi
Posts: 22
Comments: 3800

Hi Pete,

Thank you for the feedback. For custom reporting options I'd recommended you to also check the following thread where I discussed scripts to be used to retrieve backup statuses from Acronis Management Server (for example save all activities for past X hours in a text file).

Thank you.