VMware Cloud Community
DanGew
Contributor
Contributor

Reducing disk space usage of Deduplication Store in VMWare Data Recovery

Hello,

Using VDR, we occasionally need to increase the disk size of the Deduplication Store. Before version 1.2 there was no way to find out how disk space was actually used and therefore we could only presume backups took more space. As this is now possible (is shown by selecting the destination on the configuration tab under Home/Solution and Applications/VMware Data Recovery), we’ve discovered the capacity of the disk was much larger than the deduplicated size (347GB dedup of 450GB capacity).

Last weekend we run again out of disk space on the deduplication store, as there's over 20% of unused capacity this raises the following questions:

1. Why does VDR run out of disk space?

2. Is there a way to reclame / shrink / make smaller /reduce / consolidate … the size of the deduplication store to stop the need of increasing the disk?

3. Any way to decrease the disk size once point 2 was done?

Will someone please enlighten me as to what is happening?

Reply
0 Kudos
4 Replies
Stevester
Contributor
Contributor

I feel your concerns as well. VDR 1.2 is definitely more stable, but i am still seeing issues such as with space. I am not running out of space in my case, but the reclaim process is not removing restore points as it should based on my customer retention policy. This product is still not ready for production based on my opinion.

Thanks

Stevester

Reply
0 Kudos
RParker
Immortal
Immortal

1. Why does VDR run out of disk space?

De Duplication is NOT a guarantee it's a guideline. you may be able to get 50-70% duplication but it will vary. There is really no way to tell before it starts.

Also if you use more than 1 target within the same disk, the deduplication is by volume not disk. So if you add 2 destinations, even though both are in the same VDR drive, each of those desinations are separate for deduplication purposes.

2. Is there a way to reclame / shrink / make smaller /reduce / consolidate … the size of the deduplication store to stop the need of increasing the disk?

People in hades want refreshments with cold cubes to quench their thirst. You can add this to your wish list, but I wouldn't hold your breath on when to expect it. If you want these features, you have to BUY vRanger, esXpress or Veeam. VDR maybe MUCH later before you see these.

3. Any way to decrease the disk size once point 2 was done?

Yeah, if you want to manually move the contents from the previous target to the new one after you create a new one of the smaller size. It's not happening automatically. No way to set these parameters.

VDR is severely lacking, the most important thing as customers is VM Ware NEEDS to fix the problems first, before they start getting fancy....

Reply
0 Kudos
DanGew
Contributor
Contributor

Hello RParker,

Thanks for your reply. However:

  1. VDR shows you to have around 100GB of free space and then runs out of it for a backup which is significantly smaller than that size. I don't think it has to do with a difficult-to-make estimation. BTW, I only use one target so disk equals volume in my case.
    I can imagine that the reported free dedup space is made of the sum of non continuous parts and as VDR needs some continuous space to do its work is can run out of space. But this is only my guess, does someone actually know how it works? If my theory is correct then a mechanism of “defragmentation” is not a nice to have feature but a needed one.

  2. You may say that I'm a dreamer - But I'm not the only one - I hope someday you'll join us - And the world will be as one. I guess your short answer to my question is no, it can't reclaim space.

  3. I fail to see how I should be “moving" the data to another smaller store. The size of the dedup store is made by of the sum of its files, in my case this sums up to the size of the disk. I hope you don’t suggest to recover all VM’s and backup them again;-)

Dan

Reply
0 Kudos
DanGew
Contributor
Contributor

VDR is end of life with 5.1 now out. It's time to close this question.

Reply
0 Kudos