#ExaGrid and #Veeam: the results

0 Flares Twitter 0 Facebook 0 LinkedIn 0 Email -- 0 Flares ×

In the previous article, I set up ExaGrid in conjuction with Veeam to backup the servers in my test environment. Those “lab” is a production site of a customers, who kindly let me use ExaGrid on it.

In this way, data grew and changed during the tests, giving me a “real life” test experience.


The Lab

Let’s briefly repeat the servers we were going to save with Veeam:

As you can see, there is a fair amount of data, mixed guest OS, and different vmware virtual hardware version (so different behaviour regarding CBT).

This is the Veeam backup schema:

From 19.00 to 23.00, alla servers get their backup, with incremental during the week and synthetic full on Saturday night.


Veaam first run

On a Saturday, we had the first run of Veeam Backup. Since no previous backup has been made, this execution created a full backup set.

The job named “BKP-galileo” took the most part of the activity, lasting more than 7 hours, since the file server is a v4 Virtual Hardware VM, so no CBT to came and help.

On Sunday morning, ExaGrid started to deduplicate the backup set, and even if there was no previous backup to compare, the appliance used compression and gained space anyway:

While backup set was 709 GB, the used space was 447 Gb, with a Deduplication Ratio of 1,58:1.


Second day

From the second day, Veeam started its incremental backups using CBT, where available. Backup windows falled down dramatically, as the processing speed of vmdk disks was quite good:

After all the backups were completed, Veeam saved on the Exagrid about 17 GB of incremental backups (identifiable as VIB files) coming from 9 different backup sets:

On the ExaGrid, this new space was again reduced:

RAW backup space (the amount Veeam thought it was using) rose from 708.82 to 727.30 Gb (+ 18.48 Gb), while deduplicated space rose from 447.60 to 456.06 (+ 8.46).

Deduplication Ratio rose also from 1.58 to 1.59.


After Two Weeks

After two full weeks, Veaam and ExaGrid have done their respective assignements:

From this report, we can get some conclusions:

  • The total raw size of Veeam backups after only two weeks has already growed beyond the usable 2 Tb raw space of the ExaGrid device, so deduplication is effectively saving us a fair amount of space
  • As ExaGrid support told me, incremental backups does not allow to save much space, as we can see from the deduplication ration not increasing during week days, rather is get worse in some days. That’s because incremental backups saves new data that ExaGrid cannot compare to any old data it has already saved
  • At every sinthetic full backup, deduplication ratio jumps forward substantially. We use the default 14 copies retention setting from Veeam, so we can expect to reach even greater deduplication ratio extending the retention on Veeam. This means we can use ExaGrid to save a longer backup history compared to a simple NAS device.

This is the final status of our ExaGrid device after two weeks:

ExaGrid used also part of the Landing Space for saving backups, but simply waiting the next day, ExaGrid was able to regain all the Landing Space and compress all the data into the Retention Space once again:

As this test is over, I can say ExaGrid appliance is a real game changer in the disk backup market. They have a great set of features combined with a super-ease of management.

Add to the list multi-software support, grid scalability and remote replication, and we have maybe found a new all-around ultimate backup appliance.