Comprehensive data protection for all workloads
Post Reply
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Scale-Out Repositories with 2 RD1000s

Post by ahahum »

I'm wondering if this will work like I hope it will.

I have a physical server running Veeam v9 - all components running on this 1 server. Directly connected to iSCSI SAN network. Repository local.

We have 2 RD1000s plugged into this server - they show up as 2 individual USB drives.

Typically, we use 1 disk cartridge per weeknight and 2 for the weekends. There are 2 sets rotated offsite weekly. This is done with CrapupExec (going away obviously).

If I configure a separate repository for each of the RD1000 drives(configured for rotated hard drives) and then add them to a scale-out repository will this allow me to continue my process of only inserting 1 cartridge per weekday and then 2 on the weekend? I am assuming I'll be using the "Data Locality" setting. I plan to use this for Backup Copy jobs for a select few critical VMs(or more as space permits).

Is this going to work like I hope?
Dima P.
Product Manager
Posts: 14396
Liked: 1568 times
Joined: Feb 04, 2013 2:07 pm
Full Name: Dmitry Popov
Location: Prague
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by Dima P. » 1 person likes this post

Hi Adam,

I think that should work if you use both RD’s as regular non rotated extents.
Gostev
Chief Product Officer
Posts: 31460
Liked: 6648 times
Joined: Jan 01, 2006 1:01 am
Location: Baar, Switzerland
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by Gostev » 1 person likes this post

You will need to check the check box in advanced SOBR settings to create new full when previous backup file is not available, this way it should work.
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by ahahum »

Dima P. wrote:Hi Adam,

I think that should work if you use both RD’s as regular non rotated extents.
How will this work if I set them as non-rotated extents when they are in fact rotated? Can you explain how this is necessary and what this is doing behind the scenes?
Gostev
Chief Product Officer
Posts: 31460
Liked: 6648 times
Joined: Jan 01, 2006 1:01 am
Location: Baar, Switzerland
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by Gostev »

Hi, Adam. We don't have RD1000 to test this internally. Can you try yourself and let us know if it works? Thanks!
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by ahahum »

It seems without the rotate media check box on each indiviual extent I get these errors:
Error: File does not exist <path to vbk>
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by ahahum »

I have the settings configured as follows:
Each individual repository is set to be backed by rotated media.
The SOR is set to perform full when required extent is offline.

I am getting some jobs fail with an error about file does not exist.

If I manually do an Active Full then the task works correctly. Any ideas how to get this working as I want?

I don't think it is dependent on the device being an RD1000 - it is really just a removable USB hard drive as far as the OS sees it.
foggy
Veeam Software
Posts: 21069
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by foggy »

ahahum wrote:It seems without the rotate media check box on each indiviual extent I get these errors:
Error: File does not exist <path to vbk>
Do you have the check box in advanced SOBR settings to create new full when previous backup file is not available (mentioned by Anton above) enabled? In this case, new active full should be created on another extent when the required one is not available.

Basically, we haven't tested such a scenario and cannot guarantee that retention will work fine in this case. Why do you want it to be scale-out, btw? Simple rotated repository should work fine for this scenario.
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by ahahum »

Thanks for the reply, Foggy.

I do have the setting configured you mentioned to create a new full.

The reason I'm using scale-out is because on some nights there is only 1 removable disk cartridge present. Over the weekend we put 2 in there so there is enough space to cover the jobs. With 2 individual repositories I'd have to manually balance the jobs' destinations instead of letting the software put the backups where there is a disk present...or at least that was my theory going into this. If it won't work, that's fine...but it seems theoretically that it would.

I suppose I should open a support case to have a more detailed interaction with you guys - thanks for the help so far. If I'm able to get a working solution I'll post back here with those details.
foggy
Veeam Software
Posts: 21069
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by foggy »

Thanks, opening a case would be appreciated. We will also look into this scenario closer.
ahahum
Influencer
Posts: 21
Liked: never
Joined: Jul 26, 2010 8:01 pm
Full Name: Adam Bond
Contact:

Re: Scale-Out Repositories with 2 RD1000s

Post by ahahum »

Case number: 01706342

From the support staff:
The reason it isn't creating a new full backup is because this is on the G: drive which is online and is where the issue is at. That setting won't create a full backup unless the G: drive was offline in this case, which it is not. Something happened with that restore point, and that will fail regardless since there is not a setting for creating a new full backup when the job fails or when a file is missing. In the case of a backup job, this would happen I believe, as it notes on the below user guide page. If any needed incrementals are missing, the job will create a full backup as needed, where a backup copy will not process like this.


So from the support case it seems that because the drive letter is consistent the backup copy job is assuming that the backup files should be there despite the settings enabled to perform a full when previous backup files are unavailable or the extent is offline.

It seems that this behavior could be tweaked in the application so that the same logic could apply to a backup copy job that is used in a backup job (presumably). I don't necessarily want to do a new backup job directly to the removable media so that is not an option for me as a workaround. File copy won't work either because there are more savepoints in my main repository than I can fit on the removable media.

I'm going to test this without the scale-out repository and see if it works correctly as a regular repository backed by rotated media. Updates to follow.
Post Reply

Who is online

Users browsing this forum: Baidu [Spider] and 143 guests