Comprehensive data protection for all workloads
Post Reply
JamesCornell
Lurker
Posts: 2
Liked: never
Joined: Mar 19, 2014 7:34 am
Contact:

Instant Recovery Query

Post by JamesCornell »

I have a file server which has 8 (quite large) SAN LUNs (and an operating system disk which is stored in a datastore) attached to it to provide our users with their documents/archives etc. Because the sum of these disks is around 1.9TB we have broken down the job into 4 backups. The main reason for this is if we need to restore a disk (from an off site tape) we do not need to restore a huge VBK file for one disk which is potentially quite small (200-350gb).

The other reason for separating the jobs is because our backup is performed over a 100mb link to an offsite repository, if there is an issue with the link or the job fails the backup will retry but it starts from the beginning of the disk which can be hugely frustrating and can have a serious impact to our backups.

It has happened in the past that the task (when it was one job) would get to 96% of the 9 disks, fail and then start again from scratch.

The backup is broken up into four tasks using the disk exclusion tab in the job setup wizard

Job 1 = C:\Operating System Disk & 1 Data LUN
Job 2 = 2 Data LUNs
Job 3 = 2 Data LUNs
Job 4 = 3 Data LUNs

My question is regarding instant recovery

If I use instant recovery to access files from job 1 it is fine and works correctly as the opperating system disk is present and the server can be started in VMware, however if I try to recover the remaining jobs using the same approach it does not work because the recovered VMware machine only has the data LUNs and no operating system.

I was hoping that Veeam would be able to instantly recover the server using the seperate jobs, however this is not the case and it treats each Job as a seperate VM despite them all being backed up from the same server.

The only way I can see around this would be to alter the backup jobs to include the operating system disk on each job as below

Job 1 = C:\Operating System Disk & 1 Data LUN
Job 2 = C:\Operating System Disk & 2 Data LUNs
Job 3 = C:\Operating System Disk & 2 Data LUNs
Job 4 = C:\Operating System Disk & 3 Data LUNs

But this would mean backing up the 40GB OS drive four times which is a total waste of 120GB of backup space and again is adding more size/complexity/risk of failure to the backup tasks.

Can anyone shed any light on this or make and suggestions?
veremin
Product Manager
Posts: 20270
Liked: 2252 times
Joined: Oct 26, 2012 3:28 pm
Full Name: Vladimir Eremin
Contact:

Re: Instant Recovery Query

Post by veremin »

You're backing up to remote location through the limited, unstable link, right? Have you though about utilizing backup copy job, instead of backup one. The backup copy can survive temporarily network drops, and, with WAN Accelerators in place, it even has resume functionality. Also, this way, you would avoid conducting restore via limited bandwidth.

As to Instant VM Recovery, it's unlikely to work with the several disks spread across different backup jobs.

Thanks.
dellock6
Veeam Software
Posts: 6137
Liked: 1928 times
Joined: Jul 26, 2009 3:39 pm
Full Name: Luca Dell'Oca
Location: Varese, Italy
Contact:

Re: Instant Recovery Query

Post by dellock6 »

Also, I don't see the reason to split disks in differnet jobs for restore purposes. When you start a restore you can choose which virtual disk you want to restore, you do not need to restore all of them.

Luca.
Luca Dell'Oca
Principal EMEA Cloud Architect @ Veeam Software

@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
JamesCornell
Lurker
Posts: 2
Liked: never
Joined: Mar 19, 2014 7:34 am
Contact:

Re: Instant Recovery Query

Post by JamesCornell »

Hello Luca

The main issue is that we have our tapes taken off site so a 300GB VBK is a lot easier to restore over a 1.9TB one (in the case of users flapping their arms around regarding document availability) but it also allows us to repeat failed jobs without them having to back up all of the disks again in the job.

The backup is conducted over a fairly stable 100mb business link but the issue we have with using the WAN accelerator is that we do not have enough SAN storage (8TB) to complete a backup onsite and use the copy job to send it offsite .

It is a pity because I waited with bated breath for the new version of Veeam with WAN acceleration and then realised that I could not utilise it :-(
dellock6
Veeam Software
Posts: 6137
Liked: 1928 times
Joined: Jul 26, 2009 3:39 pm
Full Name: Luca Dell'Oca
Location: Varese, Italy
Contact:

Re: Instant Recovery Query

Post by dellock6 »

Thanks James for additional infos.
What is your need for WAN accelerator? To be used in regular backup jobs?
Luca Dell'Oca
Principal EMEA Cloud Architect @ Veeam Software

@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
veremin
Product Manager
Posts: 20270
Liked: 2252 times
Joined: Oct 26, 2012 3:28 pm
Full Name: Vladimir Eremin
Contact:

Re: Instant Recovery Query

Post by veremin »

The backup is conducted over a fairly stable 100mb business link but the issue we have with using the WAN accelerator is that we do not have enough SAN storage (8TB) to complete a backup onsite and use the copy job to send it offsite .
May be you have some disks that can be locally attached to the onsite machine? Or any decommissioned or existing physical server? These are good and cost-effective candidates for local backup repository.

Thanks.
Vitaliy S.
VP, Product Management
Posts: 27055
Liked: 2710 times
Joined: Mar 30, 2009 9:13 am
Full Name: Vitaliy Safarov
Contact:

Re: Instant Recovery Query

Post by Vitaliy S. »

On top of what has been said above, since you're using 100 mb WAN link, then WAN accelerators most likely will not help you, as they were designed specifically to be used with slow, unstable WAN connections.
Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 219 guests