Discussions specific to the VMware vSphere hypervisor
Post Reply
VEBC
Novice
Posts: 9
Liked: never
Joined: Mar 09, 2015 10:56 am
Contact:

BE2014 migration and lots of queries!!

Post by VEBC »

Hi,

I am seriously contemplating VEEAM but had some queries which I wonder if someone could address with some "real world" answers... I apologise if this seems long but basically I am putting down many frustrations I am currently having!

I currently have a physical backup sever attached to our fibre fabric and am thus using SAN Transport to backup (With good results 8000-12000 MB/min). I am currently backing up to LTO5/LTO6 Libraries. Deduplication on current product is unreliable!

Restores are currently being performed using NBD and these are slow (3000 MB/min)... The backup server has a 4Gb LACP trunk, and the VMware hosts have 1Gb connections. (I am wondering where the bottleneck is with this currently as I do not believe it is storage and the network does not appear to be being pushed)... Either way I believe VEEAM utilizing hot-add for restore may be a faster solution due to potential use of compression etc.?? Does anyone have any kind of reference results?

I am also unsure of how the logical deployment of VEEAM should occur... I am thinking a VM as the main control centre, and a backup proxy utilizing the physical host for SAN Transport and tape drive connectivity (As it is SAS)?!

Is it possible to replicate vCenter/any VEEAM VM servers?? If so do I need a full VEEAM server at the remote end (In case I lose the "source" VEEAM server?

Does the VEEAM vPower Instant Recovery require the backup repository to be on the same subnet as the vSphere hosts - or does it require any vSphere configuration (VMkernel NICs etc.?)

Can I backup and restore VM's (Full and granular) that are on a separate AD forest (But same vSphere cluster) to the VEEAM server (Shared hosting)...

Can I confirm VEEAM does not require the installation of Exchange Management Tools to backup/restore Exchange?

Can I backup and restore VM's (Full and granular) on servers that are using Microsoft DFS Replication...

Is there any word on the potential of using EMC VNX Hardware Based Snapshots?

Is it possible to do a granular restore from tape via staging area?

Is Windows Server 2012 R2 deduplication supported on backup repository?

In the event of total oblivion of local disk backup repository storage is it possible to push backups that are on tape, back on to a local disk repository (Seeding I guess?!)

Are there any replication limits? Also it is possible to replicate VEEAM data across AD forests? (We have a few offsite areas we support that are separate business units to us, and it would be nice to offer offsite backups to them by backing up local and then replicating to us)...

Also for these remote sites... if we licence their VMware environment with VEEAM as well as our own... what is the difference between performing a full local VEEAM deployment, vs. utilizing remote proxy and remote repository? I would like to be able to use Enterprise Manager to view backup/restore across all sites in case that has any bearing!

Currently I have a separate backup job per service we provide, this is because backups were unreliable when grouped together and a single failure caused the entire job to fail... (grrr)... also Symantec's first help was always to tell us to create a separate job when we had trouble! Also the use of separate jobs enabled us to utilize priorities to ensure running order as when we switched off deduplication we could only run 2 jobs simultaneously (2 tape drives)... I believe however in order to benefit from VEEAM deduplication we need to merge these back because deduplication is performed at the job level?

One of my biggest gripes about BE is GRT... Especially when using deduplication... We originally planned to go D2D2T... However on our 2TB file server, the initial full backup (And subsequent weekly fulls) ran using SAN Transport to deduplication which was quite quick... Then a verify took place which re-hydrated all the data (longer time than the backup LOL), then a GRT catalog operation ran which again re-hydrated the data (17 hours plus)... then of course duplicating to tape caused yet another re-hydration... so now we have GRT off! I see VEEAM uses a temporary in-guest indexer (Brilliant idea)... does anyone have any indication of how long this takes to run a given dataset and also does it run independent of backup jobs (As in if the backup finishes, does it wait for this index before starting the next backup, or does it leave it running and carry on?)

foggy
Veeam Software
Posts: 20535
Liked: 1973 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: BE2014 migration and lots of queries!!

Post by foggy » 2 people like this post

VEBC wrote:Either way I believe VEEAM utilizing hot-add for restore may be a faster solution due to potential use of compression etc.?? Does anyone have any kind of reference results?
Hotadd restores are indeed typically faster due to being performed over the ESXi storage stack (bypassing network stack). You can set up a test virtual proxy and compare restore speeds.
VEBC wrote:I am also unsure of how the logical deployment of VEEAM should occur... I am thinking a VM as the main control centre, and a backup proxy utilizing the physical host for SAN Transport and tape drive connectivity (As it is SAS)?!
The described setup is reasonable and used by many of our customers.
VEBC wrote:Is it possible to replicate vCenter/any VEEAM VM servers?? If so do I need a full VEEAM server at the remote end (In case I lose the "source" VEEAM server?
Yes, it is possible to replicate both vCenter server and Veeam B&R server. Please review this thread for methods of protecting Veeam B&R management server. Basically, you do not need a second full Veeam B&R installation, however a good practice is to make Veeam B&R instance located in the DR site responsible for replication jobs.
VEBC wrote:Does the VEEAM vPower Instant Recovery require the backup repository to be on the same subnet as the vSphere hosts - or does it require any vSphere configuration (VMkernel NICs etc.?)
There are no any special requirements in this regard, provided all the required connections are allowed.
VEBC wrote:Can I backup and restore VM's (Full and granular) that are on a separate AD forest (But same vSphere cluster) to the VEEAM server (Shared hosting)...
Yes.
VEBC wrote:Can I confirm VEEAM does not require the installation of Exchange Management Tools to backup/restore Exchange?
Installation of management tools is not required.
VEBC wrote:Can I backup and restore VM's (Full and granular) on servers that are using Microsoft DFS Replication...
Please review this thread for information.
VEBC wrote:Is there any word on the potential of using EMC VNX Hardware Based Snapshots?
Currently only HP and NetApp storage systems are supported in terms of using storage-based snapshots.
VEBC wrote:Is it possible to do a granular restore from tape via staging area?
Yes, you can perform any type of recovery after copying the backup file back from tape to disk.
VEBC wrote:Is Windows Server 2012 R2 deduplication supported on backup repository?
Yes.
VEBC wrote:In the event of total oblivion of local disk backup repository storage is it possible to push backups that are on tape, back on to a local disk repository (Seeding I guess?!)
Yes.
VEBC wrote:Are there any replication limits?
Nothing that I could warn you about off the top of my head. But still everything depends on the particular deployment.
VEBC wrote:Also it is possible to replicate VEEAM data across AD forests?
Yes.
VEBC wrote:Also for these remote sites... if we licence their VMware environment with VEEAM as well as our own... what is the difference between performing a full local VEEAM deployment, vs. utilizing remote proxy and remote repository? I would like to be able to use Enterprise Manager to view backup/restore across all sites in case that has any bearing!
It's just a matter of management an additional Veeam B&R instance (which can be also added to Enterprise Manager for monitoring). Also please keep in mind that if you assign the proxy/repository role to the same server in both Veeam B&R instances, they will not be aware of that while assigning processing tasks to this proxy (just keep the load in mind).
VEBC wrote:Currently I have a separate backup job per service we provide, this is because backups were unreliable when grouped together and a single failure caused the entire job to fail... (grrr)... also Symantec's first help was always to tell us to create a separate job when we had trouble! Also the use of separate jobs enabled us to utilize priorities to ensure running order as when we switched off deduplication we could only run 2 jobs simultaneously (2 tape drives)... I believe however in order to benefit from VEEAM deduplication we need to merge these back because deduplication is performed at the job level?
Deduplication is performed at the job level, right. However there're also other concerns, so whether to have all the VMs in a single job solely depends on your preferences.
VEBC wrote:I see VEEAM uses a temporary in-guest indexer (Brilliant idea)... does anyone have any indication of how long this takes to run a given dataset and also does it run independent of backup jobs (As in if the backup finishes, does it wait for this index before starting the next backup, or does it leave it running and carry on?)
Depends on the number of files, however typically is pretty fast due to getting data directly from MFT. Starting v8, indexing is performed in parallel with backup, so backups are not put on hold until indexing finishes.

Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 27 guests