Veeam and ExaGrid

[dcolin]

Hi folks,

We're looking to deploy Veeam in a VMware environment and we'd like to leverage our Exagrid. After much reading, I'm thinking the best approach might be to set up one or more backup proxies in virtual appliance mode, but in this scenario, can those backup proxies write to the CIFS / SMB shares for Veeam that I'd create on the Exagrid? I'm a little unclear on whether or not the proxies have to do their writing via the ESX I/O stack as well as their reading.

If we can write to the Exagrid this way, does it make any sense to have the backup server itself sitting on a physical box versus a VM? I'm a little hesitant to have the backup server itself live inside the very environment we're backing up, but I'm not keen on buying unnecessary hardware either. I suppose a third option could be to put the backup server on our vCenter box.

Thanks for any input!

[Manfriday]

[merged]

Hi,

What is the correct way to set up BR6 with Exagrid?

Would I need to set up a repository and point it at the Exagrid?
Or does the Exagrid become the repository, and I just point the backup proxies directly at the Exagrid?

Thanks!

[foggy]

Hi dcolin, HotAdd for writes is only possible when restoring and replicating VMs - in this case, a virtual backup proxy can use Virtual Appliance mode to populate virtual disks with data on VMFS LUN. Thanks!

[foggy]

Hi, Jason. You could create a backup repository as CIFS share on your appliance to store backups there. Thanks.

[Eamonn]

Hi,

I had previously heard from ExaGrid support (23rd Dec 11) that BR6 is not yet supported? Can anyone comment as to what the outstanding issues are? I was eager to upgrade to BR6 since the hotfixes became available.

Thanks

[foggy]

Eamonn, our solution is completely storage agnostic meaning that you can use any storage as a backup target and there's no need to speak about any sort of VBR support (except in terms of performance). We have never heard of any issues concerning this type of storage devices since the release of v6. Though maybe ExaGrid support guys do know more about this... Could you please provide more details on what you've heard from them as it could be some kind of miscommunication. Thanks.

[dellock6]

Hi,
we have ExaGrid units at our datacenter and we also sell them.
We tested the new B&R 6 saving into ExaGrid machines since v6 release, with CIFS share configured for Veeam 5. No problem at all, and thanks to new Veeam we also saw increased performances. Basically Veeam backup files are the same as in v5, so I cannot see how v6 would not work.
I talked about the v6 support with ExaGrid Europe, basically the "official" support is only a matter of deeper tests before claiming v6 is supported, and this also involves Hyper-V support that at the moment is NOT supported.
But as long as we talk about VMware backups, v6 is supported.

regards,
Luca.

[Eamonn]

Thanks for the info gents - I was a little confused since I was also under the belief that Veeam was storage agnostic.
The only info I got at the time was "Veeam 6 is not yet supported. I have no ETA on supportability." As you mention, (and I also suspect) this may be down to further ExaGrid compatibility testing (and possibly also the training/familiarisation of ExaGrid support personnel with VBR6?)

I now look forward to my v6 upgrade next week

[Eamonn]

With some careful use of storage vMotion I'm considering recreating my LUN's with VMFS 5.0 (to take advantage of the new unified 1MB block size and larger LUN capacities). Does anyone know how this will affect my current Veeam backup's on the ExaGrid ie. will Changed Block Tracking suddenly decide all the blocks have changed on a VM (that was storage vMotioned to a VMFS5 LUN), then back the entire thing up with the net result that I chew up more ExaGrid space compared to if I just stuck with VMFS3.46?

Also, (when setting up the ExaGrid as a Veeam repository) should I be ticking the box 'align backup file data blocks' - ie. does the ExaGrid leverage constant block size for deduplication?

Cheers.

[dellock6]

I'm not sure 100%, but for what I remember CBT data are kept even during a Storage vMotion. Do not know anyway it this is true even when we move vmdk from vmfs3 to vmfs5. Can you test on a smaller VM? My datastores are all upgraded to 5.

Luca.

[dellock6]

I asked to ExaGrid guys, alignment is not needed.

Luca.

[Eamonn]

Thanks Luca. I've created a VMFS5 LUN and storage-vMotioned (online) a single VM over for tonight's test. I'm thinking CBT tracks virtual blocks within the VMDK as opposed to the underlying blocks used in the VMFS?

According to http://kb.vmware.com/selfservice/micros ... Id=1020128 :-

For CBT to identify altered disk sectors since the last change ID, the following items are required:
The host must be ESX/ESXi 4.0 or later.
The virtual machine owning the disks to be tracked must be hardware version 7 or later.
I/O operations must go through the ESX/ESXi storage stack. So NFS is supported, as is RDM in virtual compatibility mode, but not RDM in physical compatibility mode. Of course VMFS is supported, whether backed by SAN, iSCSI, or local disk.
CBT must be enabled for the virtual machine (see below).
Obviously, virtual machine storage must not be (persistent or non-persistent) independent disk, meaning unaffected by snapshots.

For CBT to identify disk sectors in use with the special "*" change ID, the following items are required:
The virtual disk must be located on a VMFS volume, backed by SAN, iSCSI, or local disk. RDM is not VMFS.
The virtual machine must have zero (0) snapshots when CBT is enabled, for a clean start.
In some cases, such as a power failure or hard shutdown while virtual machines are powered on, CBT might reset and lose track of incremental changes. Likewise offline Storage vMotion, but not online Storage vMotion, could reset but not disable CBT.

[dellock6]

VMware KB is saying basically:
- if you use storage vmotion while VM is powered ON, you keep CBT informations
- if you use storage vmotion while VM is powered OFF, you loose CBT informations

So, online storage vmotion will be good for you.

Luca.

[tvvendors]

[merged]

I am looking to getting the fastest throughput rate from my physical Veeam server to my Exagrid appliance. Adding a 10Gb NIC to each is out of the question for now.

My Veeam server has a FC HBA attached to a Clariion CX4 @4Gbps. The Veeam server also has 2 physical 1Gb NICs teamed together using HP's teaming software for backup traffic to the Exagrid appliance. The Team is set to use Transmit Load Balancing with Fault Tolerance (TLB) and Windows reports a 2Gb connection speed. The Exagrid has 2, 1GB NICs each with separate IP addresses.

All the Veeam jobs use the SAN method for backups. Half the jobs are setup to use 1 Exagrid IP address, the other half use the 2nd Exagrid IP address. A lot of the jobs run concurrently. I want Veeam to fully utilize the 2Gb NIC Team during backups and make the backups as fast as possible. Is this the best way to accomplish this?

I have a similar setup and I'm seeing 19mbps - 917mbps throughput. I have received documentation from Veeam but some of it is contradictory to what is in this thread. We will continue to tweak and test and I will post my results here.

[tvvendors]

In the document: Deduplication Hardware tips for using with Veeam Backup 5.x.pdf page 3 recommends "Disable inline data deduplication".

In the document: VMware Backup: Best Practices for using Veeam with ExaGrid page 7 and page 8 Figure 3 recommends "Enable inline data deduplication".

In the document: Exagrid Systems: Using Veeam Backup and Replication Software with an Exagrid System page 9 it says "Enable inline data deduplication".

Which one is the correct setting?

I got the documentation directly from Exagrid support and Veeam support.

[tsightler]

Enable inline data deduplication would be correct. This is Veeam "source side" dedupe which basically means that if Veeam detects an identical block of a fixed size, it will only store it once on the Exagrid. On the other hand, Exagrids dedupe is far more advanced and will dig deeper into this block and dedupe data both within it, and any identical data blocks/data from other Veeam jobs.

[tvvendors]

In the document: Deduplication Hardware tips for using with Veeam Backup 5.x.pdf page 3 recommends "Disable inline data deduplication".

In the document: VMware Backup: Best Practices for using Veeam with ExaGrid page 7 and page 8 Figure 3 recommends "Enable inline data deduplication".

In the document: Exagrid Systems: Using Veeam Backup and Replication Software with an Exagrid System page 9 it says "Enable inline data deduplication".

Which one is the correct setting?

I got the documentation directly from Exagrid support and Veeam support.

Had a call with the Veeam guys today. There is a note in the first document that says "EXCEPT FOR EXAGRID". No conflicts just an RTFM error.

[fuzzie]

We just implemented an Exagrid Dedupe appliance and now i am trying to create a sure backup job for item recovery for exchange. I have in the Surebackup job a domain controller, Exchange 2010 DB Server and 2010 CAS server. When I run the surebackup job the domain controller takes about 30 minutes to come up and I have had it sit there in an up state for 2 hours and veeam never moves on past that vm. As a test I create another backup job of the domain controller to a local drive on the virtual veeam server. After running a test surebackup job to the backup stored local to the veeam server the SB job completed in less than 6 minutes. Does anyone have any ideas on why the slowness or failure to get the job to complete from the backups on the exagrid? Has anyone experienced a similar problem and do they know how to resolve this?

[Gostev]

Make sure your landing zone is sized properly. As soon as backups leave the landing zone, random read performance will become terrible due to all the latency caused by data rehydration.

[Daveyd]

We just implemented an Exagrid Dedupe appliance and now i am trying to create a sure backup job for item recovery for exchange. I have in the Surebackup job a domain controller, Exchange 2010 DB Server and 2010 CAS server. When I run the surebackup job the domain controller takes about 30 minutes to come up and I have had it sit there in an up state for 2 hours and veeam never moves on past that vm. As a test I create another backup job of the domain controller to a local drive on the virtual veeam server. After running a test surebackup job to the backup stored local to the veeam server the SB job completed in less than 6 minutes. Does anyone have any ideas on why the slowness or failure to get the job to complete from the backups on the exagrid? Has anyone experienced a similar problem and do they know how to resolve this?[/quote]

I've run into slowness issues when our Exagrids are actively deduping the share on which our Veeam backups reside and trying to do an Instant Recovery. Mkae sure you are not eating into your landing zone with retention.

[vgeek]

Hi,

I'm looking for people using Veeam and Exagrid and possibly offloading to tape.

How much data are you backing up?
What dedupe ratios are you getting?
What software are you using to backup to tape?
How is the performance?

Thanks in advanced for your help.

JD

[dellock6]

Hi Josue, I've used ExaGrid in the past and some customers are still using it, biggest one is a single EX7000 holding about 18 Tb of backup (6 Tb full backup, 3 weeks retention, forward incremental). Dedupe is about 12x, but I think we could go much higher by extending retention. No vaulting to tape.
I do not have direct access to the system to grab some numbers, sorry.

Luca.

[vgeek]

We are considering purchasing a couple of Exagrid devices to use as Veeam repository and to replicate and would like know other people's experiences using Veeam and Exagrid. Performance, support, etc. Thanks.

[Gostev]

One thing you might be interested to learn is that starting from firmware version 4.7, ExaGrid supports running Veeam Data Mover directly on itself, which is a game changer in terms of performance and reliability. We will support this configuration in B&R v8 (to be released next month). Thanks!

[intekular]

I currently have two 68tb Exagrid grid's at separate locations that replicate between themselves and have to say that the combination of Veeam 7 and the Exagrids is working great for us. I am waiting on the Veeam 8 so that I can upgrade the environment which will allow me to switch things around and increase my retention capabilities. Performance has been great (we have a little over 150 VM's and do weekly fulls of about 7tb every Saturday). Support has been good.

[coopsam]

We use exagrids for our backups. We have 4 in one location which then replicate off to another site to give geo resilience.

Only thing I find with them is they can be slow for synthetic fulls. I think this is caused by them not holding the 7 days in the landing area and the data needing to be re-hydrated for the synthetics to occur.

[coopsam]

Just after a little advice from those who use Exagrids with their Veeam B&R.

Lately we have been having some troubles with servers connecting to a backup share on one of our Exagrids. We have spent a fair amount of time with an Exagrid engineer trying to resolve the issue. One of the changes he has made to the share now means only one job can backup to the share at anytime. This seemed strange to us as we have other exagrid shares and boxes that we have had no issues with that we send several jobs to at once.

I was wondering if anyone had come across similar advice about 1 share 1 job at time with regard to Exagrid or if anyone has come across unable to connect to share with Veeam and Exagrid?

We have also been in touch with Veeam support but sadly the person at Veeam didn't really have any knowledge of Exagrid so was not very helpful.

Thanks

[foggy]

We have also been in touch with Veeam support but sadly the person at Veeam didn't really have any knowledge of Exagrid so was not very helpful.

I would recommend to escalate the case to the person who has better knowledge of Exagrid then.