SRM5 vs Veeam Bakup and Replication

[zak2011]

Hi,
Recently some vendors came to me and told me there is no real need to go in for VBR6 and all what VBR6 does can be done with Vsphere5 enterprise with SRM5.
Since we are now in the decison making stage and i am trying to convince the need to go in for VBR6, is there any comparison chart or any document for VBR6 vs SRM5.
Since its not possible to go in for both the products, we need something that is reliable and cost effective.
If there is some comparison chart or info between both the products, it would really help.
Thanks!

[Gostev]

Sorry to tell, but these vendors you are talking to are clueless. Sounds like a sales speech of someone who never ever installed neither own nor competitive solution they are so readily discussing, let alone designed and implemented enterprise-ready and working backup and replication strategy (otherwise, they would not make such recommendations).

No comparisons as we don't really compete these day (we are night and day), but plenty of info on these forums. For example:
Veeam vs vmware DR
Is this product any good?
Veeam Replication vs SRM 5.0 replication.

Basically, VMware backup tool is only good for smallest shops and can do most basic backups only (no any advanced functionality, which is the main selling point for Veeam). VMware SRM has very poor functionality for host-based replication and is hugely expensive (plus, you would still have to shop for a proper backup tool - because again, VMware backup is no good beyond smallest shops).

However, Veeam is not a replacement for SRM in case you need to have site failover capabilities based on storage-level replication (replication between similar SAN hardware).

[dellock6]

I agree with Anton, and in fact I have many customer that uses both for different goals (obviously, if they can afford both...).
SRM has host-based replication in its latest version but is basically aimed at small shops, while it works at best with certified storages using hardware replication and SRA drivers inside vSphere. Also, it keeps only the last version of any protected VM, so you would still need a backup solution for versioning and retention policies, offloading of backups, and most of all restore capabilities!!!!

[zak2011]

[merged]

Recently we came across one of our vendors who have told us that going for vpshere5 and SRM5 would do the same job as VBR6 and even better. I have been evaluating VBR6 and quite happy with the product. However after seeing a couple of product reviews of SRM5 and the capabilities of being able to do automatic failover and failback, i am really not sure which product to go for. The requirements of our company would be to have a good backup and disaster recovery solution. Is there any sort of comparison chart between both the products? Also is there any way to get advice on what kind of storage i could go for at the production site. Currently the HP SAN used in the prioduction site is quite old and we are thinking of putting it in the DR site and buying a new SAN for the production site.

[zak2011]

Sorry..i didnt see the replies above. The question i was asking was a repetition.
I personally really like VBR6 a lot, it was just that i needed some thing before me to show to the management and convince them about going in for VBR6.
Are there any storage vendors, or SAN models Veeam would recommend.

[Vitaliy S.]

There is no preferred vendor or SAN model, but you may want to take a look a this poll results, which will give information about the most popular vendor/SAN model used by our community members.

[Gostev]

Our closest production storage partner has been HP, we are currently doing a lot of work together with HP, so this would be particularly good choice and my recommendation going forward. However, generally speaking, we are completely storage agnostic - as the poll linked above proves.

[dellock6]

Well, we need anyway to understand if Arun was talking about source storage or backup storage

[sergeadam]

Really depends on your needs. SRM is a great product. And now that you don't need to rely on the SAN to do the replication, it adds a good features.

However, SRM is a replication and DR product. One of the strength of SRM is the failover scripting. Now ask yourself if you need immediate automatic failover.

VBR also does replication. But it also does backups. When a single VM gets hosed and your only copy is a replica at a DR site, that copy does you no good right now. When Suzy in accounting comes to you because she just deleted that 1GB Excel file, that replica does you no good. You still need a backup solution. VBR gives you all.

[zak2011]

Thanks Gostev and all the Veeam experts, for highlighting the differenced between VBR6 and SRM. It was very informative to know. Just had a discussion with the same vendors and they did agree upon the replication differences. However they said..that even though the pricing was big..since SRM integrates well with Vshere..again administration would be easier as administering veeam and vmware would require more than one window.
Also since we use two products..incase of issues with Veeam backups and replication, Veeam would point out that the problem has to do with Vmware and vice versa.
Currently we have vmware standard 4 and the recommendations gives by the vendor is to go for vmware 5 enterprise plus with SRM5 and as it is a huge investment we are contemplating to go for this or for VBR6.
My thoughts on this is again by havnig vsphere4 standard and buying VBR6 would i be able to meet all the requirements of replicating to a DR site with simplicity.
Does VBR6 have its own features similar to vmotion, SDRS and DRS which are in vmware enterprise?
Also would VBR6 be able to automatic tiering within itself ??We are considering to buy an EMC VNX storage at our primary site. As per my understanding EMC VNX has this feature but would just like to know if VBR6 has this feature.

Thanks!

[dellock6]

Uhm, Veeam is a backup and replication products, it as quick migration as a similar feature of storage vmotion (it's not completely live and the VM is freezed for a small amount of time..), but the other features are more related to your running environment, they are inside VMware and you need to understand if you need them without taking into account either Veeam nor SRM or other backup solutions. Those features are for live and running environments.

Having to deal with two consoles sorry but sounds really silly to me, you would have anyway more console to administer vSphere, the storage, some ssh sessions for networking, and so on. Why bother about one more console? Also about support and ping-ponging responsabilities and faults, it would be the same about the storage-related problems or network (is it VMware or EMC problem, or Cisco)?

About tiering, again is a feature of the underlying storage you are using, here you are comparing Veeam with vsphere SDRS technology, and they are two completely different solutions made for different purposes. Veeam would take advantage of the underlying storage features, just like vSphere or SRM would.

Forgive me for beeing rude, but it sounds like your vendor is trying to pull money from you filling the offer with its catalog made of VMware/EMC solutons. I'm not saying you do not need them at all, but both vSphere licensing or VNX storage has nothing to do with Veeam vs SRM, which was the original idea.

[zak2011]

Hello Luca,
Thanks for your response and comments. What i am confused about is at this point of time would it make sense to stick with vmware 4 standard+ veeam backup and replication+ EMC vnc. When it comes to backup, there is no doubt that VBR6 is in the lead. However the debate comes in the matter of DR. Could i manage having vmware4 standard+ VBR6+ EMC repl without having to upgrade to VMware 5 enterprise and SRM5 at this stage to have a proper backup and DR solution in place.

[dellock6]

I like SRM but in your situation, you only have two options:
- stay on vSphere 4, and buy a double storage with SRM compatibility to do storage replication
- upgrade to vsphere 5 and use the new SRM5 host based replication
both solutions are expensive and require a huge effort in design and deploy. With Veeam you can have both backup (which lacks in SRM beeing a pure replication tool) and replication and solve your actual problems.
then you can nayway evaluate where you can improve your environment and decide to upgrade storage, vsphere or other elements but without the hurry to fix the DR lack. The only big thing is missing right now in Veeam and SRM has is orchastration of the failover process, and the complete site failover functions.

Luca.

[sergeadam]

I'd stick with 4 and VEEAM.

ESX 4 standard gives you vMotion. DRS is just vMotion automated. Improperly configured DRS can spend most of it's time vMotioning VMs. I run 2 cluters, 5 hosts and 2 hosts. Took a little while to balance my workload, I now very seldom have to rebalance what runs where. No need to DRS.
I have to evacuate a SAN to reconfigure it. I sorely miss Storage vMotion. But I can live with VEEAM Migration, for the few times I'll actually have to move storage around.

[zak2011]

Thanks Lucas for your reply and suggestions. Have some queries for some points which you recommended.
- stay on vSphere 4, and buy a double storage with SRM compatibility to do storage replication ( Could veeam do storage replication? What did you mean by double storage? Since i would think of of buying Veeam instead of SRM, would i need storage with SRM compatiblity ?)
"The only big thing is missing right now in Veeam and SRM has is orchastration of the failover process, and the complete site failover functions"- Did you mean automatic failover and failback of SRM? Could you explain what you meant by orchastration of the SRM.

[dellock6]

- storage: before vSphere 4, SRM totally relies on underlying storage replication, so you would need a storage in production site, and another one in the DR site. BUT, the two storages needs to be tha same since every vendor sells replication tools that obviously work only between similar storages. So you would need two VNX, or two Compellent, or two NetApp, an so on. And yes you have to check on VMware compatibility list for storage that can work with SRM (the need a so-called SRA, or srm replication adapter)
- with vSphere 5 you can use host based replication, so SRM can use ESXi hosts and replicate between them, ignoring storage differences. But to do this you need to upgrade all your infrastructure tu vSphere 5.
- also Veeam replica is host based, so you can use different storages at the two sites, as long as they are connected as datastores to your vSphere environment
- for the last one, yes you can configure SRM to do so, and orchestration means you can group VMs to have specific boot orders, so during failover for example DCs boot first, then all the other VMs.

Luca.

[zak2011]

Thanks a lot Luca. I really appreciate taking time to answer my queries. Hope you dont mind my asking many questions, as i am quite new to Veeam and to this area.
Thankyou very much!

[dellock6]

No problem, hopefully further readers will have a complete thread to read before asking it again.

Luca.

[kgmartin365]

From what I've been able to find it doesn't appear that Veeam Replication can be used as an SRA (like) alternative for SRM. Is that the case, or am I missing something?

If it can, please point me in the right direction. If it can't, can Veeam be automated to the point to function similar to SRM?

Thanks.

[Gostev]

Hi Keith - yes and no. While Veeam does not provide SRA (and so is not "plug-able" into SRM), you can indeed use it as an alternative for SRM in certain scenarios. In some cases, SRM would be a better choice - for example, when you want to base your disaster recovery on SAN replication (approach that has its own pros and cons), or when you need complete site failover orchestration. And for other scenarios and smaller deployments, Veeam B&R will be a better choice than SRM. Thanks!

[kgmartin365]

Thanks Gostev.
We are basically looking for the complete site fail-over orchestration.
Does Veeam have the ability to automate some or all of that, or is it more of a manual process to get the DR site up and running?
I'm trying to avoid SAN replication, simply because I've tried it in the past and saw very large amounts of data changes that the SAN was trying to replicate.
They were Windows VMs and we couldn't pinpoint the data changing, but as far as the SAN was concerned there was 30 - 40 GB of data changing within a matter of a few hours, so the replication just wouldn't happen.

That client is now using Veeam for replication and having much more success. So now we'd like to see what's involved with bringing a DR site online if needed.
By the way, the SAN in both situations is HP Lefthand.

[veremin]

That client is now using Veeam for replication and having much more success. So now we'd like to see what's involved with bringing a DR site online if needed.

Based on what I know, Veeam Backup and Replication has the ace in the hole regarding replica orchestration, especially compared with SAN-based one. In general, it includes replica Failover and Failback functionality.

Failover is a process of switching over from the original VM on the source host to its VM replica on the target host. Failover is considered a temporary stage that should be further finalized by one of the following options:

•Permanent Failover (the VM replica ceases to exist as a replica and takes on the role of the original VM)

•Undo Failover (process of switching back to the original VM, reverting replication operations and discarding changes made to the working VM replica)

•Failback (the process of switching from the VM replica to the production VM. During failback, Veeam Backup & Replication uses the working replica to recover the original VM and switch back to it. Please, also notice that during this process replica VM can be failed back either to an existing VM or to entirely new location or even you can use previously made backup to restore VM in any location you want and then transfer only changes from replica VM. Failback itself provides several options: Commit Failback, Undo Failback.)

For further, more detailed information please use User Guide (p. 50).

Hope you’ll find this information useful.
Thanks.

[Daveyd]

We use both SRM 5 and Veeam B&R. We have 2 Clarrion CX4s and use SRM with EMC's RecoverPoint appliances to provide us with pretty impressive point in time failover capabilities, the only caveat being SRM will only failover to the most current point in time (which, during a disaster is usually the one is used). If we montior the SRM protected LUNs via RecoverPoint, we have the ability to presemt LUNs from a specific point in time.

Since we use SRM and RecoverPoint and the RecoverPoint SRA, we replicate entire LUNs to our DR site. We use Veeam to replicate individual VMs. Both work very, very well. SRM being (SITE Recovery manager) is more geared towards recovering entire sites during a DR scenerio.

SRM 5 does have a reprotect function which automatically protects VMs that were moved over to the recovery site so the replication is automatically reversed. If you to a permanent failover in Veeam, you have to manually re-create Replication jobs.

Each product has its pros and cons but both are top notch in my book.

[sgravel]

I'm trying to avoid SAN replication, simply because I've tried it in the past and saw very large amounts of data changes that the SAN was trying to replicate.
They were Windows VMs and we couldn't pinpoint the data changing, but as far as the SAN was concerned there was 30 - 40 GB of data changing within a matter of a few hours, so the replication just wouldn't happen.

We were facing something along those lines, and discovered that moving vswp page files (per SRM best practice) and Windows page files to unreplicated storage made a HUGE difference. SRM 5 lets you map these unreplicated resources to local resources at the recovery site. Seeing as how the page file is cleared on boot, the actual data on the paging disk is not important. Just needs to have the same signature so Windows knows which disk the clone at the recovery site is.

We also moved some drives dedicated to System State backups to unreplicated storage as they were big culprits. Those will probably go away when we adopt Veeam anyway, as they don't seem to be necessary based on what I have read.

For one LUN with a lot of paging going on, we went from worst case 24 hour + replication cycles down to about 1-2 hours on average.

Not trying to be an SRM fanboy, but like others, we wanted the full orchestration so it was the best fit for us.

[chrisdearden]

just to add that Veeam ignores page files during backup and replication (for windows boxes ) - so you dont have to reconfigure all of your VM's