The real world use of Veeam Replication

[albertwt]

Hi Everyone,

Can anyone share here about what is the benefits or real world scenario of Replicating VM using VBR 4.1 ?
is this something that works like VMware site recovery manager ?

Thanks.

[TrevorBell]

Hi,

In my enviroment i have a 3 vm cluster 45 Vm`s with 7 of them being Test Vm`s ..... all being backedup with Veeam 4.1 to a SAN partiton using VA Mode. I also have another ESX server about 500 meters away in a different part of the building incase of fire etc ( gas company high risk )... once a week i replicate some mission critical Vm`s to this server as an extra measure..

I know people say " i have a 3 node cluster" if one node goes down i have another 2 etc... but what about if you have fire or water damage ? for this reason i use the replication feature to replicate about 20 Vm`s to the other location... this way if we have a major failure i can goto the DR Room and just power on 20 ready made replica`s...

At our last DR offsite testing day i took a Dell R710 to the site 30 miles away, plugged it in, powered it on...connect from another server with VC installed and powered on all 20 machines...

It all worked 100% first i powered on both DC`s they took around 10 mins to boot as they have to reboot once and my basic network was then up and running

i then powered on exchange and sql and other servers..

Yes you can goto your DR site with a USB drive with a 130gig .VBK file and do restores ( which i also did to prove it worked ) but for quickness and proof of concept i can say to any colleagues " i can build my entire critical infrastructure within 30 mins of powering the DR server on"

How many people without Veeam can say they can do this ? Backup Exec takes an hour to catalog a new tape so there is no comparison there!

If you need any other info please let me know,

Trev

[albertwt]

wow, that's cool

yes previously i was doing backup with VCB and then archive it to tape but then VBR 4.1 makes my life easier in automatically replicating my VM.
(without having to buy very expensive enterprise solution.).

[TrevorBell]

Yes many uses.... my main backup is still .vbk but replication is good also..just depends on what works best in your enviroment

cheers

Trev;

[donikatz]

We're in the process of setting up a similar config to Trev's. Right now our Veeam backups are to a cheap off-SAN array which is then backed up weekly to tape for offsite DR (Backup Exec), but we've budgeted to add more disks to have on-SAN backup and/or replication for much quicker individual restore & DR. As Albert mentions, one of the biggest wins with Veeam is the cost vs typical SAN-level solutions and/or SRM.

[albertwt]

exactly yes
this is why I love VBR 4.1 and highly recommend it to everyone.

previously I tried to replicate between 2 ESXi hosts on top of each other (data from SAN into local SATA disk) in case the SAN HDD failed.

[msebel]

We purchased Veeam strictly for the replication option. We have a DR site about 40 miles away and currently don't have the correct SAN hardware to replicate storage across a WAN link. We have about 30 vm's that are very high priority (low RTO) and need to be available to turn on in under an hour. Luckily these systems do not have a very high RPO. We continue to use our standard backup software to take backups but that does not mean the recovery time would be acceptable. Veeam replication was the best option we could find to provide us what we needed. It provides an automated and scheduled way to have replicated, ready-to-go version of the vm's without having to hire another person to do daily restores of the servers from backup.

[albertwt]

ah yes, that is one of the good thing in using VBR 4.1

[TrevorBell]

Here is a genuine real world Use....

Yesterday Production exchange server...someone was looking around the directories and decided to delete the routing group connectors and bridghead connectors!! So now i was stuck.. havent got all the settings written down ( althought i have now ) what could i do to quickly restore the Exchange 2003 server ? Email was incoming but outgoing had no connector to tell it where to send it to and emails were building up steadily.

Solution....

I replicate once a week the Dc`s, Exchange and other mission critical servers....So i powered on the 2 DC`s and exchange server on a private network... all booted with a few minutes and then i went into ESM and then saw the connector settings and manually resetup the connectors up again.

Could of been very but all worked out great.

So that is a real life LIVE reason i use Replication

Trev...

[albertwt]

great, now I want to get ready to P2V my existing Exchange Server as it went broke after the automated Windows Update after 3 AM in the past 3 days it went down twice.

[TrevorBell]

With Exchange 2003 in my case what i did was make a new exchange VM server... then copy all mailboxes across..test they are working.. then VM the physical server ..test all fine and move mailboxes back.. i know you could just have some downtime and VM the physical but i did it this way for zero downtime ( as i moved the mailboxes out of hours ) didnt have no problems

Trev..

[albertwt]

Cool so I'm pretty confident to use this Replciation technology.

the only difference compare to that Veeam Backup is that Replicated VM is always ready to execute while the backup must be restored first.

One quick question, does replication only create one copy / automated clone of the VM ? if that is yes, then great, no need to keep the multiple duplicate (retention) as we've got the real one and the _replica VM

Thanks for the reply.

Cheers.

[Vitaliy S.]

Albert,

Actually, you can configure retention policy for replicas as well. It might be useful if you replicate already corrupted VM (for ex. infected by a virus/any malware), in this case you have source and replicared VM damaged, but with the retention policy you can go back and failover to any point in time that was healthy.
Anyway, if you don't want to keep rollbacks you can configure retention policy to 1 value, that will keep only current state of the replicated VM.

Thank you!

[albertwt]

Wow it is even better )

ok Supose I set the replica retention policy to 3 so there will be VM_replica1, VM_replica2, VM_replica3 and the original VM itself ?

cmiiw

[Vitaliy S.]

Albert,

Yes, that's better for sure By setting the retention policy to 3, you will have 1 current state and 2 rollbacks.

[albertwt]

great this is why I like Veeam and the waiting time to get ESXi replication is worthed

Thanks to all who share your response here.

Cheers.

[donikatz]

With Exchange 2003 in my case what i did was make a new exchange VM server... then copy all mailboxes across..test they are working.. then VM the physical server ..test all fine and move mailboxes back...

Trev-- A little off-topic, but curious why you bothered to P2V the Exch 2003 server if you'd already built out a new one and migrated mailboxes? Why not just run the fresh VM in production? Personally I simply P2Ved Exch 2003 as-is during off-hours, I'm just curious re your procedure. (Actually, I wound up rebuilding the data partitions later to align them, but that's a different story...)

[albertwt]

hm.. that sounds interesting too.

I'm also thinking to create another new Exchange Server 2007 from scratch because the current Physical server wasn't as healthy as I 'd expected.

[albertwt]

I have one box of Windows Server 2003 installed with Exchange Server 2007 SP1 (HT-CAS-MBx role) in my headquarter office, suppose I want to create another Exchange server 2007 box with similar role in the remote office (geographically different connected with 1 MB WAN link),

I'm also looking to replicate our exchange server across sites for disaster recovery. We have two goals in mind.
1. Replicate Exchange 2007 server to VM Exchange 2007 server at different site
2. Allow users at other site to use the VM excahnge server as their local server to speed up their Outlook.

Are both of these goals possible to do on the same VM Exchange 2007 server?

i'm now thinking to replicate the VM using VBR 4.1 hourly with the initial seeding using Removable HDD over the weeekend.

[TrevorBell]

Guys,

I made an extra Exc vm as i didnt want to change the IP of the server through our firewall ( i didnt want to cange for our isp to make the changes which would of took a day or so )... it was a case of moving the mailboxes to a new VM to keep the exchange up and running then VM the original.. and move the mailboxes back,May sound abit strange but this process took only 2 hours from start to finish thats build vm and move mailboxes.

Im sure if i just cloned the machine 12 gig c: 55 gig d: and 250 gig e: it would of took longer and i didnt want to miss any emails that would be incoming between the clone start and when it had finished cloning.


as a side note: i have just Vm`d my Vcenter that was easier! made a new VM then shut down old databases copied to new serevr and overwrote exsisting then renamed the Vcener new vm to old IP and its all running sweetly

Have a good weekend

Trev

[TrevorBell]

albertwt

yes of course that is possible...

I have just one exchange 2003 FE webmail server and one BE exchange 2003 server, we have 9 sites over MPLS ( internet connection ) this works fine for me maybe a little slow retrieving data sometimes but we only have 1 pc in each remote site.

If you add another server at your other site it would indeed speed things up and be quicker for the users.

As for replication of the link yes that would work fine too

[tsightler]

I'm also looking to replicate our exchange server across sites for disaster recovery. We have two goals in mind.
1. Replicate Exchange 2007 server to VM Exchange 2007 server at different site

Well, you don' t say how many users, and you also mention a 1MB link. WAN links are usually measured in Mb, not MB, so can you clarify if you actually have a 1Mb link? I'd be amazed if you could replicate anything but a very small, lightly used Exchange server with Veeam over a 1Mb link.

2. Allow users at other site to use the VM excahnge server as their local server to speed up their Outlook.

Certainly if you add an Exchange server at the remote site the remote users can use it, but then you confused me again. You ask "are both of these goals possible to do on the same VM Exchange 2007 server?" I'm not sure what that means. How can you have the "same VM Exchange 2007 server" at two locations? Obviously you can have two Exchange 2007 servers, one at each site, and replicate each of those to the other site with Veeam, but your statement seems to imply that you want to use the Veeam replicated VM at the remote site. This is not possible as Veeam is not a clustering solution, it's a simple, active/standby replication product, and only a one-way replication product at that, if you fail over you have to jump through hoops, promote the remote replicate to a normal VM, and preforming a full replica back, at least if you want to save the data that changed during a failover. If you have slow links that require seeding, you need to think very carefully about you failover and recovery scenario as this is a currently a weak spot for Veeam.

[icebun]

I am not sure if this is the right place but here goes.

Can anyone let me know what sort of Processing Rate, speed wise you are gettting based on the WAN link size?

Mines running at a painful 37 KB/S over a 6MB line.

Thanks.

[Gostev]

Hello, this topic would be a better place to discuss networking issues:
Problems with making replication work across a WAN/VPN?

[albertwt]

tsightler,

thanks for your response I really appreciate it

yes it is supposedly 1 Mbps (sorry for the typo that was late night). Now I understand there should be just one Active Exchange Server VM which is replicate once per 6 hours
the number of mailbox is under 60 with the average size of 1.1 GB per mailbox according to to the ExBPA.

and yes, you're right that it doesn't make sense to have VMware FT-like behaviour with Veeam, but anyway I'm still happy with Veeam.

Cheers,

AWT

[tsightler]

Right, so a 1Mbps link, assuming no WAN acceleration devices, can only transfer about 11GB of traffic per day assuming 100% utilization. Our Exchange server is quite a bit larger than yours (550 users, ~400MB per mailbox), but it generates about 75GB of replication traffic per day, and that's with only a single replication a day, it would be more if we replicated more often because Exchange touches a lot of the same blocks during the day. Have you been running Veeam backups of your Exchange environment? You can judge how much replication traffic you might need by looking at the VBR sizes of your backups.

Even Vmware FT doesn't allow both systems to have active connections and doesn't work across a slow WAN link.

[donikatz]

That may still be a tough incremental to expect of a 1 Mb link (which is likely less with overhead). Don't forget even seemingly small Exchange deployments can have a lot of incremental changes, so you may be replicating more than you expect. You'll just have to test it to see how it performs before you set expectations. Good luck and please let us know how it goes!

EDIT: Looks like Tom beat me to it, while I was composing...

[kevindv10]

Yesterday was my first day using Veeam so please bear with me. I will say that it was the quickest install and test of critcal software that I've experienced in years. I installed the demo and had my first VM backup running in less time it takes me to drive to work! Love it!

A couple of posts mention replicating Exchange once a week. What are you doing during the rest of the week for store backups? Bringing a week old Exchange box online would result in an enormous amount of data loss. I'm sure you've though of this, but I'm curious what others are doing. We use Backup Exec 12.5 with an Exchange agent to backup stores, and in the event of a crash we would (well, this will all change now that I've found VBR!) build a new Exchange box and restore from our daily backup.

Are there any best practices when using VBR and Exchange or other databases? I'm scouring the forum for any gotcha's that other people have experienced but happily haven't really found anything earth shattering yet.

[donikatz]

The posts re weekly replications are for offsite DR purposes, not daily backups. For that, you could backup and/or replicate Exchange daily like you do with BackupExec, or you can do it multiple times a day, as long as your storage can handle the load during business hours. BE is still easier for brick-level restores, but you can certainly restore brick level from Veeam, with a few more steps. Lots of threads re Exchange & VSS, so if you search around I'm sure you'll find a lot of good advice already here.

[Gostev]

BE is still easier for brick-level restores, but you can certainly restore brick level from Veeam, with a few more steps.

We should talk more about this on Monday