Replication Job and backup job overlaps

[ordinary]

I guess when there are Replication Job and backup job overlaps veeam getting confused . i got some warnings when i start continuous replication , Warnings are below

1 ) Cannot use CBT: Soap fault. Unable to communicate with the remote host, since it is disconnected.Detail: '<HostNotConnectedFault xmlns="urn:internalvim25" xsi:type="HostNotConnected"></HostNotConnectedFault>', endpoint: ''

2) Removing snapshot

Or is the mentioned error are just ok ?

Second thing , How can i run the continuous replication job in a time window (only business hours )

If the job could not finish in the time window , replication will go inconsistent state ?

What is the best practice to modify a running job

Thanks

[Gostev]

With v6, you should not have any issues when jobs overlap, as one job waits for another one to finish before starting to process VM.
To run the continuous replication job in a time window, you need to specify that window in the job schedule (in the Periodically line).
Best is to stop the job before modifying.

[Berniebgf]

Have seen these issues as well.

In my case it is two different Veeam servers accessing the same VM. One (DR site) for replication and One (Production Site) for backups.
DR site Is using Virtualised Proxies (Hot add mode)(Source and destination), production site is using "Prod Site" Veeam server + Proxy for backup "SAN Mode".

they do not play nicely together, have to set exclusion periods for replica Jobs.

Bernie.

[Gostev]

Bernard - actually, while this scenario would definitely cause problems with B&R v5, it should work absolutely fine in v6. In fact, this was causing our very own IT some problems and v5, which is why addressed it in v6. If you are seeing any issues with this scenario and v6, then please open a support case as we would certainly like to investigate this. Thanks!

[ordinary]

Do you Recommend that multiple servers for veeam BR server , one for backup and one for replications like that .........

[dellock6]

In this way you are going to loose coordination between the jobs. A better solution, similar to what you have in mind, would be to have a single Veeam Server to coordinate jobs, and then multiple proxies, some of them dedicated to backup jobs, and other to replication jobs.

[Berniebgf]

Having your Veeam Server at the destination side for your Replica's gives you more options / control in the event that the primary site is lost.
So I prefer to have a veeam server at DR for Replica's (and backups for DR) and one at production for Production Backups.

I have found that an important item when performing LOTS of replica's......~30-40 VM's, you need to resource your VirtualCenter Server Correctly with performance in mind.
With allot of Replica Jobs there is allot of work performed by the VirtualCenter Server...especially when using Virtual Proxies with Hot Add. In my case there is constant tasks being performed by virtualcenter/s.

I have a site with the following.

1. Production VirtualCenter Linked mode to DR VirtualCenter (Production Veeam Physical and DR Veeam Physical)
2. 6 x Veeam Virtualised Proxies at Production (only for replica's) (3CPU, 3GB)(All Hot Add Replica Jobs)
3. 6 x Veeam Virtualised Proxies at DR (only for replica's) (3CPU, 3GB)(All Hot Add Replica Jobs)
4. The Physical Veeam servers ONLY handle Backups with "SAN mode".

This config is working well.....Many replica's occuring, some 15min, some 1 hourly, some 2 hourly...

B.

[tsightler]

Having your Veeam Server at the destination side for your Replica's gives you more options / control in the event that the primary site is lost.
So I prefer to have a veeam server at DR for Replica's (and backups for DR) and one at production for Production Backups.

Yep, this is always my recommended approach as well. There's certainly nothing wrong with having a single server, but for sites with a lot of backups/replication, having a local server dedicated to backup, and a server in the DR site dedicated to replication is a really nice setup. The server in DR is there, available when you need to failover, and can then be used to perform backups while systems are operating at the DR site.

A lot of customers are replicating only a small subset of their VMs with Veeam, but are then using a dedupe appliance to replicate their Veeam backups to the DR site for the rest of their VMs. Having the second server at the DR site is also nice in this scenario because you can write a script that manually keeps the backups imported into the DR site's Veeam server, making restores at the DR site very easy as well.

[guillermo.lozano]

HI Community,

I have this setup (VB&R for backups on primary site and VB&R for replication on DR Site).
I need to be able to sync jobs so certain backups launch immediately after a couple of replication jobs on the DR VB&R.

Is this possible?

Thanks.
Regards,
Guillermo.-

[Vitaliy S.]

Hi Guillermo,

Yes, that's possible to do via a PowerShell script. Please check out this topic for the script example: Remote PowerShell

Thanks!

[veremin]

Hi, Guillermo. The bottom line is that you will have firstly to connect to VB&R instance in DR site, using Remote PowerShell execution mentioned by Vitaliy. Then, check the status of the corresponding job using .GetLastState()method and execute the required backup job, only if the status isn’t equal to “Working”.

Although the following example doesn’t have remote execution part, it still might be useful for you:

http://forums.veeam.com/viewtopic.php?f ... 558#p71558

If you need any further assistance in PS scripting, don’t hesitate to let me know.
Thanks.

[Vitaliy S.]

I'm not sure it is even required to check job status, you can trigger a PS script (to trigger the 2nd job) as a post backup job script for the 1st job. Second job on another server will most likely be in the idle state, but querying job status will keep you on a safe side for sure.

[guillermo.lozano]

Hi thank you both for the reply. I’m setting this up to do the job as a post action script (chained).

When the backup job completes, the replication jobs is launched and when the replication job completes the prior backup job is started and so on. This way no job has a schedule, and the first time it must be manually started.

I hope in a near future version Veeam supports chaining jobs in different servers for a scenario like mine to avoid all this scripting.

Thanks again.
Regards,
Guillermo.-