• EN
Solved

Replication window did not finish

E
Userlevel 4
Badge

Hi All,

I have a question about completing replication jobs if the replication windows exceeds. 

 

For the initial replication, Veeam will transfer the base full size of the replicated VM. Let's suppose it's a huge size VM and initial replication did not finish in the repli action window.

When the next schedule run, will Veeam start transferring from the beginning for the initial full or incremental replica that does not finish, or it will continue from the stopped point abd replication does not take another time and size?

 

I have read about (resume disconnect) but need to make sure from the experts? 

https://helpcenter.veeam.com/docs/backup/vsphere/replica_resume_disconnect.html?ver=110

Let's suppose the replication window from(17:00-23:30) only.

 

Thanks

icon

Best answer by regnor 14 January 2023, 09:14

View original

7 comments

regnor
Userlevel 7
Badge +14

If the job gets terminated or exceeds the backup windows, it will start from the beginning and not pickup its progress from the previous run. So for the initial replication you'll have to disable the backup window. If that's not possible then as an alternative you could restore the VM to the target environment and use the replica mapping feature. That way you don't need to replicate all data on the initial run. Instead Veeam will compare the state of both VMs and only replicate the changes.

Max M. | Systems Engineer @Veeam | VMCE/VMCA | former Legend & Vanguard
HunterLAFR
Userlevel 7
Badge +8

I got the same issue long time ago,

we run the initial replica stopping the backup tasks, just to ensure the full replica was done ok, and then we continue our norma operation, since then, all replicas were incremental, so no problem at all.

We did this procedure one Saturday night, ensuring our production environment was backed up, and everything was very smooth.

Also @regnor solution is good, but if you don't feel confident enough to run it, its ok, there is another alternative.

cheers.

Luis F.- HunterLF - HunterLAFR - lfconsulting.org
E
Userlevel 4
Badge

@regnor Many Thanks for your help and providing the alternative solutions for using replica mapping.

I wanted to double check this, the issue if you make backup and restore to remote site, it will be done over wan link as well and we have to wait for restore time on the target environment. So if the machine is huge in size, the will take some time to be restored. 

 

The customer does not want to consume the wan link out of the mentioned time (17:00-23.30).

 

So, Right now the only available options according to the customer needs is to let the initial replica (full) to be finished discarding setting the replication window.Right? 

E
Userlevel 4
Badge

@HunterLAFR Thanks a lot for your help, what I have understood from your reply that you stopped backup operations over wan link to save the bandwidth for Replication job to be finished.

If I understood Is right, I agree with your solution @HunterLAFR ,but the issue now is the full replication job that run only one time did not finish in the replication allotted window, it will be stopped.

So next time the replication job runs, it will start from the beginning since it did not finish. Right? 

MicoolPaul
Userlevel 7
Badge +20

@regnor Many Thanks for your help and providing the alternative solutions for using replica mapping.

I wanted to double check this, the issue if you make backup and restore to remote site, it will be done over wan link as well and we have to wait for restore time on the target environment. So if the machine is huge in size, the will take some time to be restored. 

 

The customer does not want to consume the wan link out of the mentioned time (17:00-23.30).

 

So, Right now the only available options according to the customer needs is to let the initial replica (full) to be finished discarding setting the replication window.Right? 

A workaround for this would be to throttle the bandwidth on your WAN link for Veeam outside of the permitted hours, this way the job can keep processing, even with dramatically reduced throughput. This would work with discarding the replication window for your full. A reminder however that you’ll have a snapshot on the source VM for the duration of this replication job. If you have insufficient bandwidth for your replication window the two simple answers are increasing the bandwidth/performance, or increasing the replication window.

 

 

 

Do you have backups that are copied to this site? You could use a backup as a seed if this is the case?

Michael Paul - #VeeamLegend | #VeeamVanguard | Veeam Certified Architect | VMware vExpert | https://micoolpaul.com | Twitter: @MicoolPaul | Mastodon: @micoolpaul@masto.nu | Bluesky: @micoolpaul.bsky.social
E
Userlevel 4
Badge

@MicoolPaul Thabnk you for your help, this is very good option (network throttling) outside of the permitted window to keep transferring to the target site.

 

Unfortunately, we do not have backup or backup copy job located on the target site to be used as replica seeding. 

 

That's why I highly agreed with network throttling option.

 

Really appreciate all the mentioned ideas,those will help a lot @regnor @HunterLAFR @MicoolPaul 

 

 

HunterLAFR
Userlevel 7
Badge +8

@HunterLAFR Thanks a lot for your help, what I have understood from your reply that you stopped backup operations over wan link to save the bandwidth for Replication job to be finished.

If I understood Is right, I agree with your solution @HunterLAFR ,but the issue now is the full replication job that run only one time did not finish in the replication allotted window, it will be stopped.

So next time the replication job runs, it will start from the beginning since it did not finish. Right? 

Hi, If Im not wrong, replica if is canceled, then it will start over again, answering your question, if you make a full replica once, then the future replicas will be differentials, not full again, until you break the chain or make so much changes that the amount of data would be huge.

The best, in my personal opinion, as always, try, test, play, enjoy. and then choose the most suitable workaround for you.

mora than happy to help, do not hesitate to raise any other question, and please post your final decision and tests, we all love that!

cheers!

Luis F.- HunterLF - HunterLAFR - lfconsulting.org

Comment


Terms & Conditions