-
- Influencer
- Posts: 11
- Liked: never
- Joined: Sep 02, 2011 6:36 am
- Full Name: Yajith Dayarathna
- Contact:
Improving replication performance
Hello All,
First of all, sorry if some questions below has been already asked and answered on the forum, But I thought of re-confirming with v7 capabilities.
One challenge I'm facing at this customer the amount of changes are too heavy(for the WAN link) on a set of DBs(mostly MS-SQL) and if i stop the replication for backups changes can get accumulated. I would like to know if it is possible to run backups and replications of the SAME VM at the same time. To further clarify what I mean, Is it possible to have replication jobs to run "Continuously" and also perform daily backups on the same VMs without stopping the replication ? When i checked on this sometime ago, it was not recommended to do so.
The other observation is that the amount of changes detected by CBT seems to be far higher than the actual amount of changes inside the VM. As I understand, this is due to the fact that VMware is tracking the changes at VMFS blocks while actual changes are happening inside the VM in a different manner. We have excluded some non-essential components like TEMPDB from backup for now, I would be interested to know what can be done to improve on such situations ?
Regards,
Yajith
First of all, sorry if some questions below has been already asked and answered on the forum, But I thought of re-confirming with v7 capabilities.
One challenge I'm facing at this customer the amount of changes are too heavy(for the WAN link) on a set of DBs(mostly MS-SQL) and if i stop the replication for backups changes can get accumulated. I would like to know if it is possible to run backups and replications of the SAME VM at the same time. To further clarify what I mean, Is it possible to have replication jobs to run "Continuously" and also perform daily backups on the same VMs without stopping the replication ? When i checked on this sometime ago, it was not recommended to do so.
The other observation is that the amount of changes detected by CBT seems to be far higher than the actual amount of changes inside the VM. As I understand, this is due to the fact that VMware is tracking the changes at VMFS blocks while actual changes are happening inside the VM in a different manner. We have excluded some non-essential components like TEMPDB from backup for now, I would be interested to know what can be done to improve on such situations ?
Regards,
Yajith
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Actually there are no issues with running simultaneous backup and replication jobs against the same VM.yajith wrote:I would like to know if it is possible to run backups and replications of the SAME VM at the same time. To further clarify what I mean, Is it possible to have replication jobs to run "Continuously" and also perform daily backups on the same VMs without stopping the replication ? When i checked on this sometime ago, it was not recommended to do so.
Yes, generally, your understanding is correct, this comes down to NTFS design. Please review this topic for some hints on how to optimize backup sizes for those VMs.yajith wrote:The other observation is that the amount of changes detected by CBT seems to be far higher than the actual amount of changes inside the VM. As I understand, this is due to the fact that VMware is tracking the changes at VMFS blocks while actual changes are happening inside the VM in a different manner. We have excluded some non-essential components like TEMPDB from backup for now, I would be interested to know what can be done to improve on such situations ?
-
- Influencer
- Posts: 11
- Liked: never
- Joined: Sep 02, 2011 6:36 am
- Full Name: Yajith Dayarathna
- Contact:
Re: Improving replication performance
Hi foggy,
Thanks so much for the clarification. I will look in to the threads you have mentioned.
Regards,
Yajith
Thanks so much for the clarification. I will look in to the threads you have mentioned.
Regards,
Yajith
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
You're welcome. Feel free to ask any additional questions, in case you have them after reviewing those threads.
-
- Expert
- Posts: 162
- Liked: 15 times
- Joined: Nov 15, 2011 8:47 pm
- Full Name: David Borden
- Contact:
Re: Improving replication performance
foggy wrote: Actually there are no issues with running simultaneous backup and replication jobs against the same VM.
Yes, generally, your understanding is correct, this comes down to NTFS design. Please review this topic for some hints on how to optimize backup sizes for those VMs.
I don't really agree with this statement. If the backup and replication jobs run on the same server sure they will be aware of each other and not run at the same time but if they are run from different servers(backups from prod veeam and replicas from dr veaam) we have run in to quite a few fires backup up SQL and Exchange VMs when a replication was running at the same time. We ALWAYS make sure there is no backup job running at the same time as a replication job.
Now this is not Veeam's fault. This is just Exchange/SQL VSS not playing nice with 2 running at the same time. I don't think this problem would occur on a file or web server.
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Yes, in the scenario with two different backup servers running jobs against the same VM without carrying them over in time is not recommended. In this case the two servers will not be aware of each other and there will be no coordination between them, causing possible conflicts. So making sure the jobs don't overlap in this case is a best practice. However, no issues in case of a single backup server.
-
- Expert
- Posts: 162
- Liked: 15 times
- Joined: Nov 15, 2011 8:47 pm
- Full Name: David Borden
- Contact:
Re: Improving replication performance
When are guys going to make B&R servers aware of each others jobs via enterprise manager? I think most people have a different B&R server for replication. This would be AWESOME!foggy wrote:Yes, in the scenario with two different backup servers running jobs against the same VM without carrying them over in time is not recommended. In this case the two servers will not be aware of each other and there will be no coordination between them, causing possible conflicts. So making sure the jobs don't overlap in this case is a best practice. However, no issues in case of a single backup server.
-
- Influencer
- Posts: 11
- Liked: never
- Joined: Sep 02, 2011 6:36 am
- Full Name: Yajith Dayarathna
- Contact:
Re: Improving replication performance
In this particular setup I'm talking about, we have two Veeam servers, one in the production environment which does all the backups and another at the DR site which takes care of the replication. What would be the proper way to do both backups are replications in this case ?foggy wrote:Yes, in the scenario with two different backup servers running jobs against the same VM without carrying them over in time is not recommended. In this case the two servers will not be aware of each other and there will be no coordination between them, causing possible conflicts. So making sure the jobs don't overlap in this case is a best practice. However, no issues in case of a single backup server.
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Just make sure you're planning the jobs schedule in a way to avoid overlapping.
-
- Influencer
- Posts: 11
- Liked: never
- Joined: Sep 02, 2011 6:36 am
- Full Name: Yajith Dayarathna
- Contact:
Re: Improving replication performance
Now I'm in doubt if it was a good idea to have a separate Veeam server in DR to handle replications in the first place.davidb1234 wrote: When are guys going to make B&R servers aware of each others jobs via enterprise manager? I think most people have a different B&R server for replication. This would be AWESOME!
Basically it is not possible to have simultaneous backups and replications on same VMs while retaining the convenience provided by a DR site Veeam server to manage fail-overs ?
-
- Influencer
- Posts: 11
- Liked: never
- Joined: Sep 02, 2011 6:36 am
- Full Name: Yajith Dayarathna
- Contact:
Re: Improving replication performance
Already doing that with the backup window settings. Trying to improve a bit more since accumulated changes during the backup window takes longer to replicate and RPO is impacted at times.foggy wrote:Just make sure you're planning the jobs schedule in a way to avoid overlapping.
-
- Expert
- Posts: 162
- Liked: 15 times
- Joined: Nov 15, 2011 8:47 pm
- Full Name: David Borden
- Contact:
Re: Improving replication performance
Your last statement is correct. Until Veeam makes multiple servers aware of each other this will be a drawback. We prefer the failover and REIP functionality to work so we have seperate servers and schedule our jobs accordingly.yajith wrote: Now I'm in doubt if it was a good idea to have a separate Veeam server in DR to handle replications in the first place.
Basically it is not possible to have simultaneous backups and replications on same VMs while retaining the convenience provided by a DR site Veeam server to manage fail-overs ?
-
- Enthusiast
- Posts: 88
- Liked: 2 times
- Joined: Jul 31, 2013 12:05 pm
- Full Name: Si
- Contact:
Re: Improving replication performance
Whats the recommendation when running local backups and replication offsite, to have B&R installed on the primary site for the local backups and also installed at the DR site to run the replication?
I've noted that because the installs of B&R don't know about each other you have to be careful with scheduling so they don't conflict.
I didn't realise that failover and re-IP doesn't work if you don't run the replication job on the DR site B&R install.
Am I understanding this correctly?
I've noted that because the installs of B&R don't know about each other you have to be careful with scheduling so they don't conflict.
I didn't realise that failover and re-IP doesn't work if you don't run the replication job on the DR site B&R install.
Am I understanding this correctly?
-
- VP, Product Management
- Posts: 27377
- Liked: 2800 times
- Joined: Mar 30, 2009 9:13 am
- Full Name: Vitaliy Safarov
- Contact:
Re: Improving replication performance
Yes, that's correct.CaptainFred wrote:Whats the recommendation when running local backups and replication offsite, to have B&R installed on the primary site for the local backups and also installed at the DR site to run the replication?
RE-IP and failover option can only be performed by a "live" Veeam backup server, so if you install a secondary Veeam backup server on the DR site (using the same license key as you have on the primary site), then you will always be able to use these features.CaptainFred wrote:I didn't realise that failover and re-IP doesn't work if you don't run the replication job on the DR site B&R install.
Am I understanding this correctly?
-
- Enthusiast
- Posts: 88
- Liked: 2 times
- Joined: Jul 31, 2013 12:05 pm
- Full Name: Si
- Contact:
Re: Improving replication performance
Oh this is a pity. Not sure how I will avoid the jobs conflicting/overlapping each other without leaving a big time gap between them which will impact the RPO.
However, that said, I'm not sure we will use failover or re-IP because the link is slow and I could use instant restore from the local backup anyway.
So if I ran the backup and replication jobs from the primary site B&R install and then there was an incident in the primary site, surely I could just manually power on all the VMs at the DR site? Or could I have a full copy of B&R installed on the DR site backup server but without any jobs, then use that to manage things if the primary site was no more?
However, that said, I'm not sure we will use failover or re-IP because the link is slow and I could use instant restore from the local backup anyway.
So if I ran the backup and replication jobs from the primary site B&R install and then there was an incident in the primary site, surely I could just manually power on all the VMs at the DR site? Or could I have a full copy of B&R installed on the DR site backup server but without any jobs, then use that to manage things if the primary site was no more?
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Yes, you can safely start the replica VMs in the DR site manually in case of disaster (but losing the ability to failback using Veem B&R).
Having Veeam B&R replicated to the DR site you will be able to power on the replica VM manually and then perform failover/failback tasks from it. Another option is to use Veeam B&R configuration backup functionality. There is an existing topic on the possible ways to protect Veeam B&R server.
Having Veeam B&R replicated to the DR site you will be able to power on the replica VM manually and then perform failover/failback tasks from it. Another option is to use Veeam B&R configuration backup functionality. There is an existing topic on the possible ways to protect Veeam B&R server.
-
- Enthusiast
- Posts: 88
- Liked: 2 times
- Joined: Jul 31, 2013 12:05 pm
- Full Name: Si
- Contact:
Re: Improving replication performance
Ok thanks, I will have a read of that existing topic but can you clarify what you mean by:
Unfortunately I'm planning to use a physical B&R server at the primary site because I have an existing backup server that is high spec and has direct SAN connectivity.
Do you mean use a virtual B&R server at the primary site then replicate that to the DR site? Then boot it up manually in the case of a disaster and use it to manage failover/failback?foggy wrote:Having Veeam B&R replicated to the DR site you will be able to power on the replica VM manually and then perform failover/failback tasks from it.
Unfortunately I'm planning to use a physical B&R server at the primary site because I have an existing backup server that is high spec and has direct SAN connectivity.
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Yes, I was meaning right that.CaptainFred wrote:Do you mean use a virtual B&R server at the primary site then replicate that to the DR site? Then boot it up manually in the case of a disaster and use it to manage failover/failback?
Then you can use configuration backup to restore your Veeam B&R server in case of disaster.CaptainFred wrote:Unfortunately I'm planning to use a physical B&R server at the primary site because I have an existing backup server that is high spec and has direct SAN connectivity.
-
- Enthusiast
- Posts: 88
- Liked: 2 times
- Joined: Jul 31, 2013 12:05 pm
- Full Name: Si
- Contact:
Re: Improving replication performance
Ok is that my only option left then? Is there a KB article explaining how to do that? I have done some searches...foggy wrote:Then you can use configuration backup to restore your Veeam B&R server in case of disaster.
-
- Veeam Software
- Posts: 21139
- Liked: 2141 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Improving replication performance
Configuration backup functionality is described in Veeam B&R documentation in pretty good detail.
-
- Enthusiast
- Posts: 88
- Liked: 2 times
- Joined: Jul 31, 2013 12:05 pm
- Full Name: Si
- Contact:
Re: Improving replication performance
Ah thats where I should have looked! Thanksfoggy wrote:Configuration backup functionality is described in Veeam B&R documentation in pretty good detail.
Who is online
Users browsing this forum: No registered users and 17 guests