-
- Service Provider
- Posts: 24
- Liked: never
- Joined: Jan 11, 2012 4:22 pm
- Full Name: Alex
- Contact:
Query on CBT Performance
Hi,
I'm new to Veeam, but have been testing it on and off when I get time. I really like the product and interface, but have also always been confused by the speeds I get when using it. In particular, CBT doesn't seem to be as fast as I would have imagined it to be.
As an example, I have a Dell Equallogic SAN populated with 16 x 750Gb Sata drives which is lightly used (not used in a production environment, just used for test VM's).
This San is used as storage for a VMWare vsphere infrastructure which has a 30Gb Windows 2003 virtual machine in it, and also a Windows 2008 R2 virtual machine running Veeam 6. Everything is gigabit.
I ran my first backup job (which therefore would have no change block tracking enabled yet), and achieved a throughput of 25Mb a sec. Admittedly I'm reading from one SAN back onto the same SAN, but with 3 x gigabit links in the SAN I would have expected a faster throughput than 25Mb a sec. Even more confusing, I ran the job a second time straight away after the first backup completed as this would be where CBT should kick in and this time the job achieved 28Mb a sec, but still took 18minutes to run. The statistics said "Processed 32Gb, Read 30.5Gb, Transferred, 372.1Mb)" which makes it seem that it read through the entire disk. I thought changed block tracking kept track of which blocks on the disk have been updated, so when it's time to run the backup it doesn't have to read through the entire disk - it can just grab those specific blocks that are needed???
Could someone clarify if I'm doing something wrong with CBT, but also advise on the speeds if possible please. The final statistics on the job for the bottleneck was;
First (full) backup = Source 98% > Proxy 78% > Network 4% > Target 18%
Second (change block tracking incremental) = Source 99% > Proxy 54% > Network 0% > Target 0%
Backup mode appeared to be via the vmware hot add method as veeam is running in a vm.
Thanks in advance.
I'm new to Veeam, but have been testing it on and off when I get time. I really like the product and interface, but have also always been confused by the speeds I get when using it. In particular, CBT doesn't seem to be as fast as I would have imagined it to be.
As an example, I have a Dell Equallogic SAN populated with 16 x 750Gb Sata drives which is lightly used (not used in a production environment, just used for test VM's).
This San is used as storage for a VMWare vsphere infrastructure which has a 30Gb Windows 2003 virtual machine in it, and also a Windows 2008 R2 virtual machine running Veeam 6. Everything is gigabit.
I ran my first backup job (which therefore would have no change block tracking enabled yet), and achieved a throughput of 25Mb a sec. Admittedly I'm reading from one SAN back onto the same SAN, but with 3 x gigabit links in the SAN I would have expected a faster throughput than 25Mb a sec. Even more confusing, I ran the job a second time straight away after the first backup completed as this would be where CBT should kick in and this time the job achieved 28Mb a sec, but still took 18minutes to run. The statistics said "Processed 32Gb, Read 30.5Gb, Transferred, 372.1Mb)" which makes it seem that it read through the entire disk. I thought changed block tracking kept track of which blocks on the disk have been updated, so when it's time to run the backup it doesn't have to read through the entire disk - it can just grab those specific blocks that are needed???
Could someone clarify if I'm doing something wrong with CBT, but also advise on the speeds if possible please. The final statistics on the job for the bottleneck was;
First (full) backup = Source 98% > Proxy 78% > Network 4% > Target 18%
Second (change block tracking incremental) = Source 99% > Proxy 54% > Network 0% > Target 0%
Backup mode appeared to be via the vmware hot add method as veeam is running in a vm.
Thanks in advance.
-
- Chief Product Officer
- Posts: 31419
- Liked: 6992 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
- Contact:
Re: Query on CBT Performance
According to the statistics of the second run, CBT is not being used at all. Most likely, the virtual hardware version of this VM does not support CBT. Thanks.
-
- VeeaMVP
- Posts: 6152
- Liked: 1960 times
- Joined: Jul 26, 2009 3:39 pm
- Full Name: Luca Dell'Oca
- Location: Varese, Italy
- Contact:
Re: Query on CBT Performance
Are you using virtual hardware V7? If it's V4, CBT is not an option.
Luca.
Luca.
Luca Dell'Oca
Principal EMEA Cloud Architect @ Veeam Software
@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
Principal EMEA Cloud Architect @ Veeam Software
@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
-
- Service Provider
- Posts: 24
- Liked: never
- Joined: Jan 11, 2012 4:22 pm
- Full Name: Alex
- Contact:
Re: Query on CBT Performance
Thanks for the updates. The machines are all version 7 so cbt should be used. I'll double check the logs tomorrow but couldn't see anything that said cbt wasn't being used.
-
- VeeaMVP
- Posts: 6152
- Liked: 1960 times
- Joined: Jul 26, 2009 3:39 pm
- Full Name: Luca Dell'Oca
- Location: Varese, Italy
- Contact:
Re: Query on CBT Performance
Have a look at the datastore, if you see the ctk files. they are the CBT tracking files. If they are present and you still suffer slow speed, you can try to reset CBT status and start again from scratch. There are some issues on VM if CBT was enabled when the VM had a snapshot, do not know if this is your issue, but better check anyway.
Luca.
Luca.
Luca Dell'Oca
Principal EMEA Cloud Architect @ Veeam Software
@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
Principal EMEA Cloud Architect @ Veeam Software
@dellock6
https://www.virtualtothecore.com/
vExpert 2011 -> 2022
Veeam VMCE #1
-
- Service Provider
- Posts: 24
- Liked: never
- Joined: Jan 11, 2012 4:22 pm
- Full Name: Alex
- Contact:
Re: Query on CBT Performance
Thanks all, seems that the VM was running version 4 even though I was literally certain it was verison 7.
After updating to version 7, it's working much faster now thanks.
One final question. I'd like to test the replication component as well as the backup component. Would I be right in saying it is best to place the Veeam Backup server itself in the REMOTE office, i.e. the replication target? As if the Veeam server is in the primary network and this fails, you won't have access to Veeam to initiate the failover?
After updating to version 7, it's working much faster now thanks.
One final question. I'd like to test the replication component as well as the backup component. Would I be right in saying it is best to place the Veeam Backup server itself in the REMOTE office, i.e. the replication target? As if the Veeam server is in the primary network and this fails, you won't have access to Veeam to initiate the failover?
-
- Chief Product Officer
- Posts: 31419
- Liked: 6992 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
- Contact:
Re: Query on CBT Performance
Up to you. There is really nothing preventing you from replicating Veeam server to remote site along with other VMs, if you install it in the main site.
Who is online
Users browsing this forum: No registered users and 30 guests