Map Disk Region??

[ec-inflxx]

I just switched from Veeam running as a VM to Veeam in a phsyical environment directly attached to the SAN. In my first run of backups I few of the VMs have been completed fine. But with one when I look at the vCenter client I see a ton of "Map disk region" events. Can anyone tell me what it is doing and if it will affect production? None of the other VMs have shown this and this never happened when I ran the Veeam server as a VM.

I am also seeing "Release disk lease" which I have never seen before but it seems to happen for each one and hasn't had any negative affects.

Thanks!

[Gostev]

This is normal, see here for the explanation:
http://kb.vmware.com/kb/1007331

vStorage API is no different from VCB as to how it reads data.

[ec-inflxx]

Great thanks!!

[ec-inflxx]

Wondering if I can come back to this topic for a minute. I am starting to backup our QA machines. Generally these are all home supported OSes (7, Vista, XP) and they use snapshots on these because they are reverting to different builds of our software, etc. I am starting to do the first one and I am getting TONS of Map disk region. It seems like it is almost constant. I am assuming it is because these VMs have snapshots and are probably framented (as per KB above). Wondering if the backup will ever kick off and if there are alternatives to backing up this VM?

Thanks!

[Gostev]

Snapshot presense should not really affect backup. Yes you will get these messages in vCenter, but this is not a problem. I backup VMs with snapshots all the time without any issues. Also, from the topic about infamous CBT bug 1 year ago I learnt that (surprisingly) a lot of customers are using VM snapshots heavily in various scenarios, not just in dev but also production environments, so it is not uncommon scenario at all for Veeam Backup to protect such VMs. Thanks!

[ec-inflxx]

Ok, well I started the backup at 11:05 and it is still doing the Map disk region thing. So almost an hour and the backup hasn't even started and I have 40 of these type machines. I just opened up a support request because this doesnt seem right.

Thanks!

[larry]

I just ran into one of these, the VM has no snap shots, but the disk is thin and has expanded many times. Each backup was taking an extra 15 minutes just to "map region". The fix is to vmotion to a new datastore. This cut the backup time by 15 minutes ( no more map region ). This is a VMware issue not veeam but it affects Veeam.

Just for everyones info.

[Vitaliy S.]

Larry, thanks for sharing!

[gvinpin]

Also, if you are using dynamic disks and span it across different vmdk's - it will generate a lot of "map region", but backup will be OK anyway.

[kte]

Is there a way to stop these messages, because I have a big file server (15TB ) and in the events I only see Map Disk Region events more the 1000 lines

K

[Vitaliy S.]

Koen, have you tried applying the workaround mentioned by Larry in the post above?

[veremin]

I’m wondering whether you have (old) snapshots on this VM or not. Thanks.

[kte]

no snapshots , i did a consolidation

[foggy]

Have you tried to vMotion the VM to a new datastore? It helped one of the previous posters here.

[kte]

it is 15 tb server i don't have that much free space anymore

k

[maverick964_uk]

We have veeam 6.5 patch3 installed. All has been working fine via virtual proxies.

We are now testing backups with physical proxy servers and we get a ton of "Map Disk Region" messages in Vcenter for those VM's that have user snapshots.

How do we get around this problem without having to remove user snapshots. Obviously its not a problem when using virtual proxies!!

thanks!

[foggy]

Mike, please review the thread above for some hints that helped others in similar situations. Thanks.

[maverick964_uk]

do these messages go eventually? Why should this be a problem with physical proxies and not virtual proxies?
It would be good to suppress these messages or understand a fix .....

[foggy]

Have you tried to vMotion these VMs to a new datastore?

[tsightler]

Physical proxies access the disk directly, so they need to request leases for each region they are backing up. When backup up via other modes the I/O is going through the ESXi I/O stack which already "owns" the underlying disk. There are not errors and are just a normal part of backing up via direct SAN mode, especially with thin disk that have grown a lot over time. Performing a storage vMotion should help since it will effectively consolidate all of the chunks that have been added over time.

[dellock6]

And I can't find it right now, but if I remeber correctly there is also a way to suppress those messages to appear, and most of all to fill vCenter database...

Luca.

[maverick964_uk]

hi, I need to run a test and see if the backup of a VM (with user snapshots) does eventually work. I'm concerned about the number of messages that it generates...

we cant vmotion to another datastore as the VM's need to reside in certain datastores.

Also it is possible for user snapshots to be there so I dont want to potentially break vcenter if an overnight backup causes 1000's of "map disk region" messages to appear....

If anyone knows how to suppress these messages, please let me know.

thanks and appreciate your replies.

[dellock6]

my bad, I checked again and I was wrong, there is no way to suppress those errors.
The only permanent solution is to run a storage vmotion of the thin provisioned disk in order to remove its fragmentation, that is the cause of those kind of errors.

Luca.

[maverick964_uk]

many thanks!

[jimboblair]

I am also plagued by this. Did you get a resolve? Stuck with snapshot on the VM and cannot do storage vmotion it while powered on with a snapshot present. It has been almost an hour of Map Disk Region tasks.

[foggy]

Jim, have you tried all the suggestions given in this thread already?

[jimboblair]

Working on migrating the VMs to alternate storage volumes today to see if it cleans them up.