Hotadd feature that worked in v5.02 is not working in v6.1

[kklo]

[ID#5203663]

I've replaced my 'old' Veeam B&R v5.02 vm with a new one with v6.1.
With v5.02 I was able to run replica jobs using hotadd, but with v6.1 hotadd feature is failing.

Checking logs there is an error claiming that a vmdk file is bigger than the maximum file size supported on datastore where 'Veeam6' vm resides.

[18.07.2012 12:47:45] < 3360> dsk| diskLibPlugin: 2012-07-18T12:47:45.942+02:00 [03360 verbose 'transport'] Mounting VM vim.VirtualMachine:736 using transport hotadd failed: Disk [OpenE:target1] MAIL_replica/MAIL-000016.vmdk is bigger than the maximum file size supported on datastore Cabina1:Raid10.\n\n
[18.07.2012 12:47:45] < 3360> dsk| diskLibPlugin: 2012-07-18T12:47:45.958+02:00 [03360 verbose 'transport'] Cannot use mode hotadd to access [OpenE:target1] MAIL_replica/SISTEMA-000016.vmdk: Cannot mount using this method.\n\n


And this is right, the datastore where 'Veeam6' vm resides has a 1MB blocksize (256 GB max. file size) and the vmdk filesize is 900 GB. But my old 'Veeam5' vm was in the same datastore than the new one and was capable to mount the vmdk file using hotadd without problems.

Is Veeam reducing their software capabilities? Why, using the same scenario, hotadd feature worked without problems in v5 and with v6.1 does not? Is this blocksize check necessary to mount a file that resides in other datastore?

You can tell me that I could change the blocksize in datastore where ‘Veeam6’ vm resides. And I've done it, but now ALL my replica jobs are not able to use hotadd, even that jobs that have small disks and previously to this blocksize change to 4MB were replicating using hotadd correctly.

Regards

[joergr]

Hi,

yeah, hotadd has some specialities, but this relies completely on vmware vadp api´s, nothing what veeam could do about that. B+R 6.1 uses newer original vmware vadp api´s than B+R 5.

you already read this one?

Do you have by any chance an esxi 5 machine deployed? If yes, could you try to move the virtual veeam server or proxy to a FRESH (1mb block size gpt) formatted vmfs5 datastore and try the backup again? I heard of installations where the hotadd then went successfull, even if backing up vmfs3 1mb/2mb/4mb/8mb block datastore vmdks.

Best regards,
Joerg

[kklo]

I've solved my problems with hotadd in B+R 6.1, so here are some tips if someone is in the same situation.

B+R v6 checks if there is enough free space in datastore where veeam vm resides to mount all the disks that will be copied, so you need a blocksize large enough to mount your largest vmdk file.
B+R v5 didn't check the blocksize, so hotadd feature worked for me when using B+R v5 despite datastore's blocksize. But when upgrading to v6, hotadd feature fails after checking datastore's blocksize parameter.

Reading carefully hotadd requirements at http://www.veeam.com/KB1054 I've had to move my Veeam 6 vm to a datastore with a large enough blocksize and after that all the replicas are able to use hotadd feature.


Thanks for your suggestions,
kklo

[Gostev]

Small correction, this check is performed by VMware vStorage API (VDDK) - not by Veeam B&R.
Veeam B&R v6 uses newer VDDK version than v5 did (newer version was required to support vSphere 5).