HotAdd Error: The specified transports are not available.

[tfleener]

I too am having a problem with this same error, and have been working with tech support the last 3 weeks. Network mode works fine, however Appliance mode or SAN mode does not work. The vmdk-flat file is Hotadded to the server fine, and I can copy files from it, however it seems it seems the Veeam client can not access it for some reason.

The ESX host I have tested it on are HP DL380 G6, and HP DL360 G5.

So far I have:
1) Setup ESX connection using IP address, bypassing VC
2) Created another root user on ESX for authentication.
3) Verified Disks are under 1TB
4) Remove Veeam server from domain (disabling all GPOs)
5) Moved server to remote site, and ran it against an older SAN.
6) Sent log file a dozen times.
7) Made sure both Veeam server and VM Guest Host are on the same server.
Create a VM server from scratch, install VBR and tested.
9) No IDE disks.

This one is a real tough one.

Any ideas ?

[Gostev]

Hello, are you using DataDomain as target? Although it should not matter for this issue, I would like to split this into separate topic to prevent confusion and mix up. As for your issue, please continue working with our support on this, as we cannot assist on troubleshooting technical issues over forum. By the way, what is your support case number - I would like to check on its status.

We have also been working compiling a list of issue which may prevent hot add from working, I will check on Monday to see what other things you may need to check.

Thanks!

[tfleener]

I am not using DataDomain, we have a EMC Cx4 SAN, the error messages appear to be the same, plus it works in network mode fine(similar to us) , so it appears to be related to using the VStroage API so I thought I would post to it.

My case number is : [ID#526439]. Any assistance greatly appreciated.

Tom.

[tfleener]

Gustov, can you please send the list of what issues may prevent the the following errorr :

Client error: The specified transports are not available. Available transports: [nbd]. Specified transports: [hotadd]. VMDK path: [[datastore1] pctest/pctest.vmdk].

To date, my issue has not been resolved.

[Vitaliy S.]

Hello Tom,

The issue might be related to the Secondary IDE Channel presented on the Veeam Backup server. It would help us a lot if you could send this registry key from the backup server to our technical support team for further troubleshooting:

Code: Select all

HKEY_LOCAL_MACHINE\HARDWARE\DEVICEMAP\Scsi

Thanks!

[Gostev]

Hello Tom, here is the list:

• Hot Add transport may fail while mounting large (1 TB) VMDK files.

• LUN with Veeam Backup server is located should have capacity more than (or equal) the capacity of the disk which is being hot added. Actual free space can be less than hot added virtual disk size.

• Hot Add transport is not supported for backup server VM with IDE virtual disks. Use SCSI virtual disks instead.

• Hot Add transport is not supported for backup server VM with SAS adapters.

• Hot Add may fails if Veeam Backup VM has one of the following configurations:
- The proxy virtual machine on ESX 4.x has an IDE disk attached at IDE 0:0 and IDE 1:0.
- The proxy virtual machine does not have any IDE CD-ROM device.

• LUN where Veeam Backup VM resides should be formatted with proper block size to be able to mount the largest virtual disk of the hot added VMs.
1MB block size – 256GB maximum file size
2MB block size – 512GB maximum file size
4MB block size – 1024GB maximum file size
8MB block size – 2048GB maximum file size

• Hot Add and SAN transport fail, falling back to NBD transport, when a backed-up virtual machine has independent disks.

• With Hot Add transport, backup of a virtual machine fails when multiple disks have the same name, as they could if on different datastores.

• A single SCSI controller can have a maximum of 8 disks attached. To run multiple concurrent jobs with more than 8 disks, you should add more SCSI controllers in your Veeam Backup VM virtual hardware.

• In case of standalone host connection (no vCenter presented), you can only hot add VMs which are located on the same host running Veeam Backup VM.

• Virtual Appliance backup job fails after you’ve cloned or restored your VBR server. The reason for that is missing UUID for Veeam Backup VM virtual machine on the ESX server. Please refer to this KB article for the instructions on fixing the issue.

[drbarker]

Tom: If you're using a CX4 SAN, why not use SAN mode? The CX4 has the added advantage in that you can use one set of Ultraflex modules for production traffic (FC?) and a second set for backup traffic only (iSCSI?)

[tfleener]

Interesting, the registry key " HKEY_LOCAL_MACHINE\HARDWARE\DEVICEMAP\Scsi" has values, but there are no entries in the Device Manager.

Gustov list definitely has some stuff I was unaware of, however in testing, I think we meet all the criteria as the server we are trying to backup is only 10GB, and the VBR server is 16GB on the sysem drive. The secondary drive is smailler, but we added that later to add an additonal scsi controller during troubleshooting.

I am assuming the Proxy Virtual Machine does not apply to appliance mode/SAN mode, but I am not absolutely sure. I do not have a CD ROM attached to my VBR server, but I plan to add one and test it out regardless.

On drbarker question, we have tried SAN mode, but have been unsucessful with similar type errors. Currently we are not utilizing the ISCSI connection, but it does look to be an interesting option.

[Vitaliy S.]

Tom,

If you have the values in the registry key, it means that we might have found the reason for your issue. Please continue working with our support team on this matter.

[murda]

how is this issue solved?
I always run towards this issue when I upgrade client environment towards esxi4.1.

I can't remember how I fixed this issue last month :/

[Gostev]

Veeam Backup 4.1.2 is able to detect invalid registry entries automatically, and use correct ones instead (documented under Resolved Issues in the Release Notes).

[murda]

Alright. Newer version fixed the issue.

[sofatime]

A short addition to this old thread. Gostev mentions the following:

LUN where Veeam Backup VM resides should be formatted with proper block size to be able to mount the largest virtual disk of the hot added VMs.
1MB block size – 256GB maximum file size
2MB block size – 512GB maximum file size
4MB block size – 1024GB maximum file size
8MB block size – 2048GB maximum file size

The release notes for 5.0.2 say the following (p.10):

Virtual Appliance mode fails if ESX datastore hosting Veeam Backup VM has different VMFS block size than datastore where backed up virtual disk is located.

This is contradictory, so I asked support about this: gostev is right, the release notes are wrong. The block size has to be big enough for the hot added VMs, and not the same as the datastore with the backed up disks.

[Vitaliy S.]

Thank you for noticing it, much appreciated.

[vmnewbie1]

Gostev,

Thank you for posting this information. This is helpful as I try to configure our new install of Veeam BaR v6. Just a couple of quick, and hopefully simple questions/clarifications.

• LUN with Veeam Backup server is located should have capacity more than (or equal) the capacity of the disk which is being hot added. Actual free space can be less than hot added virtual disk size.

If I understand this correctly, if my Veeam BaR Server is a VM on a 1TB datastore (FYI, I'm in an NFS environment) which has 400GB free space, I can still backup a VM with an 800GB vmdk file? How/why is this possible? And if two jobs are running simultaneously, this means the Veeam VM's datastore needs to have the capacity for the disks from both jobs? (I'm assuming only 1 disk at a time is mounted from a VM, not all at once. Meaning I would need capacity equal to the sum of the largest disk from each VM in both jobs, not all disks from all VMs in the jobs.)

• Hot Add transport is not supported for backup server VM with SAS adapters.

This is for the Veeam BaR Server, not the VMs being backed up, correct? And what about paravirtual adapters? Are they supported for this mode?

Many thanks!

[Gostev]

If I understand this correctly, if my Veeam BaR Server is a VM on a 1TB datastore (FYI, I'm in an NFS environment) which has 400GB free space, I can still backup a VM with an 800GB vmdk file? How/why is this possible?

Because nothing is ever copied there to that datastore. This is (or was?) just a bug with some incorrect disk space check. May have been fixed by now, frankly this topic is quite old (soon 2 years old). May be I should lock it at this time...

This is for the Veeam BaR Server, not the VMs being backed up, correct?

Correct, just like your quote says: "Hot Add transport is not supported for backup server VM with SAS adapters"

And what about paravirtual adapters? Are they supported for this mode?

Out of common sense, I recommend that you go with default adapters, or you are just calling for issues. There is really no point to use different adapters. Remember, with any software at all, standard configurations usually get 100 times more QC time than non-standard ones. This means that the chance of running into bugs or issues becomes proportionally higher with non-standard configs.

[martins]

This started happening with one of my customers last week. So I applied the update 2 patch to VEEAM B&R 6, and it fixed it. However the customer ran another recovery job this week after I patched, & lo & behold the error started again. So I deleted the registry entries (backed up key first) & ran a backup job, and viola, no errors. Hot add working fine.