Instant Recovery

[theviking28]

Here is my backup infrastructure layout.

2 HP dl360p gen8 servers in esxi cluster. 5.5

1 HP p2000 SAN directly attached via SAS

1 HP x1600 G2 NAS <----- this is my backup repository. It is a server running Windows 2008 storage server with a disk storage system attached, etc. So that means I should be able to install Veeam's agents/vpower stuff, and I did.

Servers are all virtual (including veeam b&r server) EXCEPT my NAS, which is a physical box.

The issue I have had since day one, is that when I try to do an instant recovery is the following

8/22/2014 9:08:50 AM Starting VM "" recovery
8/22/2014 9:08:58 AM Connecting to host ""
8/22/2014 9:09:39 AM Error Checking if vPower NFS datastore is mounted on host
8/22/2014 9:09:39 AM Error Failed to publish VM kittens Error: Unable to mount vPower NFS volume (nas:/VeeamBackup_nas). nas: An error occurred during host configuration.
192.168.100.240: An error occurred during host configuration.

Now, if I go into my backup repository (nas) settings within veeam and I switch my vpower nfs server to my veeam backup server itself, then instant recovery "works" but its horribly slow. I have been told that the reason is that the vm is started up out of the backup file on my nas BUT then it is piped over a single 1gb Ethernet to my veeam backup server. (which makes sense, as the vpower server is set to my veeam backup server at that time)

I believe that my overall performance of instant recovery would be much improved if I can run my VMs directly off the NAS storage itself, as then all of that process is "local" to the nas.

Are there any things to check easily about why my Windows 2008 NAS fails to mount a vPower NFS datastore on my hosts? As I said, the Veeam backup server itself has no issue doing that.

[foggy]

Derek, you can try to mount the vPower NFS datastore manually. If that does not help, I suggest contacting support for deeper analysis, since logs review is required in this case.

[theviking28]

When I tried to manually mount the datastore, I found that there is somehow a "ghost" past datastore that says it is unmountable and unreachable. I am not able to "unmount" it. I am guessing this is the problem, but I will have to work with VMware to figure out how to forcefully unmount it. When I tried using a standard command, it told me the datastore was not mounted with NFS so it could not unmount it.

[foggy]

Please keep us informed on the results of you communication with VMware.

Btw, restarting the host management services might help.

[theviking28]

Ok so, I was able to work with VMware to get the old datastore removed. Then, I found out another key thing... I had Microsoft services for NFS installed on my backup repository server. That was somehow taking precedence over Veeam's vPower NFS agent and answering "no" to any requests for using NFS.. I uninstalled Microsoft's NFS services, rebooted the repository and tested.. BAM! Now, instant recovery is much faster and is a viable option.

This has real world use for me, as before this worked, I had to make our business wait 4 hours to restore an entire file server, (which worked great btw) but now I could use IR to boot it and let them work during the day.

Love VEEAM! Thank you!

[foggy]

Right, it is not possible to use both vPower NFS server and Windows NFS server on the same machine due to the port conflict.

Anyway, I'm glad you're fine now, thanks for finding time to update the thread!