Change "VeeamBackup_" datastore name

[m.costantino]

Hi!
I've a very annoyng problem.
All of my surebackup "backup" jobs are not working anymore due to the fact that "VeeamBackup_myserver" is not accessible anymore by Veeam.
That datastore has been created by surebackup.
I cannot delete VeeamBackup_myserver from ESXi (it says that the datastore is busy).
I've tryed everything, no vm on that datastore, no files opened, the datastore is not browsable,no way to unmount the datastore from cli,etc...
On ESXi forums I found that the last thing to try is to rebbot the hypervisor but I cannot do that.
I would like to know is there's a way to let my veeam server to create and use another datastore. I noticed that the datastore has the name = VeeamBackup_hostnameoftheserver...
What happen if I rename the Windows server hostname (from "myserver" to "veeamsrv"), will Veeam create "VeeamBackup_veeamsrv" on the first attempt to use SureBackup?
I really cannot reboot ESXi.
Any suggestion is really appreciated!

[Andreas Neufert]

Can you try to enable SSH at the ESXi host.
Then login with root and execute the following command:
esxcfg-nas -d <datastore_name>

[Andreas Neufert]

As well please check on these:
https://knowledge.broadcom.com/external ... re-wi.html
https://community.spiceworks.com/t/zomb ... 0u2/801897


In the end this one is more of an VMware issue when we can not remove the NFS share datastore form the ESXi.
I would open a ticket at VMware first and look at the root cause. To avoid that there is a bigger issue and your rename action would end up in same result.

[m.costantino]

Can you try to enable SSH at the ESXi host.
Then login with root and execute the following command:
esxcfg-nas -d <datastore_name>

Already tryed thanks, but it says that the datastore is busy (i've tryed but no files are opended on the datastore, atm i don't remember the command but there's a way...)

[Andreas Neufert]

Did you check what the other links said, for example with the SWAP file?
As well the other link from VMware/Boradcom has a more detailed check on what prevents the unmount.

[m.costantino]

Thanks Andreas,
I've tryed for two days trying every posts I found.
Tryed EVERYTHING, other people with a config like mine found that rebooting is the solution (I've only a local datastore)
Seems that the only way is to reboot.
This is not the place for discussing how Broadcom assistance works (really don't wanna discuss), I'll open a ticket...but not today.
One day I'll reboot but not in these days.

A test for me would be to let Veeam create a new datastore on the fly when using Surebackup, using a different name.

Really, many thanks for trying to help me giving links.

[Andreas Neufert]

Yes, you can rename the VBR server and reboot it (to start all Veeam services again).
Maybe as it is a temporary workaround do it differently.
Create a VM or use another Windows Server to use this server as mount server (including the vPOWER NFS) and select this server at the Repository properties wizard. Potentially you need to first add it as managed server before you can select it there.

This way the NFS share is started from another server. Then when you rebooted the ESXi you can just turn back the mount server to original without bigger changes.

[Servior]

Most „busy“ errors are some open files or references in some VM. I think you checked already, but is there some reference (disk attached to a VM, ISO specified in a VM configuration, etc.) visible?

If you use templates (directly visible in inventory or in content library), check these as well. You have to convert them into a VM or check these file contents manually.

Running RVTools and view the output is very helpful to find such ghost entries.

[m.costantino]

Hi Servior,
just installed RVTools, where should I check ?

PS: As I wrote, the datastore created by Veeam is inaccessible

[Servior]

Hi again,

the datastore can be inaccessible, yes. This doesn't mean that you don't have a file referenced somewhere in a vm configuration. In RVTools you have to check each tab for the veeambackup datastore name. For example:

In the vInfo tab the "Path", "Log directory", "Snapshot directory", "Suspend directory" columns.
vDisk - Path Column.
vCD - Device Type Column.

I would check the vHealth tab Messages for zombie files.
Easiest approach: Export it as Excel and search the complete workbook for the veeambackup datastore name. If there is any, remove it (like attached as ISO into a VM DVD drive) and try to unmount the datastore again.

[m.costantino]

No references to that datastore...Searched manually with RVTools and in the exported Excel, no way
I'm going mad...Opened a ticket with Broadcom 4 days ago, still no answer

[Servior]

If you want, you can share the RVTools export via private message and I take a look myself.

[dabrigo]

Also, if you have HA enabled, check in the Cluster HA settings that it's excluded from the Datastore for Heartbeating.

[m.costantino]

If you want, you can share the RVTools export via private message and I take a look myself.

Thank you very much, I'll let you know if my "boss" will allow me to send that data outside.

[m.costantino]

A "little" update.
I've deployed another VM (different hostname), reinstalled Veeam and imported the configuration from the old Veaam.
Everything is now working perfectly (another datastore is created on the fly and it works perfectly!).
The old datatore is still there, I'll try to see if it disappear when i can reboot the host.

I would like to say thank you to everyone that helped me, i know that it's not a Veeam issue.

[RubinCompServ]

maybe I missed this earlier, but if you reboot the Veeam server, VMware should see the NFS connection drop and change the datastore to "(Inaccessible)", at which point I'd think you could remove it.

[m.costantino]

Hi Rubin, already tryed also powering off the Veeam server...no way.
At the moment I've that datastore lying on my esxi doing nothing, no way to unmount/remove

[m.costantino]

Happened again today...
Now I have 2 un-mountable NFS datastore.
Nothiung changed in these days, really...NOTHING!
It SEEMS that Veeam for some reason leave the datastore in that state.
I've renamed the Veeam hostname and now I'm working on a third NFS datastore with Surebackup.
I cannot find where's the problem...ESXi? Veeam?

[RubinCompServ]

Is it the same host/vcenter?

[m.costantino]

Yes, exactly the same host.
Last timeI reinstalled Veeam on another VM with a different hostname, this time I simply changed the hostname of the Veeam VM.
It's really interesting that it it seems that i'm the only one with this "problem".

[RubinCompServ]

Do you have a different host and/or vcenter you can test with? It sounds to me like it's VMware preventing you (and Veeam) from unmounting the datastore.

[Andreas Neufert]

Did you give Veeam a full admin user on the vcenter or did you define the role yourself with the permissions. Maybe our vcenter user just have not enough rights there?

[m.costantino]

Do you have a different host and/or vcenter you can test with? It sounds to me like it's VMware preventing you (and Veeam) from unmounting the datastore.

Unfortunately not...no other hosts to test.

[m.costantino]

Did you give Veeam a full admin user on the vcenter or did you define the role yourself with the permissions. Maybe our vcenter user just have not enough rights there?

Full permission to the ESXi host.
The "problem" is random. If I did not grant full admin, the problem (I think) should be "persistent".

[Andreas Neufert]

I think opening a support ticket and let someone read through the logs on both VMware and Veeam side should be a good idea.'
As well check on if there are any ESXi host and vcenter updates are available for your version.

Did you had a chance to reboot the ESXi host now?