cancel
Showing results for 
Search instead for 
Did you mean: 

nbrestorevm -R

cc1702004
Level 4

Hello,

We need to do a VM instance recovery test for a customer. May I know for the nbrestorevm -R (rename_file).

Can anyone please share with me how to use this (rename_file) ?

(1) What file extension should this file ended with ? Can it be a txt example, RenameConf.txt..or it should end with other extension ?

(2) Where should this file be placed ? Can I placed it on any location, example C:\temp, or it should be one of the Netbackup location, example, the Netbackup log directory ?

Can someone please be kind to share with me?

Thank you

26 REPLIES 26

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

@cc1702004 

I see good explanation and examples in NBU for VMware Admin Guide.

Maybe you want to have a look?

NetBackup for VMware Administrator's Guide

I did go through the options in the Netbackup VM admin guide..but not to sure of certain things return..anyway we did try to run nbrestorevm few times but failed to do the instance recovery.

Basically customer environment have one master server (running on Windows 2008) running version 7.7.3 and a 5230 nbu appliance running NBU OS 2.7.3

Run nbrestorevm -vmw -C tuaap181iis -ir_activate -temp_location PRD_DF_VMFS_02  -R "E:\Program Files\Veritas\Netbackup\logs\restoreinput.txt"

Run nbrestorevm -vmw -C tuaap181iis -ir_activate -temp_location PRD_DF_VMFS_02 -R "E:\Program Files\Veritas\Netbackup\logs\tuaap181iis.txt"

The rename file, we also tested to use only one entry to change the vmname to RES_tuaap181iis.

We attempted nbrestorevm for VM instance recovery but so far failed in our few attempt. Struggling to get things working..we are not sure what is the problem. Is it some permission problem, if yes, where should we start to check ?

 

12/03/2019 11:30:38 - Info bprd (pid=13924) Starting to export a backup image on the media server tuaap03msbackup
12/03/2019 11:30:39 - Info bprd (pid=13924) Backup image has been successfully exported on the media server tuaap03msbackup
12/03/2019 11:30:45 - Info bprd (pid=13924) Backup image has been successfully mounted on the server 10.178.130.57
12/03/2019 11:30:45 - Info bprd (pid=13924) Starting to activate a virtual machine on the server 10.178.130.57
12/03/2019 11:30:58 - Error bprd (pid=13924) Failed to activate a virtual machine on the server 10.178.130.57
12/03/2019 11:30:58 - Info bprd (pid=13924) Starting to unmount a backup image on the server 10.178.130.57
12/03/2019 11:30:58 - Info bpVMutil (pid=241539) INF - vmwareLogger: CreateVirtualMachineExAPI: SYM_VMC_ERROR: FAILED_TO_GET_DS_MOR
 

sdo
Moderator
Moderator
Partner    VIP    Certified

I googled for:

SYM_VMC_ERROR: FAILED_TO_GET_DS_MOR

...and came up with:

Virtual machine snapshots failing with Status 4239 "Failed to get datastore mor for the vmx location"

https://www.veritas.com/support/en_US/article.100014401

...which says:

Solution

Added the Permissions for the User Account at the vSphere level

I checked with my colleague the account used to connect to the Vcenter server is already Administrator account with all the privileges..so not sure how to proceed to check for permission issues.

sdo
Moderator
Moderator
Partner    VIP    Certified

1) Exact version of ESXi, please?

2) Exact version of vCenter, please?

3) Can we see the contents of the XML file that bpVMreq created, please?

4) Can we see the contents of the rename file please?

5) Have you first proven that a normal restore works before trying a "instant + rename" type of restore?

Hello SDO,

I will check on your enquiries..but I also want to consult you one thing. The incumbent operator mentioned they use the NBU appliance (media server) as NFS. Do we need to mount anything in the NBU appliance before running the VM instance recovery :

Example : -temp_location : I understood that the backup image needed to be mounted on a NFS datastore..so if someone said they used the NBU appliance as NFS to do the VM instance recovery, what should we check next?

For example : the -temp_location is a location of a VMFS datastore in ESX host.

n nbrestorevm -vmw -C tserver1 -ir_activate -temp_location PRD_DF_VMFS_02  -R "E:\Program Files\Veritas\Netbackup\logs\restoreinput.txt"

 

Thank you

 

 

sdo
Moderator
Moderator
Partner    VIP    Certified

If you read this:

About Instant Recovery for VMware

https://www.veritas.com/content/support/en_US/doc/21902280-127283730-0/v75741283-127283730

...then let's think about what we know... that the restored VM has to be running on ESXi (as it cannot run anywhere else), and so ESXi has to be able to read the VM backup... and so we can infer that NetBackup provides an NFS export for ESXi to mount... so, to answer your question, then no, you yourself do not have to pre-mount or pre-export anything... because it is NetBackup that exports to ESXi, and it is ESXi that mounts the export... and we know that the temporary storage area within ESXi environment is used for caching writes and updates as the VM runs... and anything old that the VM needs to read is read from the read-only NFS data-store which is auto-presented out from NetBackup... all until you then storage vMotion the VM... which then combines the "temporary write area" with the "old read-only backup area" all of which effectively migrates the newly restored VM wholly onto VMware managed storage.

I've not had to debug this myself, so you'll either need to open an official support case, or hope that some kind soul on here will direct you through log capture and analysis..

.

...or...

.

Maybe you could start here:

How to troubleshoot backups/restores on SLES/RHEL/Appliance VMware Backup Host

https://www.veritas.com/content/support/en_US/article.100038096

...and see what you can find first, and then come back to us if you get stuck during the log analysis.

 

sdo
Moderator
Moderator
Partner    VIP    Certified

You still haven't told us what the exact version (and update / patch level) is for the VMware vSphere ESXi that you are running.   So, technically, you still have not proven (from the NetBackup SCL) whether your interop is actually a supported configuration or not.

Hello,

Will check and let you know..Netbackup version is 7.7.3 just EOS in May 2019. Now we are taking over operation from another vendor.

Hello SBO

 

ESX host  :6.5

ESX vcenter : 6.7

sdo
Moderator
Moderator
Partner    VIP    Certified

no patch levels ?  "U" levels ?

sdo
Moderator
Moderator
Partner    VIP    Certified

I've cross-checked the interop, and NetBackup v7.7.3 was compiled with the VMware VDDK v6.0, and this version of VMware's VDDK supports ESXi 5.1 to ESXi 6.0.

So, technically your interop of NetBackup v7.7.3 is unsupported for use with VMware v6.5 or v6.7.  This is not a NetBackup limitation.  This is a VMware VDDK limitation.  i.e. it is VMware themselves that restrict / limit the interop support to a narrow range of versions, and there's nothing that you can do about it except to upgrade NetBackup to a version of NetBackup which contains a version of the VMware VDDK which supports your deployed version of VMware vSphere.

sdo
Moderator
Moderator
Partner    VIP    Certified

how to cross-check NetBackup version to VMware VDDK version to VMware vSphere versions and features :

https://vox.veritas.com/t5/NetBackup/How-to-find-VMware-API-version-used/td-p/611999

 

Hello SOD,

I don't think so. Customer feedback their incumbent vendor had done VM instance recovery using this Netbackup 7.7.3 environment..so I cannot tell that to customer unless we are very sure this is the root cause

https://www.veritas.com/content/support/en_US/doc/NB_70_80_VE

NetBackup 7.7.3
NetBackup 7.7.3 supports vSphere 6.5, subject to the following limitations:
■ Backup and restore of encrypted VMs are not supported.
■ VMs that reside on VMFS 6.0 cannot be backed up by using the san transport
mode. As a workaround, use nbd, nbdssl, or hotadd transport mode to back up
such VMs. Note that the san transport mode for VMs on VMFS 6.0 is supported
by NetBackup 8.1.
Note: NetBackup support for VMFS 5 in vSphere 6.5 environments is unchanged.

sdo
Moderator
Moderator
Partner    VIP    Certified

Well you might be in luck, you might not.  Just because a particular feature is not explicitly called out as not being supported then this does not necessarily imply that any given feature is actually explicitly supported.  I guess you just have to hope that whichever VMware vSphere v6.5 features that are required to support instant+renamed VM recovery are indeed supported by the VDDK 6.0 in your NetBackup v7.7.3.  Or, you could try to cross-check for peace of mind... i.e. have you checked the VDDK 6.0 release notes in detail, and also checked in detail all of the subsequent versions of VDDK release notes to make sure that nothing which might be related to instant VM recovery is not explicity recinded by later versions of the VDDK which is embedded on the vSphere side ?

(i.e. some later versions of VDDK do not support all features in all earlier versions of VDDK, and you can only find this out by reading the release notes of all of the interleaving VDDK versions - so, do you see how, for example, the VDDK v6.5.2 EP1 in vSphere v6.5 U2 might not support some back level features of the VDDK v6.0 in NetBackup v7.7.3 - and so this is another reason why you need to know exactly what your vSphere and vCenter patch levels are).

We still don't know your actual patch levels of vSphere and vCenter ?

Does a normal simple VM recovery work ?

Did you follow the troubleshooting link that I posted earlier ?

Hello SDO,

Thank you for being patience with me. I haven't have opportunity to work on the logging as I need to do more reading on the logging. 

As of now our concentration is checking on the ESX hosts. The problem now I am facing is the NBU appliance is like a black box to us. In our own environment we don't use NBU appliance. In this customer environment, they do. The master server, a Windows server has no client for NFS enabled. We checked with the customer's existing vendor, they said they use the NBU appliance as NFS server. So I have two question pertaining to this :

1) How do we check whether NFS is running on the NBU appliance ?

2) I read there is option to mount remote NFS ? Do we neeed to do this remote mounting NFS in the NBU appliance in order for VM instance recovery to work?

Hello SDO,

I would like to ask one more question, as highlighted, the customer environment consists of one Windows master server 7.7.3 and a media server nbu appliance 2.7.3/7.7.3. The Windows master server has no client for NFS service enabled. Is the client for NFS service required on the Windows master server?

I see in the vmware policy for that client, the backup host is the nbu appliance.

I ran the nbrestorevm command from the Windows master server.  Do I need to use the parameter 

-vmproxy restore_host, where restore_host = nbu appliance ? Do I need to explicitly specify the restore host since the Windows Master server has no Client for NFS service enabled ? 

sdo
Moderator
Moderator
Partner    VIP    Certified

1) I would have thought something like this :

ps -ef | grep -i "nfs"

service nfs status

service nfslock status

service rpcbind status

But you shouldn't have to do this anyway.  Why do you think you need to do this?

2) Certainly the appliance can export a mount (for example the software incoming area and the logs area), and can also mount an export from another host (indeed I sometimes do this to drop the catalog backup DR file somewhere else).  But again, you shouldn't need to do any of this yourself for VM instant+rename recovery.  Why do you keep asking this?

(top tip... do *not* mess around with any NFS (exports or mounts) manually from the shell CLI otherwise you will very likely break the CLIsh scripted environment - which would also very likely break the NFS elements of instant recovery)

3) Have you tried a simple VM recovery to prove that it works?

4) Why won't you tell us your actual vSphere and vCenter patch levels?

5) I suspect that your only way to find out what is failing is to enable logging and to inspect those logs.

sdo
Moderator
Moderator
Partner    VIP    Certified

Yes, to me it would seem reasonable to try using the -vmproxy switch.