Instant Recovery failed with 'socket open failed (21)' error message

Hello Smiley Happy

I am working on IR_VMware feature with NBU 8.0
My Environment is:
Master Server: Windows 2012 R2
Media Server : HPUX host
Client Server : RHEL VM
VMware ESXi server version : 6.5


Before starting IR, I have renamed VM and Powered off in ESXi server.

I have backed up RHEL VM successfully. while performing Instant Recovery I am facing "socket open failed (21)" error.

For restore I am executing following command on NBU Master (windows host) host.

nbrestorevm -vmw -ir_activate -C RHEL_VM -backupid RHEL_VM_1000001836 -temp_location ABC<created NFS datastore on ESXi server>

JOB Details:

VM Instant Recovery Job detail:

Aug 9, 2017 4:07:26 AM - Info bprd (pid=6456) Found the backup image RHEL_VM_1502171836 supported for Instant Recovery
Aug 9, 2017 4:07:35 AM - requesting resource @aaabC
Aug 9, 2017 4:07:35 AM - granted resource MediaID=@aaabC;DiskVolume=AAA_HPUX_CoE_IR;DiskPool=HPUX_CoE_IR;Path=AAA_HPUX_CoE_IR;StorageServer=10.X.X.X;MediaServer=HPUX_HOST
Aug 9, 2017 4:07:47 AM - Info bprd (pid=6456) Releasing the allocated resources
socket open failed (21)

 

ACTIVATE Instant Recovery Job detail:

Aug 9, 2017 4:07:45 AM - Info bprd (pid=6456) Waiting for Instant Recovery job to end
Aug 9, 2017 4:07:45 AM - Info bprd (pid=6456) Starting to export a backup image on the media server HPUX_HOST
Aug 9, 2017 4:07:47 AM - Error bprd (pid=6456) Failed to export a backup image on the media server HPUX_HOST
socket open failed (21)

Tags (2)
15 Replies

Re: Instant Recovery failed with 'socket open failed (21)' error message

Is HPUX backup host supported one for vmware? I doubt

All 64 bit only: 
Windows Server 2012 R2, 2012, 2008 R2 
Red Hat Enterprise Linux (RHEL) 7.0, 6.6 
NetBackup Appliance 3.0 
SUSE Linux Enterprise Server (SLES) 12, 11.3

after https://download.veritas.com/resources/content/live/SFDC/116000/000115691/en_US/nbu_80_scl.html?__gd...

Re: Instant Recovery failed with 'socket open failed (21)' error message

Can you please share all the steps that you have followed before you issued the command?

The NetBackup for VMware Administrator's Guide   contains quite a bit of steps and pre-reqs that need to be checked beforehand.

Highlighted

Re: Instant Recovery failed with 'socket open failed (21)' error message

 

Steps followed before restore:

1.Created Virtual Server Machine (i.e. added Vsphare Center as VSM),
2.Created Storage Server,Disk Pool, Storage unit (using StoreOnce Catalyst Store)
3.Created VMware Policy
4. backed up VM (RHEL) using VMware policy.
5.Renamed VM in ESXi server.
6.Powered off VM
7.Created temporary NFS datastore for restore.
8.executed restore command.

Re: Instant Recovery failed with 'socket open failed (21)' error message

There is a whole list of steps and pre-reqs that you have missed.

Please see this topic in NBU for VMware manual:

Requirements for Instant Recovery for VMware

Re: Instant Recovery failed with 'socket open failed (21)' error message

Did you configure a restore host that is not your HP-UX media server?

The restore host can be on Windows or Linux.

■ For a restore host that is separate from the NetBackup master server or media server: You must add the restore host to the list of servers that can access the master server.

I believe you are using the HP-UX which is not supported.  Might be easier if you made the Windows Master server the Backup Host and the Media Server for quicker IR configuration.

Re: Instant Recovery failed with 'socket open failed (21)' error message

@PBY

Ensuring that the Windows Master Server is selected as Restore Host is important - similar to your post of a week ago.

Before we get to the Restore Host - have you gone through the other requirements (as per the manual)?
e.g. NFS on ESX server, NFS on the Restore Host and media server (NFS should be fine on Unix media server).

The thing that worries me is this requirement:

The NetBackup media server platform must support Granular Recovery
Technology. See the NetBackup Enterprise Server and Server OS Software
Compatibility List

My problem is that I cannot find this information in the SCL. 
Maybe @eduncan can point us in the right direction?

 

Re: Instant Recovery failed with 'socket open failed (21)' error message

Hi,

As was mentioned before is very important check all requirements for instant recovery for VMware, did you do?

Other think, I found this TB below, and there is a similar problem to you, did you already check?

Look

TCP port 7394 (nbfsd) and 111 (portmap) from the target ESX server to the media server was not open.

 NetBackup Vmware Instant Recovery "Failed to export a backup image on the media server" - socket open failed 21 - https://www.veritas.com/support/en_US/article.000115596

 

Regards,

 

Thiago

 

Re: Instant Recovery failed with 'socket open failed (21)' error message

Thank you everyone for replying!

@eduncan

My Environment is:
Master Server: Windows 2012 R2 
Media Server : HPUX host
Client Server : RHEL VM
VMware ESXi server version : 6.5
my restore host is RHEL VM.

 

@Marianne

I have restarted NFS services on all hosts/servers.
Previously I tried GRT with HPUX (i.e. media server) & it worked.

Re: Instant Recovery failed with 'socket open failed (21)' error message

Your restore host should not be a Linux VM.
It should be the Windows Master server (same concept as backup host).

Re: Instant Recovery failed with 'socket open failed (21)' error message

still facing the same issue.

Re: Instant Recovery failed with 'socket open failed (21)' error message

At this point we would at least need the bprd log to help move this forward, or you can engage NetBackup support for assistance.

Re: Instant Recovery failed with 'socket open failed (21)' error message

NBU Master Server: WINDOWS_HOST
NBU Media Server : HPUX_HOST

bprd logs:

03:15:00.538 [12906.1] <2> bprd: INITIATING bprd (VERBOSE = 0): NetBackup 8.0 2016110921 on HPUX_HOST
03:15:00.539 [12906.1] <2> bprd: Now initializing logging for libcorbaobj
03:15:00.539 [12906.1] <2> bprd: the request timeout value is 300 seconds
03:15:00.541 [12906.1] <2> nbapp_info_init_ex: path=/etc/nbapp-release
03:15:00.541 [12906.1] <2> read_keyvalue_file: failed to open file: "/etc/nbapp-release", status: 12
03:15:00.541 [12906.1] <2> nbapp_info_init_ex: read_keyvalue_file failed, path:/etc/nbapp-release, status: 12
03:15:00.542 [12906.1] <2> bprd: HPUX_HOST is not the primary server WINDOWS_HOST...exiting

Re: Instant Recovery failed with 'socket open failed (21)' error message

bprd runs on the master server.

If the log folder does not exist, you need to create it and restart NBU.

After next failure, copy the log file to bprd.txt and upload here.

Re: Instant Recovery failed with 'socket open failed (21)' error message

Hello !

Master Server: Windows 2012 R2
Media Server : HPUX host
Client Server : RHEL VM
VMware ESXi server version : 6.5

I have reinstall NBU on Media Server (i.e. HP-UX host)

Now portmapper and nbfsd services are not running

Re: Instant Recovery failed with 'socket open failed (21)' error message

Why did you reinstall NBU on the media server?

Can you confirm that the VMware backup resides on disk on the media server?

We have asked you for bprd log on the master server. 
Have you created bprd log folder on the master followed by NBU restart?

Only nfsd needs to be running on a Unix media server - that is normally enabled by default at OS-level.

See this extract in the manual:

Configuring a UNIX media server and Windows backup or restore host for Granular Recovery
Technology (NetBackup for VMware)
For backups and restores that use Granular Recovery Technology (GRT), perform the following configuration if you use a UNIX media server and Windows restore host:
■ Confirm that your media server is installed on a platform that supports granular
recovery. See the NetBackup Enterprise Server and Server OS Software
Compatibility List:
www.veritas.com/docs/TECH59978
No other configuration is required for the UNIX media server.

Enable or install NFS on the restore host.
See “Enabling Services for Network File System (NFS) on a Windows 2012 restore host (NetBackup for VMware)” on page 323.


From what I see in the manual, nbfsd is called by nbrestorevm command:

Run the nbrestorevm command* to access the virtual machine from its backup
image. The NetBackup File System Service (NBFSD) on the media server
accesses the backup image file system and mounts the image as an NFS
datastore. The datastore becomes accessible to the ESX host where the
virtual machine is to be restored.
On the same command, select a temporary datastore that is accessible to
the ESX host.