BMR Error when trying to recover Solaris 10 server

[tarun_konduri1]

Hi

I am trying to recover Solaris 10 server. I have created the CD Based SRT with Solaris 10 update 6 version. I have burned it, loaded it in CD ROM and started Prepare to Restore & the job is queued in task list. This is a straight forward BMR Restore i.e., trying to recover same server with same H/W.

Master: Windows 2003, NBU version: 6.5.5

Media/Boot Server: Solaris 10 update 6, NBU version: 6.5.5

Client Server: Solaris 10 update 6, NBU version: 6.5.5

Booted the server using the CD Based SRT and all the required details like client name, it's IP, subnet, Master name, it's IP, etc and it ran for sometime and failed with below error.

Starting Bare Metal Restore...

The configuration settings were added successfully.

Enter the client's name : rigbmrpoc01

Enter the NetBackup master server name [rigback01] :

Enter the NetBackup master server IP Address (dotted decimal form) [10.101.80.19] :

Enter the client's default gateway (dotted decimal form) : 10.101.80.253

add net default: gateway 10.101.80.253: entry exists

Reconfiguring /dev and /devices for Bare Metal Restore

V-128-749 Operation failed due to network problem.

[Error] V-122-3 Operation failed.

ERROR: Failed to retrieve "/tmp/0A6550FA.restore".

V-128-749 Operation failed due to network problem.

[Error] V-122-3 Operation failed.

Bare Metal Restore has failed.

Now, you will be dropped to a shell prompt.

You may reboot the system when you have completed diagnosis.

So I checked the tmp file and found the below entries.

# cat 0A6550FA.conf

bmrHosts="192.168.80.19 rigback01,10.101.80.70 rigback02"

export bmrHosts

clConsole=/dev/console

export clConsole

clName=rigbmrpoc01

export clName

clNodeName=rigbmrpoc01

export clNodeName

clOs=sol

export clOs

logging=YES

export logging

newConfig=

export newConfig

onEpError=1

export onEpError

runEp=NO

export runEp

runMode=RESTORE

export runMode

serverName=rigback01

export serverName

isMediaServer=NO

export isMediaServer

srtNbuType=CLIENT

export srtNbuType

It seems that BMR Restore is trying to contact 192.168.80.19 IP instead of 10.101.80.19 IP of Master server. I checked the hosts entries of Master, Media but didn't found this 192.x entry in any of these. I am confused why it is picking up that IP. We used to have two interfaces previously one 192.168.80.19 IPanother one is 10.101.80.19 IP but now only 10.101.80.19 is active. I don't know which entry should be changed so that it should contact 10.101.xx.xx IP. Is this related to SRT, if so which file should be modified so that it should contact active IP.

I really appreciate if someone let me know what needs to be done. Let me know if you need further details.

Thanks & Regards,

Tarun Konduri.

[mph999]

I'd have to look into this (very late here ), but first thoughts are ...

Starting Bare Metal Restore...

The configuration settings were added successfully.

Enter the client's name : rigbmrpoc01

Enter the NetBackup master server name [rigback01] :

Enter the NetBackup master server IP Address (dotted decimal form) [10.101.80.19] :

Enter the client's default gateway (dotted decimal form) : 10.101.80.253

add net default: gateway 10.101.80.253: entry exists

I'm not a BMR 'expert' but if I remember this file 0A6550FA.conf is the client config pulled from the BMR DB during the 'prepare to restore' which should be passed to the client .

The filename is the client IP in hex - which translates to 

10.101.80.250

Given this :ERROR: Failed to retrieve "/tmp/0A6550FA.restore".

You could be onto something as the bmrHosts entry is used for the restore.

Assuming that the restore files were not cleaned up by the failure, the following commands need to be run on the Master Server and the output files submnitted for review:  If these commands don't work, run another prepare to restore.

bmrc -o pull -res info -client rigbmrpoc01 -source 0A6550FA.info -destination /tmp/rigbmrpoc01.info

bmrc -o pull -res info -client rigbmrpoc01 -source 0A6550FA.conf -destination /tmp/rigbmrpoc01.conf

bmrc -o pull -res info -client rigbmrpoc01 -source 0A6550FA.hosts -destination /tmp/rigbmrpoc01.hosts

bmrc -o pull -res info -client rigbmrpoc01 -source 0A6550FA.restore -destination /tmp/rigbmrpoc01.restore

bmrc -o pull -res info -client rigbmrpoc01 -source 0A6550FA.bp.conf -destination /tmp/rigbmrpoc01.bp.conf

the *.info file describes the client name, ip address, and configuration name used 

the *.conf file will describe what servers are to be used for the restore (the bmrHosts line entry). 

The first entry will be the Master Server ip and hostname.

The next entry is the Boot Server.

the *.bp.conf file will be the working client's bp.conf file for the restore.

the *.hosts file will become the /etc/hosts file for the running client.

Make sure these files are correct, 

At this point, we have only looked, no changes are made.

If you need to edit you can push them back into the DB like this

NOTE:  This is 'theory only' I haven't tested this ... 

bmrc -o push -res info -client rigbmrpoc01 -source /tmp/rigbmrpoc01.hosts -destination 0A6550FA.hosts

Very important - keep a copy of the files you export.  Only edit and 'put back' a copy.

martin

[mandar_khanolka]

>> ERROR: Failed to retrieve "/tmp/0A6550FA.restore".

Are you configuring the client IP details in restore env for the correct client NIC? Is that the NIC with hex address 0A6550FA.

Sometimes it happens when client machine is booting over other NIC on that machine than the one BMR has considered as primary during PTR operation. Accordingly it creates the NIC_HEX_ADDR.* files during PTR and puts into BMRDB.

Also when it falls to the command prompt with restore failure, check /etc/hosts file to confirm the host-ip details set.

Also do you have two subnet configuration where backup happened through one and recovery via other subnet?

Thanks.

Mandar