cancel
Showing results for 
Search instead for 
Did you mean: 

oracle Solaris recover in Ldom

Deyaa
Level 4

Hi all ;

we have SPARC T4-2 servers ,configured OVM for sparc on it . one control domain and 3 LDOMs and these LDOMS configured as media servers to backup directly to NAS Storage.

we have the below environment :

* Master server

            - Hostname: Netbackup-bck , Solaris 10

            - netbackup 7.6.1.1

* Media servers :

     1- control domain

            - hostname :control-domain-bck , solaris 11

     1- LDOM 1 " database server ":  

           - hostname :dbserver-bck, solaris 11

     2-  LDOM 2 "application server ] :

           - hostname : appserver-bck , solaris 10

      3- LDOM 2  " file server ":

            - hostname : fileserver-bck , solaris 10

refering to this article :https://support.symantec.com/en_US/article.TECH162994.html

All NetBackup components supported with Solaris 11 SPARC physical servers are supported in a Solaris 11 LDoms Control Domain and I/O Domain with the exception of Bare Metal Restore (server or client). Guest domain support is limited to standard client, database agents, master server and disk media server

BMR is not supported on LDOMs , so i backup the control domain and each ldom like normal os backup means backup " /  " .

the question here , if there is a failure in LDOM and need to recover the LDOM OS backup . how to restore it ??

does i need to install solaris again and netbackup agent and restore the backup to overwrite the existing files or what ??

please provide Step by Step how-to if possible to test the restore.

thanks

Deyaa 

1 ACCEPTED SOLUTION

Accepted Solutions

Jaime_Vazquez
Level 6
Employee

Your statement about BMR and Solaris LDOMS is not totally correct.

 

BMR supports guest OS instances on Solaris 10 LDOMS with UFS attached.

Although not stated to be supported, at least one user on this forum has gotten BMR to work on Solaris 11 LDOMS, again with UFS attached.

BMR works at the LDOM guest  domain level.

See this:

Statement of Support for NetBackup 7.x in a Virtual Environment (Virtualization Technologies)   
http://www.symantec.com/docs/TECH127089

 

View solution in original post

4 REPLIES 4

Jaime_Vazquez
Level 6
Employee

Your statement about BMR and Solaris LDOMS is not totally correct.

 

BMR supports guest OS instances on Solaris 10 LDOMS with UFS attached.

Although not stated to be supported, at least one user on this forum has gotten BMR to work on Solaris 11 LDOMS, again with UFS attached.

BMR works at the LDOM guest  domain level.

See this:

Statement of Support for NetBackup 7.x in a Virtual Environment (Virtualization Technologies)   
http://www.symantec.com/docs/TECH127089

 

Deyaa
Level 4

in my environment , both solaris 10 and solaris 11 use zfs for root pool . forget BMR now.

i need to know how to restore solaris without using BMR , as i took backup of / file system and need to restore it in case there is a disaster.

 

BR

Deyaa 

jim_dalton
Level 6

This is a good question, interested to know the answer. What do Oracle/sun advise?

If you break it down, the safest route would be:

-separate data , app from OS

-backup data , app in an appropriate manner.

-backup OS.

This way at least you have your all important data, you just need to figure how to go from a freshly installed server to the server with your latest OS on it. Traditionally one could do this by running off an arbitrary OS either on part of the OS disk or from bootable media with netb client installed and then reinstalling the OS onto  another part of disk and then setting the boot disk to the restored slice but have yet to investigate this when root is zfs. I'm certain Oracle will have covered this.

Jim

Noor_Toorabally
Level 4

Hello,

 

These are the steps I would suggest for solaris 10:

 

1. Reinstall solaris 10 and netbackup client on a slice (or disk)

2. Restore the / on a root disk (or a root slice) mounted as /a for example (exclude the /tmp while restoring).

 

3. After the restore completed, the following solaris procedures need to be done:

touch /a/etc/mnttab

chmod 444 /a/etc/mnttab

mv /a/etc/path_to_inst /a/etc/ORIGINAL-PATH-TO-INST

rm /a/etc/path_to_inst*

rm -rf /a/devices/*

rm -rf /a/dev/*

cd /devices; find . | cpio -pmd /a/devices

cd /dev; find . | cpio -pmd /a/dev

 

Install the bootblk on the new root:

cd /a/usr/platform/`uname -i`/lib/fs/ufs

/a/usr/sbin/installboot bootblk /dev/rdsk/rootdisk_dysx

 

 

Adjust network interface parameters:
echo hostname > /a/etc/hostname.interfacename

 

 

Further adjustments required:

mkdir /a/tmp; chmod 1777 /a/tmp

bootadm update-archive -v -R /a

 

 

Go to OBP to perform a reconfiguration boot on the restored root_disk (slice)

init 0

boot root_disk -ras

 

Then reboot to multiusermode after the above reconfiguration boot.

 

Gud Luck