Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

BE2012 WAN optimized deduplication with a communication or Central Adminstration Server (CAS) failure

Created: 23 Oct 2012 | 5 comments

Version 14.0 1798

Problem:
I’m having trouble getting past some issues with this structure.  It comes down to either optimized deduplication works, but I can’t use the MBES/MMS when the CAS is down or both sites are fault tolerant but I can’t setup optimized deduplication between them.

Goal:
My goal is to use optimized/shared deduplication to replicate server backups between my two sites and be able to restore those backups and local backups during a vpn failure or site failure.  If site A is down, I want to recover site A servers to site B.  If site B is down, I want to recover site B servers to site A.

Details:
I thought that I could place a Backup Exec 2012 server at each site and use optimized deduplication for this, but I ran into a few problems.  Optimized deduplication, or shared deduplication as it is now called, seems to require the Central Administration Server Option and this has the following options for its Managed Backup Exec Servers:
  1. Centrally managed Backup Exec server
  1a. Centrally managed Backup Exec server with Unrestricted access to catalogs and backup sets for restore
  2. Locally managed Backup Exec server

Options 1 & 1a:
With the “Centrally managed Backup Exec server” option chosen for the Managed Backup Exec Server.  Having a CAS with a deduplication folder at site A and a MBES with a deduplication folder at site B works and staged jobs will run from one deduplication folder to the other.  The problem here is that if the CAS is down or the VPN is down, I can’t even bring up the BE2012 console on the MBES let alone run jobs or restores.  The MBES is totally dependant on a persistent communication link with the CAS under the “Centrally managed Backup Exec server” “Managed Backup Exec Server Configuration” option.

Option 2:
MBES that is a “Locally managed Backup Exec server”.  While the CAS and MBES can still function without each other under this scenario, I can not share the MBES deduplication folder with the CAS.  So I can’t do optimized deduplication between the two.
 

Am I wrong about optimized deduplication requiring CAS + Centrally managed Backup Exec server?  If so, please explain how to set it up.

Is there a way to use the Centrally managed Backup Exec server when the CAS is down?  I don’t understand how I’m supposed to recover if the CAS is gone.

Comments 5 CommentsJump to latest comment

Colin Weaver's picture

I am not 100% sure that we designed bi-directional Opt Duplicatiion into the product. It is certainly not mentioned in the white papers on use of Private Cloud. It will/should work but is almost certainly affacted by unreliable links and where the CASO server is located.

On the subject of Private Cloud. Have you enabled this against one of the servers as it can have an effect on the catalog handling in addition to the options you gave above.

Opt Dup does need the CASO. If you think your links might be unreliable then the recommendation is put the CASO at the critical/production site so you at least get your initial backup still, do not put the CASO at the DR site. However your bi-directional Opt Dup setup has a production site at both ends meaning this recommendation is only effective for one location.

Mike55's picture

If there was a way for me to run deduplication to deduplication jobs across sites with CAS/Locally Managed Servers, I would do that.  I haven’t found that, so I’m trying to work with the Centrally Managed type of managed servers.

With a Central Administration Server at one site and a Centrally Managed Backup Exec Server at the other, I’ve ran deduplication to deduplication jobs in both directions.  For example if I backup a VMware VM to the CAS deduplication folder with a deduplication stage to the Centrally Managed Backup Exec Server deduplication folder, I’m able to restore the VM from the Centrally Managed Backup Exec server.  It works in the other direction as well, staging backups to the CAS.

I read through the Backup Exec Private Cloud Services Planning and Deployment Guide since the Backup Exec administrators guide only contains the following.

Private cloud server

Enables a managed service provider to locate a Backup Exec server in its data center, and then configure it for a CASO environment with other Backup Exec servers that are located across theWANat the managed service provider's customer locations. As an alternative to shipping tapes off-site for storage, backups can be run and stored locally, and then copied to the cloud server's deduplication disk storage device. Additionally, this feature can be used by customers with widely distributed networks who want to use Backup Exec servers in remote offices for local backups, and then copy the backup sets to a Backup Exec server that is located in a central data center. This option is part of the Cloud Services for Backup Exec feature.

Even after reading the guide, I’m still not sure what the actual enabling the Private cloud server setting does.

Those three situations in the guide don’t really fit what I’m doing.  You’re right, I have production that runs at both sites.  Each site is the disaster recovery location for the other.  It doesn’t do me much good to have my backups sitting at the Managed Backup Exec server if I can’t access them because the CAS is gone.  This is similar to the Private Cloud “offsite copy to cloud managed Backup Exec server configuration” where you copy the backups from your local CAS to a Managed Backup Exec server in the cloud.  Say your local site with the CAS gets wiped out.  How do you access the backups in the cloud?

I found the following recovery procedure for a Managed Backup Exec Server in the guide on page 40.  I haven’t tried it yet, but I think that the index/catalog portion could end up taking a very long time to get the Backup Exec functional.  After all of that then I would be able to run backups jobs and to start recovering servers.  It still surprises me that I lose all access to the Managed Backup Exec Servers when the CAS link or server is down.  It really seems like I should have access to the managed server with centrally managed functions unavailable.  Instead there is nothing, existing jobs don’t even run.

Restoring data from a managed Backup Exec server in the event of a central administration server failure

If a hardware failure or other disaster affects your central administration server, it makes it impossible for your managed Backup Exec server to run backup or restore jobs. You can recover the central administration server by configuring a replacement computer and reinstalling the Backup Exec central administration server. You can also, however, convert a managed Backup Exec server to a standalone Backup Exec server to restore the central administration server.

To convert a managed Backup Exec server to a standalone Backup Exec server to restore the central administration server

1 On the managed Backup Exec server, note the names and directory paths of any local disk storage.

Note: Double-click the disk storage on the Storage tab. Then click Properties in the left pane to view the storage properties.

2 If the managed Backup Exec server has its own deduplication disk storage device, note the device's name, path, logon account, and password properties.

Note: Double-click the deduplication disk storage device on the Storage tab. Then click Properties in the left pane to view the storage properties.

3 Open the Programs and Features (or Add or Remove Programs) dialog or the Uninstall a program dialog from the Windows Control Panel.

4 Select the Change option for Symantec Backup Exec.

5 In the left pane, select Additional Options, if it is not already selected.

6 Click Next until you reach the ConfigureManagedBackupExecServer panel.

7 Select the Locally Managed Backup Exec server option.

8 Click Next.

9 Do one of the following when you receive the message "Unable to contact {central administration server}. Ensure that the central administration server is running."

If the central administration server is unavailable and you want this managed Backup Exec server to be managed locally
    Click OK to continue.
If you want to retry this operation when the central administration server is running
    Click Cancel to end the procedure.

When the installation is complete, the computer is no longer a centrally managed Backup Exec server.

10 Click Next.

11 Restart the computer if you are prompted to do so.

12 Open Backup Exec and select the Storage tab.  If Backup Exec fails to connect to the Backup Exec server, restart the Backup Exec services and then try again.

13 Recreate any local disk storage by importing the original disk storage using the same names and paths that you noted in step 1.

14 Recreate any deduplication disk storage devices by importing the original deduplication disk storage devices using the same information that you noted in step 2.

Note: It might take much longer to recreate an existing storage device than it would to create a new storage device. The amount of time depends on how many backup sets the storage device contained and whether this managed Backup Exec server has access to its domain controller and DNS.

Colin Weaver's picture

Well you should be able to acces them for restore pursposes if the CAS has gone (or the WAN link is down) by setting the Managed Backup Exec Server back to Stand Alone. and then it depends on yoru catalog mode as to whetehr or not you would have to do extra steps.

The usual problem with a bi-directional opt dup setup is that whilst the CAS can still run jobs with the MBEs offline (WAN link failure etc) if the jobs are managed by the CAS then the MBE won't run any jobs meaning you not only lose your DR capability because the link is down, but the backups stop running at a local level as well.

akuma22's picture

Hi Mike,

Did you get anywhere with this? It seems we're experiencing the same problem with 2 very time consuming solutions:

My recent thread:
https://www-secure.symantec.com/connect/forums/backup-exec-replication-recovery

I'm in a position where I can try the conversion to stand alone in my lab but honestly, I don't think I'd want this to be the recovery solution we implement since it would take way too long and even then, I would also need to "repair" my MMS afterwards.

Mike55's picture

Sorry Akuma,

I've explained the issue to my manager and we now accept this flaw.  My long term solution is to waste a BE2012 CASO license and build a dedicated CAS that is replicated in such a way that I can’t lose it.

Let me know if you figure out something better.