More on WSS (MOSS) Backup and Recovery

So we are able to create our backup images like we talked about last time.

A few issues remain before our approach can be called "complete." 

  1. What's up with all those files?
  2. What to do with the back-up files as they grow in size and number
  3. How to do a restore.
  4. What to do when a restore fails. 
  5. What about configuration and Central Admin content?
  6. What about your DBAs?
  7. Migrating to a Different Farm.
  8. Gotcha's

So, using my best Google searches, I again find that our friends at TechTarget offer some pretty good content.  So we can take these in order:

1.  What's up with all those files?

Remember that, when we ran the backup, we specified and backup folder.  Remember, also, that it was a fully qualified UNC path to which I had created a share called \MyServerBackups and it was this share that needed suitable permissions for the various actors involved.

So in this folder, we'll find a file called spbrtoc.xml and a bunch of folders named using the format spbrxxxx.  The xxxx is the index number of each specific backup operation.

Inside the backup folder, you'll find some folders with GUIDs for names with a Projects folder, a Config folder, and a RegistryBlob.reg file; let's just call these "mysteries."

In addition, you'll find a spbackup.log and an spbackup.xml file.  The backup log will contain the errors encountered if you're not lucky enough to have the job complete without errors. The XML file includes metadata details about the backup operation including the top level component ID and wanrning and error counts.

Once you run a restore using a given restore folder, you'll get sprestore.log and a sprestore.xml files in the folder as well.

2. What to do with the back-up files as they grow in size and number

The restore folders are the key to your answer here.  SharePoint does not care when you put or copy the folders.  You only need the root folder with the spbrtoc.xml file and all of its component back up spbrxxxx folders.

If I had a maintenance window, I'd do a full backup to new folder and then intermittant differentials throughout the working days.  Then I'd change folders for the next full backup and diffs.  Then I could slough off fulldiff sets whenever I got tired of them taking up my disk space.

3. How to do a restore.

Okay, the restore is pretty easy working from Central Admin Operations.  Select a restore directory and check the specific content you'd like to restore and let it fly. 

4. What to do when a restore fails. 

As far as failures are concerned, I presume youi'll run into the same permissions difficulties we had with backups.  Once those are resolved, you get the job to start and the page will refresh with updates but individual segments of the operation may fail with a pointer to the backup log.

In addition, remember that the timer job will have to be deleted manually from the Central Admin Operations Time Jobs Definitions page.  Click on the Backup/Restore item and then click Delete.  Otherwise, next time you try one, you'll get thid error:

The backup/restore job failed because there is already another job scheduled. Delete the timer job from the Timer Job Definitions page, and then restart the backup/restore job.

For example, I sufferred from the fact that my Windows SharePoint Admin Service had not been started by default which makes me go "Hmmm…"   When it crapped out, I got an error in the log that said:

Error: Object Shared Search Index failed in event OnRestore. For more information, see the error log located in the backup directory.

That's nice.  So a little more digging and it told me:

InvalidOperationException: System.InvalidOperationException: This operation uses the SharePoint Administration service (spadmin), which could not be contacted.  If the service is stopped or disabled, start it and try the operation again.

So, in my Control Panel Services tool, I find the service and started it.  I deleted the timer job and kicked off the restore again.

This time, I got an error on the primary.
port 80" web app, and the SSP.  In the logs, I found this error:

SPException: The specified component exists. You must specify a name that does not exist.

5. What about configuration and Central Admin content?

In the logs, I get this message:

[Farm] The configuration database and central administration content database components cannot be restored.

I think this occurred because I neglected to tell the restore page 2 to back up to the same farm and then did not give it new farm info onto which to restore iteself.

6. What about your DBAs?

7. Migrating to a Different Farm. 

While the restore file collection can be sourced from a different farm and restored onto a new host, the SMIGRATE tool might be better suited for migrating site collections from one farm to another.

8. Gotcha's

One thing is that your site security is embedded in your site so unless your new restore host can reach the same domain controller, you're going to have trouble getting anything to work.

One thing the community is unanimous about is meddling with the contents of the backup folders.  I would hesitate to try to fool the xml into thinking that folders exist when they don't or don't when they do. 

Another is that the command line interface bypasses the Job Timer; you won't get a job to delete if you get a failure.

I'm getting some cacheing issues when I restore where my pages are not completely refreshed until I restart my browser or reset IIS.

 

TechTarget.Com

http://searchexchange.techtarget.com/general/0,295582,sid43_gci1275048,00.html

Microsoft Support:

http://technet2.microsoft.com/Office/f/?en-us/library/16a7e571-3531-4a4e-baa7-f348a9f9d1d11033.mspx

Content Deployment

http://technet2.microsoft.com/Office/en-us/library/edcdacca-8013-460e-95a0-d2b83b6cc7ef1033.mspx?mfr=true


Tags:

 
 
 

Comments are closed.