FlashBackup-Windows policy restore error(2817)

[Stanleyj]

I am trying to restore a directory that contains 350gb worth of scanned images.  Does anyone have suggestions on what could be causing me to get this error.

[Will_Restore]

See this thread for possible solution:

https://www-secure.symantec.com/connect/forums/vmdk-restores#comment-6430441

[RLeon]

In the BAR restore interface, in the "specify Netbackup machines and policy type" screen, select Flashbackup-Windows for the policy type, click ok.

Next, do not just click the "Select for Restore" button, but click the little down-arrow next to it. A list of items should show up. Make sure the topmost "Restore from Normal Backup" is selected.

[Marianne]

Please share all relevant info:

Steps taken in BAR to select filenames.
All text in Details tab of failed job.

[Stanleyj]

All settings are correct for a flash backup restore.  Normal backup, flashbackup-windows, etc...

this paticular backup runs a full on saturdays and incrementals thru out the week twice a day.  They are always successful so i know the data is there.

Steps in the BAR:

Choosing the first full, Browsing to the folder i want, then checking the box

Under restore screen:

Restore everything to different location (this is going to a new server which has been added to the client list on master)

Browsing to the location, leaving the rest defult.  Start the restore.

It takes about 20mins before this paticular restore will start processing.  (this is the only flash backup type i have)

2/18/2013 2:46:52 PM - begin Restore
2/18/2013 3:05:16 PM - restoring image p170fil09.home.gfbmic.com_1361038559
2/18/2013 3:05:16 PM - requesting resource @aaaab
2/18/2013 3:05:18 PM - Info bprd(pid=7140) Restoring from copy 1 of image created 02/16/13 13:15:59   
2/18/2013 3:05:18 PM - granted resource MediaID=@aaaab;DiskVolume=PureDiskVolume;DiskPool=dp_disk_nbuappl01;Path=PureDiskVolume;StorageServe...
2/18/2013 3:08:36 PM - end Restore; elapsed time: 00:21:44
FlashBackup-Windows policy restore error(2817)

[Mark_Solutions]

What are your client read timeouts on the media servers?

It takes a long time to read the file list and almost 20 minutes has elapsed before it tries to start the restore - perhaps it has already timedout by then?

[Stanleyj]

Timeouts: (seconds)

Client connect : 2700

Client read: 2700

Backup start: 300

Backup end: 300

media server connect: 30

us OS dependent timeouts (checked)

I only have two directories on the drive.  Image Data and sql data.  i just tried restoring the sql data folder and it is restoring without any problems.  So only the image data directory is causing the issue.  So Mark you may be on to something with the timeouts.  There is several million files in this directory.

when i say restoring sql data this is stil using the flashbackup-windows method.  Not a sql restore.

[Mark_Solutions]

maybe check the client connect and client read timeouts on the actual client too (as well as Master and all Media Servers)

[Mark_Solutions]

forgot to ask .. which version and patch level of NBU are you on?

[Stanleyj]

Sorry mark just saw your comment.  I adjusted the client read timeouts and im still getting the same errors.  I am running 7.5.0.4 NBU and 2.5.1 on the 5200 appliance.

It definitly looks like a timeout issue and that the job cant read thru the millions of files.  When i try and restore an incremtal it's successful but once i select the full to go with it then it fails.

[Stanleyj]

When I look at the verbose logging from the client side I am seeing this error: 14:56:19 (542674.001) Restoring from copy 1 of image created 3/23/2013 12:13:08 PM 14:59:29 (542674.001) Status of restore from copy 1 of image created 3/23/2013 12:13:08 PM = failed trying to allocate memory Does this actually mean that the backup is bad even thougth I get successful messages every night from it?