05-31-2016 05:51 AM
Hello,
i configured a optimized backup for my windows fileserver cluster drive on which i have enabled the ms deduplication feature. i did the configuration like described here: https://www.veritas.com/support/en_US/article.TECH216603
now i want to restore to another client. the client is a server 2012 r2 with enabled dedup option. when i chose to restore one particular folder i always get a "cannot write data to socket 10053" error. At the end job is failing with "24: socket write failed". But please see the attached job log:
31.05.2016 13:51:38 - begin Restore
31.05.2016 13:51:38 - restoring image swolfile01.office.company.dir_1463776795
31.05.2016 13:51:38 - requesting resource @aaaaf
31.05.2016 13:51:38 - granted resource MediaID=@aaaaf;DiskVolume=PureDiskVolume;DiskPool=PureDisk_WOL_3;Path=PureDiskVolume;StorageServer=s...
31.05.2016 13:51:39 - Info bprd(pid=8108) Restoring from copy 1 of image created 05/20/16 22:39:55 from policy FileserverCluster_Company-to-WOLL
31.05.2016 13:51:41 - Info bpbrm(pid=5364) smidcad01.office.company.dir is the host to restore to
31.05.2016 13:51:41 - Info bpbrm(pid=5364) reading file list for client
31.05.2016 13:51:46 - connecting
31.05.2016 13:51:46 - Info bpbrm(pid=5364) starting bptm
31.05.2016 13:51:51 - Error bpbrm(pid=5364) cannot start nbfsd on ssolbkp03.office.company.dir
31.05.2016 13:51:52 - Info tar32(pid=3852) Restore started
31.05.2016 13:51:52 - connected; connect time: 0:00:06
31.05.2016 13:51:52 - Info bptm(pid=2296) start
31.05.2016 13:51:52 - started process bptm (2296)
31.05.2016 13:51:52 - Info bptm(pid=2296) reading backup image
31.05.2016 13:51:52 - Info bptm(pid=2296) using 4096 data buffers
31.05.2016 13:51:52 - Info bptm(pid=2296) spawning a child process
31.05.2016 13:51:52 - Info bptm(pid=2296) child pid: 3096
31.05.2016 13:51:52 - Info bptm(pid=3096) start
31.05.2016 13:51:52 - started process bptm (3096)
31.05.2016 13:51:56 - Info tar32(pid=3852) done. status 0
31.05.2016 13:51:56 - begin reading
31.05.2016 13:51:58 - Error bptm(pid=3096) cannot write data to socket, 10053
31.05.2016 13:51:58 - Info tar32(pid=3852) done. status: 5
31.05.2016 13:52:00 - Error bptm(pid=3096) The following files/folders were not restored:
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2008 Reports/2008 Energia űrlap.xls
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2009 Reports/
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2009 Reports/Tárolótér SB.xls
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2009 Reports/Vasúti mozgások 2009.xls
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2010 Reports/
31.05.2016 13:52:00 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2010 Reports/2010 Termék nyilv..xls
31.05.2016 13:52:01 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2010 Reports/KiszállÃtás 2010.xls
31.05.2016 13:52:01 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2010 Reports/Tárolótér 2010.xls
31.05.2016 13:52:01 - Error bptm(pid=3096) UTF - /E/dfs_data/MABA HUN/Ãltalános/Attila Bacsárdi Y/2007-15 Reports/2007-13 Reports/2010 Reports/Vasúti mozgások 2010.xls
31.05.2016 13:53:00 - Info bptm(pid=2296) waited for empty buffer 0 times, delayed 0 times
31.05.2016 13:53:01 - Info bptm(pid=2296) EXITING with status 24 <----------
31.05.2016 13:53:01 - Info ssolbkp03.office.company.dir(pid=2296) StorageServer=PureDisk:swolbkp03.office.company.dir; Report=PDDO Stats for (swolbkp03.office.company.dir): read: 66981 KB, CR received: 32497 KB, CR received over FC: 0 KB, dedup: 0.0%
31.05.2016 13:53:01 - Info tar32(pid=3852) done. status: 24: socket write failed
31.05.2016 13:53:01 - restored image swolfile01.office.company.dir_1463776795 - (socket write failed(24)); restore time 0:01:23
31.05.2016 13:53:01 - Error bpbrm(pid=5364) client restore EXIT STATUS 24: socket write failed
31.05.2016 13:53:01 - Warning bprd(pid=8108) Restore must be resumed prior to first image expiration on 20.06.2016 22:39:55
31.05.2016 13:53:01 - end Restore; elapsed time: 0:01:23
MS-Windows policy restore error(2808)
Backup and Restore from normal Backups to this client works fine. So there has to be a problem with the optimized backup only. I uploaded the policy configuration as attachements. Policy is running without any problems.
Looking for help.
Thanks
Gerald
05-31-2016 05:55 AM
I forgot to mention that "NFS Server" service on media server is installed. I have already tried with service enabled and disable (like mentioned in the link above). NFS Client service is running on the clients.
05-31-2016 06:41 AM
What does the event viewer on the media server say ?
Please note - there are requirement on what service account the nbfsd service runs under (mentioned in the tech note).
05-31-2016 06:42 AM
This tech note might be helpfull
How to manually test NetBackup Granular Restore Technology (GRT) NFS connections between media server and client
https://www.veritas.com/support/en_US/article.TECH124810
05-31-2016 07:04 AM
Thank you Nicolai for pointing me in the right direction. Restore to another client is working now. Bit now i have another problem :(
As I mentioned above restore is working now and all files are getting created on the new client. So NFS mount should work as expected. But it looks like every deduplicated file is corrupted!!! Restore job was finished successfully. AHAH
It looks like the chunks of the dedupstore folder are not restored. I found following INFO in the job log:
begin reading
31.05.2016 15:40:24 - Info tar32(pid=2792) INF - waiting for mount mutex
31.05.2016 15:40:24 - Info bpbrm(pid=5640) from client smidcad01.office.company.dir: TRV - Starting granular backup processing for (E:\System Volume Information\Dedup). This may take a while...
31.05.2016 15:40:24 - Info bpbrm(pid=5640) from client smidcad01.office.company.dir: TRV - Granular processing failed!
31.05.2016 15:41:06 - Info bptm(pid=7872) waited for empty buffer 0 times, delayed 0 times
31.05.2016 15:41:06 - end reading; read time: 0:00:46
How is it possible to mark a job as successful when a very important part of it is failing!
Any ideas?
05-31-2016 07:39 AM
Did you see this part of the tech note
As such, should always display this same content in Backup Archive and Restore. Whether restoring the whole volume or individual files, under NO CIRCUMSTANCES should "driveletter:\System Volume Information\" ever be restored. Doing so could corrupt the NTFS Deduplication ChunkStore and Database
Youre last post show :
31.05.2016 15:40:24 - Info bpbrm(pid=5640) from client smidcad01.office.company.dir: TRV - Starting granular backup processing for (E:\System Volume Information\Dedup). This may take a while...
new volume - re-try the operation, but do not restore E:\System Volume Information\Dedup
05-31-2016 07:29 PM