Restore from client fails with error code 13: file read failed
I'm thinking there is a simple answer here, but I don't know it...
I'm testing restores from Data Domain replication to a disaster recovery server. These are data base files, so I had the DBA (logged on as oracle, since the source files are owned by oracle and the target directory is also owned by oracle), run the bprestore command from the target host:
bprestore -K -s 02/04/2011 18:00:00 -e 02/07/2011 06:00:00 -C source_server.foo.com -D target_server.foo.com -L /tmp/restore.log -R /tmp/rename -copy 2 /u04/dbfiles
This failed with the following:
Feb 10, 2011 12:50:22 PM - Error bpbrm (pid=30429) socket read failed: errno = 62 - Timer expired
Feb 10, 2011 12:50:25 PM - Error bptm (pid=30430) media manager terminated by parent process
Feb 10, 2011 1:05:26 PM - Error bptm (pid=30430) wait for child pid 30431 timeout
Feb 10, 2011 1:05:27 PM - Error bptm (pid=30431) cannot write data to socket, Interrupted system call
Feb 10, 2011 1:05:27 PM - Error bptm (pid=30431) The following files/folders were not restored:
Files listed here...
Feb 10, 2011 1:05:29 PM - Error bptm (pid=30431) more than 10 files were not restored, remaining ones are shown in the progress log.
Feb 10, 2011 1:05:29 PM - Error bpbrm (pid=30429) client restore EXIT STATUS 13: file read failed
Feb 10, 2011 1:05:30 PM - restored from image source_server.foo.com_1296866568; restore time: 0:30:16
Feb 10, 2011 1:05:30 PM - Warning bprd (pid=29076) Restore must be resumed prior to first image expiration on Mon 07 Mar 2011 07:42:48 PM EST
Feb 10, 2011 1:05:30 PM - end Restore; elapsed time 0:30:17
the restore failed to recover the requested files (5)
I did create the files in /usr/openv/netbackup/db/altnames (I created a file named source_server.foo.com with "target_server.foo.com" in it, and the other way around since I forget which way it works).
Anyway, I ran the same command as root on the master and its restoring. I'm not sure if its because it was run as a non-root user, or because it was run from the client (I plan on running those, but this restore is taking a long time).
This is a Netbackup 6.5.6 environment, on RHEL servers and clients.
thanks
To allow a single client to perform redirected restores
The NetBackup administrator can permit a single client to restore the backups
that belong to other clients. Create a peername file on the NetBackup master
server where the policy that backed up the other client(s) resides.
Note: The information in this topic applies to restores made by using the command
line, not the Backup, Archive, and Restore client interface.
Create an altnames directory in the following location, then place the empty file
inside of the directory:
/usr/openv/netbackup/db/altnames/peername
Where peername is the client to possess restore privileges.
In this case, the requesting client (peername) can access the files that are backed
up by another client. The NetBackup client name setting on peername must match
the name of the other client.
To allow redirected restores of a client’s files
The NetBackup administrator can permit a single client to restore the backups
that belong to another client. Create a peername file on the NetBackup master
server of the requesting client as described here.
Note: The information within this topic applies to restores made using the
command line, not the Backup, Archive, and Restore client interface.
Create an altnames directory in the following location, then place the peername
file inside of the directory:
/usr/openv/netbackup/db/altnames/peername
Where peername is the client to possess restore privileges. Add to the peername
file the names of the client(s) whose files the requesting client wants to restore.
The requesting client can restore the files that were backed up by another client
if:
■ The names of the other clients appear in the peername file, and
■ The NetBackup client name of the requesting client is changed to match the
name of the client whose files the requesting client wants to restore.
