>>>>> On Wed, 14 Nov 2007 16:54:57 +0100, Eric =?utf-8?Q?B=C3=B6se-Wolf?= said: > > Hello Martin, > > Martin Simmons <[EMAIL PROTECTED]> writes: > > >> >>> ./bcopy -c ../config/bacula-sd.conf -w /opt/bacula/working/ DVD-RAM \ > >> >>> Dell-TD6100-152 -i > >> >>> 'Archive-0001|Archive-0002|Archive-0003|Archive-0004\ > >> >>> |Archive-0005|Archive-0006|Archive-0007|Archive-0008|Archive-0009' \ > >> >>> -o 'Archive-Tape-1' > > >> Is bcopy broken? > > > Yes, it looks like bcopy creates broken file volumes :-( > > > However, it does read the input volume, so it is a useful test for their > > correctness. > > We still need something to copy volumes around (from tape to file, from > file to tape, without touching the catalog ...) > > > Does bls give the same "short block" error when reading the DVD-RAM volume? > > Yes.
Good, that is as expected because it uses the same code. > > Does restore work for the jobs on DVD-RAM volume Archive-0001? > > This is what I got trying to do a restore from Archive-0001: (Newlines > inserted) > > 14-Nov 16:42 hq-sd JobId 31: Error: block.c:1091 Volume data error at > 1:284417160! Short block of 4983 bytes on device "DVD-RAM" > (/media/cdrom0) discarded. > > 14-Nov 16:42 hq-sd JobId 31: Error: read_record.c:142 block.c:1091 > Volume data error at 1:284417160! Short block of 4983 bytes on device > "DVD-RAM" (/media/cdrom0) discarded. > > 14-Nov 16:42 hq-sd JobId 31: End of file 1 on device "DVD-RAM" > (/media/cdrom0), Volume "Archive-0001" > > 14-Nov 16:42 hq-sd JobId 31: End of Volume at file 1 on device "DVD-RAM" > (/media/cdrom0), Volume "Archive-0001" > > 14-Nov 16:42 hq-sd JobId 31: Warning: acquire.c:206 Read open device > "DVD-RAM" (/media/cdrom0) Volume "Archive-0002" failed: ERR=dev.c:491 > Could not open: /media/cdrom0/Archive-0002, ERR=No such file or > directory > > 14-Nov 16:42 hq-sd JobId 31: Please mount Volume "Archive-0002" or label > a new one for: > Job: Restore.2007-11-14_16.30.04 > Storage: "DVD-RAM" (/media/cdrom0) > Pool: Default > Media type: /media/cdrom0 > *mount DVD-RAM > 3001 OK mount. Device="DVD-RAM" (/media/cdrom0) > > 14-Nov 16:47 hq-sd JobId 31: Ready to read from volume "Archive-0002" on > device "DVD-RAM" (/media/cdrom0). > > 14-Nov 16:47 hq-sd JobId 31: Forward spacing Volume "Archive-0002" to > file:block 0:199. > > Archive-0001 was created by bacula-sd 2.2.5 by doing a backup job in > bconsole writing a file volume to a mounted DVD-RAM with ufs > filesystem. Maybe the out of storage detection is not good for this > combination. Yes, it looks like there is no code to remove a half-written block from the end when it gets an error. > I would expect bacula to detect that it didn't write the > whole block and repeat the block at the beginning of the next volume, > but why does the restore go on on file 0 block 199 of the next > volume? The file and block numbers reported for disk volumes are the upper and lower 32 bits of the byte position, i.e. 1:284417160 is byte 4579384456 in Archive-0001 and 0:199 is byte 199 in Archive-0002. The volume header is probably 199 bytes. > As I proposed to do some archiving here, I don't have the data backup'ed > up to those Archive-000? any more in some other storage. How can I > detect which files have been damaged? Maybe using bls to list the files? The last file listed for Archive-0001 or the first file listed for Archive-0002. You could also find the information from the catalog but it is complicated. > > I suppose that most of the backup is useful, as the restore goes on ... >>>>> On Wed, 14 Nov 2007 17:45:30 +0100, Eric =?utf-8?Q?B=C3=B6se-Wolf?= said: > > Hello Martin, > > [EMAIL PROTECTED] (Eric Böse-Wolf) writes: > > >> Does restore work for the jobs on DVD-RAM volume Archive-0001? > > > > This is what I got trying to do a restore from Archive-0001: (Newlines > > inserted) > > > [Copy and Paste deleted] > > Additionally I get (after the whole restore): > > 14-Nov 17:33 hq-sd JobId 31: End of file 1 on device "DVD-RAM" > (/media/cdrom0), Volume "Archive-0004" > 14-Nov 17:33 hq-sd JobId 31: End of Volume at file 1 on device "DVD-RAM" > (/media/cdrom0), Volume "Archive-0004" > 14-Nov 17:33 hq-sd JobId 31: End of all volumes. > 14-Nov 17:33 hq-dir JobId 31: Bacula hq-dir 2.2.5 (09Oct07): 14-Nov-2007 > 17:33:51 > Build OS: i686-pc-linux-gnu debian lenny/sid > JobId: 31 > Job: Restore.2007-11-14_16.30.04 > Restore Client: hq-fd > Start time: 14-Nov-2007 16:30:39 > End time: 14-Nov-2007 17:33:51 > Files Expected: 3,342 > Files Restored: 3,342 > Bytes Restored: 18,280,046,615 > Rate: 4820.7 KB/s > FD Errors: 0 > FD termination status: OK > SD termination status: OK > Termination: Restore OK > > > So bacula don't think, that the short blocks are a problem. I guess it is a little safer to ignore block errors in restore, because it can count the files and do checksums. Bcopy doesn't look inside the file records. __Martin ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
