From: Kern S. <ke...@si...> - 2004-02-21 23:09:27
|
Hello Jason, Thanks sounds like enough info to keep me busy for the time being. Thanks, I'll work on fixing it. Best regards, Kern On Sat, 2004-02-21 at 22:28, JML wrote: > Kern, > > mt status and tapeinfo output are attached. Default blocksize was > originally 512 bytes. I ran the following test on my machine: > > - set defblksize to 512 > - started bacula, labelled tape, ran job > - job completes with the following in log > SD termination status: OK > Termination: Backup OK > - restore attempts with blocksize set to 0 or 512 result in the familiar > "Volume data error! Wanted block-id BB02, got . Buffer discarded." > > This is redhat 9 with a 2.4.20 kernel. I can also send you the large debug > logs for the director and storage daemon off-list if you'ld like em. > > thanks, > > Jason > > > On Sat, 21 Feb 2004, Kern Sibbald wrote: > > > On Fri, 2004-02-20 at 20:05, JML wrote: > > > sorry, my mistake- a default block size was still set. It might be > > > clearer, however, if that tip moved from section 17.7 to 17.6 in the docs. > > > > Good idea. I copied it so it appears in both places. > > > > > > > > By the way, should I be concerned that bacula wrote to tape without > > > reporting errors in spite of the incorrect block size? > > > > Yes, Bacula should report all errors. If it is not, and I don't see how > > that can happen, there is a serious "bug" in Bacula. Please give me the > > details so that I can reproduce it here (I imagine it is your default > > block size, but I'd like to have the details anyway). > > > > Best regards, Kern > > > > > In other words, any > > > chance that it would write a tape that was later unreadable, even if the > > > setting were corrected after the fact? I'm thinking of the sysadmin who > > > forgets to script the mt commands for startup, or the newbie who never > > > tests restores until they're needed. :) > > > > > > regards, > > > > > > Jason > > > > > > > > > On Thu, 19 Feb 2004, Kern Sibbald wrote: > > > > > > > Hello, > > > > > > > > Before running the btape "test", run the mt command recommended in the > > > > manual that sets your SCSI driver parameters. > > > > > > > > Then please send me the full output from the latest btape 1.33.3-16Feb04 > > > > "test" command, and I'll look at it. Also please send me your > > > > bacula-sd.conf (I know you sent it before, but it is buried somewhere > > > > long ago ...). > > > > > > > > Best regards, Kern > > > > > > > > On Wed, 2004-02-18 at 22:19, JML wrote: > > > > > Hi Kern, > > > > > > > > > > a quick update since my previous post below- the new btape in 1.33.3 > > > > > reports the same problem with our DDS-4. Any thoughts on what I should try > > > > > next? > > > > > > > > > > thanks, > > > > > > > > > > Jason > > > > > > > > > > > > > > > On Fri, 13 Feb 2004, JML wrote: > > > > > > > > > > > Yes, the problem seems odd for so mainstream a drive as this. I followed > > > > > > your suggestions on tape cleaning and alternate tapes, but they check out > > > > > > okay. I can even write and read tar archives from the device with no > > > > > > problem. > > > > > > > > > > > > In case they offer any clues, I've attached the results from tapeinfo, mt > > > > > > status, and the CVS version btape, which does catch an error now BTW. It > > > > > > also seems odd that tapeinfo reports "Medium Type: Not Loaded" when it is > > > > > > loaded. > > > > > > > > > > > > thanks, > > > > > > > > > > > > Jason > > > > > > > > > > > > > > > > > > On Fri, 13 Feb 2004, Kern Sibbald wrote: > > > > > > > > > > > > > Hello, > > > > > > > > > > > > > > Fast block rejection is still turned off in 1.33. In fact, it should > > > > > > > cause problems only if using multiple concurrent jobs, so I doubt it is > > > > > > > the problem here. With a Sony DDS-4 you should be having no problems. > > > > > > > I'd recommend that you remove the: > > > > > > > > > > > > > > Fast Forward Space File = no > > > > > > > Backward Space Record = No > > > > > > > > > > > > > > as they just cripple your excellent drive. It may be that you have a > > > > > > > bad tape, or a dirty drive, so I would recommend you run a cleaning tape > > > > > > > at least once, then as I suggested in a previous email, download the > > > > > > > latest CVS code and rebuild btape. It now has a tape block read back > > > > > > > test as well as a positioning test, so maybe something will show up. > > > > > > > > > > > > > > Best regards, Kern > > > > > > > > > > > > > > > > > > > > > On Fri, 2004-02-13 at 05:35, JML wrote: > > > > > > > > Hello again, > > > > > > > > > > > > > > > > I have a separate issue with restoring from a DDS-4 changer (specifically > > > > > > > > a Sony TSL-S11000). System is RedHat 9. Changer ops and backup work fine, > > > > > > > > then I get the following during a restore job: > > > > > > > > > > > > > > > > nail-sd: Ready to read from volume "foo-06" on device /dev/nst0. > > > > > > > > nail-sd: Forward spacing to file:block 0:1. > > > > > > > > nail-sd: RestoreFiles.2004-02-12_16.36.23 Error: block.c:248 Expected > > > > > > > > block-id BB01 or BB02, got . Buffer discarded. > > > > > > > > nail-sd: RestoreFiles.2004-02-12_16.36.23 Error: block.c:248 Expected > > > > > > > > block-id BB01 or BB02, got . Buffer discarded. > > > > > > > > nail-sd: RestoreFiles.2004-02-12_16.36.23 Error: Read error on Record > > > > > > > > Header /dev/nst0: Success > > > > > > > > nail-fd: RestoreFiles.2004-02-12_16.36.23 Error: bnet.c:374 Write error > > > > > > > > sending to Storage daemon:nail:9103: ERR=Broken pipe > > > > > > > > nail-dir: RestoreFiles.2004-02-12_16.36.23 Error: Bacula 1.33 (30Jan04): > > > > > > > > 12-Feb-2004 16:38 > > > > > > > > JobId: 35 > > > > > > > > Job: RestoreFiles.2004-02-12_16.36.23 > > > > > > > > Client: nail-fd > > > > > > > > Start time: 12-Feb-2004 16:36 > > > > > > > > End time: 12-Feb-2004 16:38 > > > > > > > > Files Restored: 0 > > > > > > > > Bytes Restored: 0 > > > > > > > > Rate: 0.0 KB/s > > > > > > > > FD Errors: 1 > > > > > > > > FD termination status: Error > > > > > > > > SD termination status: Error > > > > > > > > Termination: *** Restore Error *** > > > > > > > > > > > > > > > > Seems to make no difference whether I "mark *" or only restore a subset of > > > > > > > > the job. I also followed the instructions pertaining to mt stoptions. > > > > > > > > > > > > > > > > btape's test gives a clean bill of health on this drive, though after > > > > > > > > advising that I set the last two options in my Device declaration below: > > > > > > > > > > > > > > > > Device { > > > > > > > > Name = DDS-4 # > > > > > > > > Media Type = DDS-4 > > > > > > > > Archive Device = /dev/nst0 > > > > > > > > Autochanger = Yes > > > > > > > > Changer Command = "/usr/local/bacula/etc/mtx-changer %c %o %S %a" > > > > > > > > Changer Device = /dev/sg1 > > > > > > > > LabelMedia = No > > > > > > > > AutomaticMount = yes; # when device opened, read it > > > > > > > > AlwaysOpen = yes; > > > > > > > > Fast Forward Space File = no > > > > > > > > Backward Space Record = No > > > > > > > > } > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > All of this sounds similar to a thread from last week when the advice was > > > > > > > > to upgrade to 1.32f-4, which had fast block rejection disabled. Is this > > > > > > > > disabled already in the 1.33 beta, or is there an easy way to do so? > > > > > > > > > > > > > > > > best, > > > > > > > > > > > > > > > > Jason > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ------------------------------------------------------- > > > > > > > > SF.Net is sponsored by: Speed Start Your Linux Apps Now. > > > > > > > > Build and deploy apps & Web services for Linux with > > > > > > > > a free DVD software kit from IBM. Click Now! > > > > > > > > http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click > > > > > > > > _______________________________________________ > > > > > > > > Bacula-users mailing list > > > > > > > > Bac...@li... > > > > > > > > https://lists.sourceforge.net/lists/listinfo/bacula-users > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > |