It looks like it is trying to read off the end of the tape and getting an I/O error. Possibly setting TWOEOF = yes may solve the problem.
On Tuesday 11 April 2006 10:55, Gabriele Bulfon wrote: > Hello, I upgraded and read your latests manual. > So now, I'm playing with the device flags to check for some configuration > that may work. Whil doing btape/test , I could see this output (I never saw > it because it was scrolling very fast) : This is an *essential* feature ... > btape: btape.c:795 Wrote 1000 blocks of 64412 bytes. > btape: btape.c:469 Wrote 1 EOF to "DAT72" (/dev/rmt/1) > btape: btape.c:811 Wrote 1000 blocks of 64412 bytes. > btape: btape.c:469 Wrote 1 EOF to "DAT72" (/dev/rmt/1) > btape: btape.c:820 Rewind OK. > 1000 blocks re-read correctly. > Got EOF on tape. > 11-Apr 10:44 btape: btape Error: block.c:945 Read error at file:blk 1:0 on > device "DAT72" (/dev/rmt/1). ERR=I/O error. and then it scrolls all the way > down with the "Got EOF on tape". > Have any idea? > Gabriele Bulfon - Sonicle S.r.l. > Tel +39 028246016 Int. 30 - Fax +39 028243880 > Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY > http://www.sonicle.com > --------------------------------------------------------------------------- >------- Da: Kern Sibbald <[EMAIL PROTECTED]> > A: Gabriele Bulfon <[EMAIL PROTECTED]> > Cc: bacula-users <bacula-users@lists.sourceforge.net> > Data: 10 aprile 2006 18.56.25 CEST > Oggetto: Re: [Bacula-users] restoring > > On Monday 10 April 2006 18:33, Gabriele Bulfon wrote: > > ...I'm worrying about it because the list of "Got EOF on tape." never > > ends, until I kill the btape/test; should I wait longer? > > Sorry, my error. I did not read what you wrote carefully enough. If you are > getting that during a restore, there is no problem and my previous comments > apply. > Since you are getting that during the btape "test" command as you are, then > yes, it indicates that your Device specification in the bacula-sd.conf file > does not match your OS and tape drive. You must correct this condition > before Bacula will work correctly. The btape tests must run to completion > correctly there are some "got EOF" messages but only 5-10. > Perhaps if you send your bacula-sd.conf file to the list, someone can help > you debug the problems. It would also be a good idea to send a copy of the > output of btape "test" (of course, truncating any big number of "got EOF" > messages). > By the way, Bacula 1.36.2 is quite old and no longer "officially" > supported ... > > > And I'm worrying because > > in the pdf document, you state that if you get this output, it's probably > > because of the variable / fixed-size block setting, and that I should fix > > it. I'm worrying because I cannot run a single "./bls -c > > ../etc/bacula-sd.conf -j -v -V THURSDAY2 QUANTUM" on any tape (old or > > recent), because it will fail to show me but the first job. And finally > > I'm worrying because I cannot run a single "./bextract ....." on jobs > > that ar not the first one. It seems like command line tools can work only > > on a short amount of data that stays at the beginning of the tape. ...I'm > > really worried...please help.... > > Gabriele Bulfon - Sonicle S.r.l. > > Tel +39 028246016 Int. 30 - Fax +39 028243880 > > Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY > > http://www.sonicle.com > > ------------------------------------------------------------------------- > >-- ------- Da: Kern Sibbald <[EMAIL PROTECTED]> > > A: Gabriele Bulfon <[EMAIL PROTECTED]> > > Cc: bacula-users <bacula-users@lists.sourceforge.net> > > Data: 10 aprile 2006 18.24.02 CEST > > Oggetto: Re: [Bacula-users] restoring > > I don't know why you are worrying about the "got EOF" messages. They are > > quite normal. They are not error messages. Bacula is *very* careful to > > put the word "error" in error messages or "warning" in possible error > > messages. If neither of those words are in the message, it can be > > considered an "informational" message and not normally anything to worry > > about. > > > > On Monday 10 April 2006 18:14, Gabriele Bulfon wrote: > > > I tried the btape tests on clean systems: > > > Sun v20z, AMD64, Solaris 10, self compiled bacula 1.36.2, Quantum > > > Ultrium LTO2. SunFire280R, Sparcv9, Solaris 10, self compiled bacula > > > 1.36.2, Quantum Ultrium LTO2. I always get the long list of "got > > > EOF"....but I'm sure that the system defaults to variable block size. I > > > even tried substituting the Quantum device with a certified Sun-Dat72 : > > > same results. I believe that until I can't get rid of these "got EOF" > > > during tests, I will have my problem (able to restore via bconsole, not > > > able to restore through command line). Do you think there may be > > > something wrong in the way I compiled binaries on Solaris? Is there any > > > issue about the 64 bit platform? The binaries are all 32bit ones: may > > > be this a problem? Is there any specific issue pertaining to the > > > Solaris 10 platform? Thanks for any help. > > > Gabriele Bulfon - Sonicle S.r.l. > > > Tel +39 028246016 Int. 30 - Fax +39 028243880 > > > Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY > > > http://www.sonicle.com > > > ----------------------------------------------------------------------- > > >-- -- ------- Da: Kern Sibbald <[EMAIL PROTECTED]> > > > A: bacula-users@lists.sourceforge.net > > > Cc: Gabriele Bulfon <[EMAIL PROTECTED]> > > > Data: 10 aprile 2006 11.50.50 CEST > > > Oggetto: Re: [Bacula-users] restoring > > > > > > On Monday 10 April 2006 09:25, Kern Sibbald wrote: > > > > On Sunday 09 April 2006 23:21, Gabriele Bulfon wrote: > > > > > Thanx for your suggestions. > > > > > Don't you think that it's strange that bacula can restore from > > > > > bconsole, and not from manual commands? > > > > > > > > All but the highest level code that reads a Volume is identical > > > > between all Bacula components, so there is unlikely to be any > > > > difference between jobs run from bconsole and using the standalone > > > > tools. > > > > > > > > > Is it a possible situation with a badly conficured device? > > > > > > > > Since the code is the same, my answer is: no. > > > > > > > > Your problem is most likely a bad tape or a dirty drive as I > > > > mentioned last time. An I/O error as you got means a hardware problem > > > > not a software problem. The manual lists other possibilities. > > > > > > Well, yes, I should probably be more explicit as I don't always > > > distinguish between OS problems and hardware problems. > > > I should have said that this is very unlikely to be an application > > > program (Bacula) error. If there is an I/O error, either the kernel > > > driver is buggy or there is a problem with the hardware (which could > > > also be the firmware in the tape drive). > > > > > > > > Actually I never happened to encounter this problem, because I > > > > > could always restore anything that was still in the database. > > > > > Gabriele Bulfon - Sonicle S.r.l. > > > > > Tel +39 028246016 Int. 30 - Fax +39 028243880 > > > > > Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY > > > > > http://www.sonicle.com > > > > > ------------------------------------------------------------------- > > > > >-- -- -- -- ------- Da: Kern Sibbald <[EMAIL PROTECTED]> > > > > > A: bacula-users@lists.sourceforge.net > > > > > Cc: Gabriele Bulfon <[EMAIL PROTECTED]> > > > > > Data: 9 aprile 2006 10.25.44 CEST > > > > > Oggetto: Re: [Bacula-users] restoring > > > > > > > > > > On Saturday 08 April 2006 13:02, Gabriele Bulfon wrote: > > > > > > Hello, > > > > > > yesterday I was trying to restore an old file from an old backup > > > > > > tape that is no more in the database. So I tried using first > > > > > > btape and bls to check the content. > > > > > > With btape I could read the original label. > > > > > > With bls I was trying to get the list of jobs numbers present in > > > > > > the tape (this tape should contain 5 jobs of 5 different clients > > > > > > on the same day). But this is what I got from the system: > > > > > > -bash-3.00# ./bls -c ../etc/bacula-sd.conf -j -v -V FRIDAY > > > > > > QUANTUM bls: butil.c:258 Using device: "QUANTUM" for reading. > > > > > > 08-Apr 12:47 bls: Ready to read from volume "FRIDAY" on device > > > > > > /dev/rmt/0. Volume Label: > > > > > > Id : Bacula 1.0 immortal > > > > > > VerNo : 11 > > > > > > VolName : FRIDAY > > > > > > PrevVolName : > > > > > > VolFile : 0 > > > > > > LabelType : VOL_LABEL > > > > > > LabelSize : 162 > > > > > > PoolName : FridayPool > > > > > > MediaType : LTO > > > > > > PoolType : Backup > > > > > > HostName : iserver > > > > > > Date label written: 20-Jul-2005 10:39 > > > > > > Begin Job Session Record: > > > > > > JobId : 16 > > > > > > VerNum : 11 > > > > > > PoolName : FridayPool > > > > > > PoolType : Backup > > > > > > JobName : Enterprise Backup > > > > > > ClientName : iserver-fd > > > > > > Job (unique name) : Enterprise_Backup.2006-04-07_23.00.00 > > > > > > FileSet : Full Set > > > > > > JobType : B > > > > > > JobLevel : F > > > > > > Date written : 07-Apr-2006 23:00 > > > > > > 08-Apr 12:48 bls: Got EOF at file 1 on device /dev/rmt/0, Volume > > > > > > "FRIDAY" 08-Apr 12:48 bls: bls Error: block.c:782 Read error at > > > > > > file:blk 1:0 on device /dev/rmt/0. ERR=I/O error. 08-Apr 12:48 > > > > > > bls: End of Volume at file 1 on device /dev/rmt/0, Volume > > > > > > "FRIDAY" 08-Apr 12:48 bls: End of all volumes. End of physical > > > > > > tape. 08-Apr 12:48 bls: ERROR in dev.c:1288 dev.c:1287 This > > > > > > device does not support Unknown. -bash-3.00# > > > > > > Then, to check the system, I placed other recent tapes in the LTO > > > > > > drive, and issued the same commands. Always the same error.... > > > > > > I even tried to force a "bextract", knowning what was inside the > > > > > > tape, with the same errors. So I did the same on other systems I > > > > > > have at customers.....always the same problem!.... I was worried. > > > > > > At last, I tried restoring a file from a recent tape that is in > > > > > > the database, so I could use bconsole/restore and the file was > > > > > > correctly restored!........but the same tape gives errors with > > > > > > manual commands. This problem doesn't let me restore tapes that > > > > > > are no more in the database. I did not try to run a "bscan" yet, > > > > > > on a separate and clean machine, but I assume it will be > > > > > > producing the same errors while trying to recreate the db. Any > > > > > > idea about this? > > > > > > > > > > It looks like Bacula got an error trying to read the data on your > > > > > tape. I suspect that your drive is dirty or your media is bad. You > > > > > might try cleaning your tape drive. > > > > > You should also run the btape "test" command if you haven't already > > > > > to ensure that your Device directives are compatibile with your > > > > > drive. Also, it appears that Bacula's code is attempting to issue > > > > > an unsupported ioctl() *after* the tape go an error. > > > > > > > > > > > Machines are Sun v20z amd64, Solaris 10, Bacula 1.36.2 compiled > > > > > > with gcc in 32 bit mode. Drives are Certance Ultrium LTO2. > > > > > > Gabriele. > > > > > > Gabriele Bulfon - Sonicle S.r.l. > > > > > > Tel +39 028246016 Int. 30 - Fax +39 028243880 > > > > > > Via Felice Cavallotti 16 - 20089, Rozzano - Milano - ITALY > > > > > > http://www.sonicle.com > > > > > > > > > > -- > > > > > Best regards, > > > > > Kern > > > > > ("> > > > > > /\ > > > > > V_V > > > > > > -- > > > Best regards, > > > Kern > > > ("> > > > /\ > > > V_V > > > > -- > > Best regards, > > Kern > > ("> > > /\ > > V_V > > -- > Best regards, > Kern > ("> > /\ > V_V -- Best regards, Kern ("> /\ V_V ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
