After taking a look at the manual again, I don't think there is any need to change anything. This problem is *very* clearly explained in the section that describes "Archive device = ..." for the Storage daemon conf file. I suppose I could possibly add a warning in the installation section, but the manual is already quite large :-)
On Tuesday 11 April 2006 15:57, Kern Sibbald wrote: > On Tuesday 11 April 2006 13:14, Gabriele Bulfon wrote: > > Maybe this is a silly discovery, but I did not find any writing about it > > anywhere, so I'm sending you the solution. Maybe would be nice to have a > > mention on the pdf documentation. Running Bacula on Solaris : you HAVE to > > use the "b" device on tapes! Read the Solaris 10 mtio man page extract: > > Read Operation > > The read(2) function reads the next record on the tape. The > > record size is passed back as the number of bytes read, pro- > > vided it is not greater than the number requested. When a > > tape mark or end of data is read, a zero byte count is > > returned; all successive reads after the zero read will > > return an error and errno will be set to EIO. To move to the > > next file, an MTFSF ioctl can be issued before or after the > > read causing the error. This error handling behavior is dif- > > ferent from the older BSD behavior, where another read will > > fetch the first record of the next tape file. If the BSD > > behavior is required, device names containing the letter b > > (for BSD behavior) in the final component should be used. If > > persistent error handling was enabled with either the BSD or > > SVR4 tape device behavior, all operations after this read > > error will return EIO errors until the MTIOCLRERR ioctl is > > issued. An MTFSF ioctl can then he issued. > > Yes, this *is* in the manual (in the Tape Testing chapter), but perhaps I > could stress the /dev/rmt/0b for Solaris systems a bit more. > > Anyway, I am happy you found the solution ... > > > Now, using /dev/rmt/0b will run the tests perfectly. > > If you use /dev/rmt/0 : > > - Test will not work > > - Bacula will correctly write > > - bconsole will correctly read, because it knows exactly the file number > > to reach, and ask the device to skip there before reading - command line > > utilities won't work, beacuse they will fail discovery of content after > > reaching the end of file=1 Previous tapes that I have written using > > /dev/rmt/0 can be read through /dev/rmt/0b, but may mislead you. Trying > > an "bls -j -v -V ..." on the old tape will show the first start of job, > > and then come out with a lot of: 11-Apr 13:03 bls: Got EOF at file 1 on > > device /dev/rmt/0b, Volume "THURSDAY2" 11-Apr 13:03 bls: Got EOF at file > > 2 on device /dev/rmt/0b, Volume "THURSDAY2" .... > > .... > > You just have to wait. The rest of data has been written into later > > files. Then you can create your bootstrap file and suggest bextract where > > to position. Hope this helps, > > Yes, I will put this information in the manual. > > > 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 ------------------------------------------------------- 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
