On Wednesday 28 March 2007 12:57, Brian Debelius wrote: > A workaround, which I do because i have not figured out how to correct > this, is to just make sure your backups do not fill the tape.
First, these problems are not normally fatal, but I wouldn't run my tape drive like that. Excluding tape/drive errors which produce similar problems, the most frequent causes of this are: 1. Old tape drives that don't have a logical end of tape marker (this is not likely to be the case except in drives manufactured more than 10 years ago). 2. Poor tape driver implementation. This is the most common cause. Tape driver writers often don't really know what they are doing. The driver must read the logical end of tape (this occurs before the physical end of the tape) and must signal it to the program. In the Unix world most off brand OSes do not signal the end of the tape. I won't mention their names other than to say that Solaris, Linux, and I think FreeBSD, do so correctly. Correctly signaling a logical end of tape means sending back a error status to the write with an errno == ENOSPC then allowing the program to continue writing until the hardware end of tape. Alternatively, the OS can signal with an error status and errno == EIO and continue to allow the program to write until the physical end of the tape. This is *much* less correct since Bacula does not know if it is a real error condition or an end of tape. However, aside from the worrysome error messages, the results are the same -- i.e. OK. The worst case is the OS signals an error status (errno = EIO) and does not allow the program to write to the physical end of the tape. This is a bad situation, but Bacula should handle it. At least the data can be recovered -- in most cases. The problem is when the WEOF fails, one cannot be sure whether or not the tape is properly terminated, and there is a chance that some restore operation may read off the end of the tape. 3. Since I think we are talking about Win32, it is possible that the low level system calls (via mscrt.dll or whatever the name of the Microsoft compatibility layer we use) is not well written to return reasonable/correct statuses. This would be equivalent to poorly written Unix OS drivers I mentioned above. > > brian- > > [EMAIL PROTECTED] wrote: > > Can any one help us fix this problem ? (details below) > > > > Thank you > > > > Nawfel BERAICH > > Casablanca 9h30 AM > > > > ----- Message d'origine ---- > > De : Brian Debelius <[EMAIL PROTECTED]> > > À : [EMAIL PROTECTED] > > Cc : bacula-users@lists.sourceforge.net > > Envoyé le : Mardi, 27 Mars 2007, 14h15mn 51s > > Objet : Re : [Bacula-users] Re : Windows::Permission denied > > > > You are not to the point to start backing up files yet. This needs to > > be fixed first. > > > > > > [EMAIL PROTECTED] wrote: > > > Is it that blocking ? I think that with this kind of error bacula > > will start the Job and fail while trying to finish it (because it's an > > I/O error) but my bacula dont start writing the jobs ! it always say > > that there is an error and stop without spending any time in writing > > files... :-( ! > > > > > > Thank you so much Brian for you precious help ! > > > > > > ----- Message d'origine ---- > > > De : Brian Debelius <[EMAIL PROTECTED]> > > > À : [EMAIL PROTECTED] > > > Cc : bacula-users@lists.sourceforge.net > > > Envoyé le : Mardi, 27 Mars 2007, 13h25mn 14s > > > Objet : Re: Re : Re : Re : Re : [Bacula-users] Re : > > Windows::Permission denied > > > > > > Ok, you are now beyond my experience. Can someone help with fixing his > > > EOF error? > > > > > > 26-Mar 18:35 btape: End of Volume "TestVolume1" at 148:13128 on device > > > "HPStorag > > > eTape0" (Tape0). Write of 64512 bytes got -1. > > > 26-Mar 18:35 btape: btape Error: Error writing final EOF to tape. This > > > Volume ma > > > y not be readable. > > > ../../stored/dev.c:1688 ioctl MTWEOF error on "HPStorageTape0" (Tape0). > > > ERR=Inpu > > > t/output error. > > > > > > brian- > > > > > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys-and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
