From: Kern S. <ke...@si...> - 2004-02-13 10:42:23
|
A points: 1. be sure you have also done the: mt -f /dev/nosst0 defblksize 32768 2. The block size was highly recommended by the author of the device driver, and is known to work with it. 3. Make sure you have the OnStream kernel osst driver version 0.9.14 or above. If not I *guarantee* that Bacula will not work with the drive. 4. When doing the fill test, you must select the simple or single tape option rather than the multiple tape option, which is still not working correctly. This has only been noted in the online manual for about a week or so, so you may not have noticed it. Best regards, Kern On Fri, 2004-02-13 at 10:52, H Jansen wrote: > Since bacula (version 1.33) now also supports OnStream tape devices I > thought I'll give my OnStream ADR2.60 (IDE-SCSI) tape unit another try. > Unfortunately, the btape test did not complete successfully: Rereading > the tape (after writing to a multivolume with btape's fill commeand) > showed that all written blocks were invariably bad (repeating the error > message "... bad block at record ..." for every block it was attempting > to read). > > The tape unit is accessed through /dev/nst0 and I use "Maximum Block > Size = 32768" (according to Bacula doc) although I do remember to have > read somewhere that the ADR2.60 works with tape formats of block size > 65536 (OnStream doc) --- with this latter bigger block size the btape > excercise already fails when writing. > > I feel I'm almost there but missing the last piece of the puzzle. Has > anyone some constructive ideas (it's a bit of waste to leave this $500 > device unused....) > > Thanks, > > --Henk > > -- > Henk Jansen <h.j...@fe...> > TNO Physics and Electronics Laboratory > |