Date: Fri, 20 Sep 2002 12:58:56 +0200 (CEST) From: "Hartmann, O." <ohartman@klima.physik.uni-mainz.de> To: freebsd-stable@freebsd.org Cc: freebsd-questions@freebsd.org Subject: Essentiell questions on SA-tape driver: blocksize, EOT-model Message-ID: <20020920123220.W50485-100000@klima.physik.uni-mainz.de>
next in thread | raw e-mail | index | archive | help
Dear Sirs. We use at our department of meteorology FreeBSD 4.6 and 4.7 based systems and several HP DAT tape drives in conjunction with afbackup 3.3.7pl2 for performing our backup tasks. I need to mention that afbackup worked well since the changes of FBSD 4.7 in the SA driver (early August/2002). With FreeBSD 4.7-PRERELEASE and FreeBSD 4.7-RC afbackup 3.3.7pl2 and the ports-collection outdated afbackup 3.3.5 doesn't work anymore. But this is simply the background for my outstanding questions. I try to figure out what to do to get best results and stability with tape drives and fiddling around with several parameters results in several strange behaviours of the tape drives. We use two single tape drives from HP, one DDS-4, recognized by FBSD 4.7-RC as: sa0 at sym0 bus 0 target 4 lun 0 sa0: <HP C5683A C005> Removable Sequential Access SCSI-2 device sa0: 40.000MB/s transfers (20.000MHz, offset 31, 16bit) and an older DDS-2 tape drive, recognized as: sa0 at ahc1 bus 0 target 5 lun 0 sa0: <HP C1533A 9503> Removable Sequential Access SCSI-2 device sa0: 10.000MB/s transfers (10.000MHz, offset 8) The last one, DDS-2 type tape drive, has been run all the time with varibale-block size without any kind of errors while doing backups localy and via network. The newer DAT 40 tape drive (DDS-4) reports errors when switching to varibale blocksize with 'mt blocksize 0'. this is the error message reported by the kernel of the <HP C5683A C005> drive: (sa0:sym0:0:4:0): WRITE FILEMARKS. CDB: 10 0 0 0 1 0 (sa0:sym0:0:4:0): MEDIUM ERROR asc:3b,0 (sa0:sym0:0:4:0): Sequential positioning error (sa0:sym0:0:4:0): failed to write terminating filemark(s) (sa0:sym0:0:4:0): tape is now frozen- use an OFFLINE, REWIND or MTEOM command to clear this state. (sa0:sym0:0:4:0): tape is now frozen- use an OFFLINE, REWIND or MTEOM command to clear this state. Both tape drives use EOT model 1 filemark (mt seteotmodel 1). What does this error mena and why does it occur? Why not on the older tape drive? Next problematic device is our third backup system, also a DAT drive from HP, but mounted to an autoloader (HP SureStore 40x6i): sa0 at sym1 bus 0 target 5 lun 0 sa0: <HP C5713A H910> Removable Sequential Access SCSI-2 device sa0: 40.000MB/s transfers (20.000MHz, offset 31, 16bit) ch0 at sym1 bus 0 target 5 lun 1 ch0: <HP C5713A H910> Removable Changer SCSI-2 device ch0: 40.000MB/s transfers (20.000MHz, offset 31, 16bit) ch0: 6 slots, 1 drive, 0 pickers, 0 portals This autoloader/tape unit backups our RAID system and utilize afbackup 3.3.7pl2, or better: it utilized afbackup, it is now broken and unuseable with FreeBSD 4.7! Using viarible blocksize, EOT filemark 1 and compression on with this tape drive results in this this error: (sa0:sym1:0:5:0): unable to rewind after test read (ch0:sym1:0:5:1): READ ELEMENT STATUS. CDB: b8 20 0 2 0 1 0 0 4 0 0 0 (ch0:sym1:0:5:1): NOT READY asc:3e,0 (ch0:sym1:0:5:1): Logical unit has not self-configured yet ch: warning: could not map element source address 0d to a valid element type (ch0:sym1:0:5:1): chgetparams: error getting device capabilities page (ch0:sym1:0:5:1): chgetparams: error getting element address page (ch0:sym1:0:5:1): chgetparams: error getting element address page sym1:5:control msgout: 80 6. (ch0:sym1:0:5:1): chgetparams: error getting element address page sym1:5:control msgout: 80 6. It seems also that the changer (ch0) has problems to change tapes in afbackup, but I can not figure our why and I suspect the chaos in the new sa-driver (where is the HEADS UP?), but I confess that this is out of my focus of competence. The big mystery is: why does varibale blocksize work on the older models of HP and not on the newer ones (DAT tape drives)? At the end I do not want to ask for a solution for the afbackup-catastrophy, but I want to know more about how to use variable blocksize - or to avoid them, if they are not suitable for FreeBSD. What is about the two distinguish EOT models? Which one is the one we should prefer, 2 filemarks or only one? Assuming having modern tape drives, like DAT 40 or SDLT/LTO, what kind of model is prefereable? And what kind of blcoksize? As often as I read something about blocksizes on tape drive units I read '10240', ten times more than 1024. But when I took a look on a new tapedrive unit or after a reset, I see a blocksize of 1024 (fixed). Is that correct? Should the blocksize be variable or fixed to 1024? Please refer this to FreeBSD and its backup utilities and with modern tape drive units in mind. At this moment I feel sick and really confused, especially disappointed by the weirdness of the sa driver failure with the very important 'afbackup' utility. Hope someone dares to answer, I know the problems are more specific than they seem to be. Thanks in advance, Oliver -- MfG O. Hartmann ohartman@klima.physik.uni-mainz.de ------------------------------------------------------------------ IT-Administration des Institutes fuer Physik der Atmosphaere (IPA) ------------------------------------------------------------------ Johannes Gutenberg Universitaet Mainz Becherweg 21 55099 Mainz Tel: +496131/3924662 (Maschinenraum) Tel: +496131/3924144 (Buero) FAX: +496131/3923532 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020920123220.W50485-100000>