Date: Sun, 21 Jan 2001 11:20:17 -0800 (PST) From: Matthew Jacob <mjacob@feral.com> To: Odhiambo Washington <wash@iconnect.co.ke> Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: CANNOT ACCESS T20 TAPE Message-ID: <Pine.BSF.4.21.0101211115450.1784-100000@beppo.feral.com> In-Reply-To: <20010121192427.K589@poeza.iconnect.co.ke>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 21 Jan 2001, Odhiambo Washington wrote:
> * Matthew Jacob <mjacob@feral.com> [20010121 02:26]: writing on the subject 'Re: CANNOT ACCESS T20 TAPE'
> Matthew>
> Matthew> I have a T20 with 3.00 f/w which works just peacy in -current.
> Matthew>
> Matthew> I know you said things worked fine under 3.4 and now don't work when you
> Matthew> upgraded to 4.2- but I have to say that this seems very unlikely somehow. It
> Matthew> seems like your tape drive has died. Please retry 3.4 and see what's up with
> Matthew> that and let me know.
>
> Is it easy for a tape drive to die? I still hope it is some other issue
> because I did a full backup 2 days before the upgrade and we've not had
> any power failures or something that I can suspect. Okay the only problem
> here now is that it is a SCSI tape and I do not have another readily
> available system with a SCSI adapter to attach it to for a test.
It wouldn't be the adapter.
> Well, whatever it is lemme think hard about it. I'll take one of the last
> options I always resort to when i have something like this. You know the
> Powell doctrine? (Yes, that Powell - General Colin Powell) - "Think hard
> before you act, act decisively when you do, and finish quickly" . Okay the
> last part might be a bit out of range but the first 2 definately are
> called for at this moment.
Okay.
>
> I also hope that someone somewhere could have an idea or so. Or do you
> suppose I can try -CURRENT?? ;-) ..hmmm.. I am not at par with those
> running -CURRENT and this is a server that is relied upon by hundreds of
> clients - so I am just tied for now.
>
I wish I had something better to tell you. One things you *might* try is this
change:
Change, in scsi_sa.c:samount:
/*
* *Very* first off, make sure we're loaded to BOT.
*/
scsi_load_unload(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG, FALSE,
FALSE, FALSE, 1, SSD_FULL_SIZE, REWIND_TIMEOUT);
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
&softc->device_stats);
QFRLS(ccb);
/*
* In case this doesn't work, do a REWIND instead
*/
if (error) {
scsi_rewind(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG,
FALSE, SSD_FULL_SIZE, REWIND_TIMEOUT);
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
&softc->device_stats);
QFRLS(ccb);
}
if (error) {
xpt_release_ccb(ccb);
goto exit;
}
to
/*
* *Very* first off, make sure we're loaded to BOT.
*/
scsi_rewind(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG,
FALSE, SSD_FULL_SIZE, REWIND_TIMEOUT);
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
&softc->device_stats);
QFRLS(ccb);
if (error) {
xpt_release_ccb(ccb);
goto exit;
}
It's *just* possible that this tape drive is stupid as to not only fail the
LOAD/UNLOAD but then get sent into the weeds and fail to do the rewind.
The reason the LOAD/UNLOAD is used first is to catch the case of tape inserted
but not spun up and loaded to BOT (rewind would fail on those drives). But
this is something of an edge case.
-matt
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0101211115450.1784-100000>