I'll have to check to see what the BIOS version and driver version are,
but I can tell you that when the problem happens, it typically manifests
itself in one very obvious way. When I come in the next morning, I'll
notice that one, sometimes two (not always the same ones, though) of the
drives will have an error message in the devices window on the nwadmin
GUI. It will say something to the effect of:
read open error, Device or resource busy
read open error, Invalid argument
or some such thing. If you attempt to run /etc/LGTOuscsi/inquire then
the propgram will fail to report the picker, sometimes the one on the
ATL library, too, but always on the STK. Somehow, Legato just doesn't
see the picker anymore. Of course, a quick check reveals that there's at
least one, sometimes several, nsrjb processes still running. Go figure.
They never completed because the picker disappeared. The only fix is to
stop the running groups, reboot the storagenode server and then re-start
the groups. Just rebooting the storagenode server alone will not resume
the groups. If everything is running smoothly, the inquire tool is
always happy.
George
"McDougal, Philip H" wrote:
>
> George, what bios and driver version are you running for the 39160?
>
> One version, u160_xp_2000_sp4 (I think, it's been a while, could have been
> sp3a) gave me a beautiful BSOD. I finally got things moving with version
> u160fms40_sp5_s1 (driver version 6.1.6300 on 7/15/03) for the 39160 (for
> windows 2000). The Bios I'm runnig is 3.10.0
>
> I've also seen the "\Device\scsi\adpu160m4,did not respond". The m4 tells me
> what bus/drive is affected. Unfortuntately, as George so eloquently put it,
> it's hard to tell what is causing the point of failure and it's quite
> frustrating to continously deal with an issue thats been around for a while.
>
> Regards,
> Phil.
>
> -----Original Message-----
> From: George Sinclair [mailto:George.Sinclair AT NOAA DOT GOV]
> Sent: Wed 10/8/2003 11:41 AM
> To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> Cc:
> Subject: Re: [Networker] LTO Tapes being marked full prematurely
>
> We have Seagate, first generation LTO drives, and we occasionally see
> this problem. When it occurs, there is typically always a SCSI
> error(s)
> in the system log somewhere around the time the tape was prematurely
> marked full. In other words, something happens on the SCSI bus,
> NetWorker sees it, freaks out and marks the tape full because in its
> little mind, it's doing you a favor by regarding the tape as suspect,
> as
> in "I'm not gonna write to this tape any more because something
> happened, and I'm not convinced that the tape was not the culprit". I
> guess I can remotely understand the reasoning, but it is annoying when
> the tape has plenty of space, and you know it was just a SCSI error.
> The
> real question is, what causes these SCSI errors? They seem sporadic
> and
> unpredictable at best.
>
> For a while we were under the impression that it was a drive problem,
> but then the attention turned to the Adaptec SCSI driver as has been
> observed and commented on by several people in this forum. We checked
> cables, replaced cables with the proper LVD ones, replaced the
> terminators, and at least two drives. Same problems. We're running
> Linux
> on the storagenode server, and the library is a StorageTek L80. We
> finally replaced the kernel code with a new version of the Adaptec
> driver code to see if that would fix the problem. The problem has
> occurred once since the update which I think was maybe two weeks ago.
> I
> should note that we're also running an ATL P1000 SDLT tape library on
> this same storagenode server, but it runs on a separate Adaptec card.
> Maybe having both libraries on there, even though they're running on
> separate cards, is a possible problem? I should also note that in the
> case of the StorageTek library, two LTO drives and the picker share
> one
> channel, and the other two drives share the other channel, each
> channel
> has one SCSI cable going to its respective port on the back of the
> dual
> channel Adaptec 39160 card. The ATL has two drives and one picker
> which
> share a single channel (one SCSI cable) to its dual Adaptec 39160
> card.
> Two separate cards. We also have the latest BIOS on both cards. Who
> knows?
>
> George
>
> "Willeat, Todd" wrote:
> >
> > Hi Phillip,
> >
> > I'm also having problems with the LTO-2 drives in my i2000. We keep
> getting
> > and error 6 on the drive, which is supposed to be a media problem.
> > Unfortunately, when that happens, the library takes that drive
> offline and
> > NetWorker just sits for hours trying to access it and doesn't do
> anything
> > with the other drives that still work.
> >
> > BTW, the drives are IBM, not HP...
> >
> > -----Original Message-----
> > From: McDougal, Philip H [mailto:PMcDougal AT JENNER DOT COM]
> > Sent: Wednesday, October 08, 2003 10:04 AM
> > To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> > Subject: Re: [Networker] LTO Tapes being marked full prematurely
> >
> > Is it me or are there quite a few people who keep having LTO-2 drive
> > failures??? I have an ADIC Scalar i2000 with 6 LTO-2 drives and
> had 1 DOA,
> > another failure about a week later. Last night (almost 2 months
> later), I
> > have 2 more dead drives.
> >
> > Do you think it's the ADIC library or the HP drives that are the
> problem (or
> > both)? I'd like to know who I should be venting to before I go off
> on both
> > of them ;-)
> >
> > Phil.
> >
> > -----Original Message-----
> > From: Rada Uspensky [mailto:ruspensky AT RA.ROCKWELL DOT COM]
> > Sent: Monday, October 06, 2003 4:38 PM
> > To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> > Subject: Re: [Networker] LTO Tapes being marked full prematurely
> >
> > We had similar situation due to the tape drives malfunction.
> > It started slowly but in a couple weeks we had hundred tapes marked
> full
> > prematurely.
> > Check the event log for the drive related errors and contact the
> vendor to
> > fix or replace drives.
> >
> > Davina Treiber <Treiber AT HOTPOP DOT COM>@LISTMAIL.TEMPLE.EDU> on
> 10/06/2003
> > 04:22:39 PM
> >
> > Please respond to Legato NetWorker discussion
> > <NETWORKER AT LISTMAIL.TEMPLE DOT EDU>; Please respond to
> Davina Treiber
> > <Treiber AT HOTPOP DOT COM>
> >
> > Sent by: Legato NetWorker discussion <NETWORKER AT
> LISTMAIL.TEMPLE DOT EDU>
> >
> > To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> > cc:
> >
> > Subject: Re: [Networker] LTO Tapes being marked full prematurely
> >
> > Maine, Daniel wrote:
> > > Hello!
> > >
> > > We are using Networker 7x on a Win2000 server. It is attached to
> a
> > STKL180
> > > by SCSI.
> > >
> > > We are experiencing problems with the LTO tapes being marked
> "full"
> > before
> > > they have used it's allotted space. Some are being marked "full'
> with
> > only
> > > 2g.
> > >
> >
> > What brand of drives and what brand of tapes?
> >
> > If they are Seagate drives, they will only work well with certain
> tape
> > brands. Maxell/Fuji/Seagate are the only ones certain to work,
> although
> > some others are Maxell or Fuji rebadged.
> >
> > --
> > Note: To sign off this list, send a "signoff networker" command via
> email
> > to listserv AT listmail.temple DOT edu or visit the list's Web site
> at
> > http://listmail.temple.edu/archives/networker.html where you can
> > also view and post messages to the list.
> >
> > =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
> >
> > --
> > Note: To sign off this list, send a "signoff networker" command via
> email
> > to listserv AT listmail.temple DOT edu or visit the list's Web site
> at
> > http://listmail.temple.edu/archives/networker.html where you can
> > also view and post messages to the list.
> > =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
> >
> > --
> > Note: To sign off this list, send a "signoff networker" command via
> email
> > to listserv AT listmail.temple DOT edu or visit the list's Web site
> at
> > http://listmail.temple.edu/archives/networker.html where you can
> > also view and post messages to the list.
> > =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
> >
> > ---
> > Incoming mail is certified Virus Free.
> > Checked by AVG anti-virus system (http://www.grisoft.com).
> > Version: 6.0.515 / Virus Database: 313 - Release Date: 9/1/2003
> >
> > ---
> > Outgoing mail is certified Virus Free.
> > Checked by AVG anti-virus system (http://www.grisoft.com).
> > Version: 6.0.515 / Virus Database: 313 - Release Date: 9/1/2003
> >
> > --
> > Note: To sign off this list, send a "signoff networker" command via
> email
> > to listserv AT listmail.temple DOT edu or visit the list's Web site
> at
> > http://listmail.temple.edu/archives/networker.html where you can
> > also view and post messages to the list.
> > =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
> --
> Note: To sign off this list, send a "signoff networker" command via
> email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
