Mailing List Archive

On Sat, Nov 30, 2002 at 06:49:08PM -0700, Ambrose_Earle@shamrockfoods.com wrote:
> Does anyone else experience frequent backup failures?
>
> We have an F840 cluster and we use NetVault to perform ndmpd backups to an
> overland tape library. Many times we receive messages like
>
> Mon Nov 25 02:43:55 MST [filer: ispfc_main:error]: Tape (switch:0.1L3)
> hard media error: Sense_Key (3) ASC (c) ASCQ (0)

We've been fighting this problem on both F840 filers with Quantum ATL
L200 libraries attached for about a year now. We've replaced the
libraries, SCSI cables and terminators, and tapes. The problem would
go away, and then re-appear again. To the date, the jury is still out
on what is the culprit. It still could be the from with ONTAP or
filer hardware. Do you have a case open with NetApp on this?

Igor

Ambrose_Earle@shamrockfoods.com wrote:

>Does anyone else experience frequent backup failures?
>
>We have an F840 cluster and we use NetVault to perform ndmpd backups to an
>overland tape library. Many times we receive messages like
>
>Mon Nov 25 02:43:55 MST [filer: ispfc_main:error]: Tape (switch:0.1L3)
>hard media error: Sense_Key (3) ASC (c) ASCQ (0)
>
>and the backup fails.
>
>Our same system is also used by a UNIX server and it never fails, so is
>DataONTAP more sensitive to these types of issues?
>
>Technically, what triggers this?
>
>Any help would be much appreciated,
>
>
>
>
>
By this time, we encounter strange problems impliying F85, Overland LXL
librairy, and the SCSI card QLogic inside the F85
we use to get some resetting device messages
- here it was a firmware problem on the Overland trying to deal with a
slightly defectuous Qlogic card (with Adaptec cards, there is no problem)
Also on another site, we got some inpredictable failure using an F810
filer, a NeoSerie Overland librairy and the Netvault software,
- here was a dumb cable problems, you have to understand that SCSI
cables are week and could be damaged rather quiclky; always keep some
SCSI cables appart and carfully put them on machines in order the cable
wont be to much twisted
- also, if the backup work well for at least a while, try to shake the
cable and see if the error appear at this time. If yes it definitely
demonstrate a cable failure.

hth ( = hope this helps)