Amanda-Users

Re: scsi-changer debugging...

2004-03-13 02:40:54
Subject: Re: scsi-changer debugging...
From: Christoph Scheeder <christoph.scheeder AT scheeder DOT de>
To: Kristian Rink <krink AT pm-planc DOT de>
Date: Sat, 13 Mar 2004 08:36:33 +0100
Hi,
seems like chg-scsi does not know your changer/tape at the moment,
and as a result can't handle the messages it get's back from it.
if you have a little experience with programming, have a descrition
of the sense-codes your changer generates and a little bit time left,
you can take a look at chg-scsi.c/h in the sources of amanda.
There are some tables defining the sensecodes returned by the changer
and tape and the actions to be taken after them.
These tables are quite forward, so it is not much work to adapt them.
And if you get it working don't forget to send a patch to
amanda-hackers, so in a future release your changer will be supported
out of the box.
Christoph


Kristian Rink schrieb:
Hi all,...

...after still having some weird issues with my backup runs not
running as expected (scheduled) running chg-scsi, I was playing
around with the debug files a little, and here's the result (see
attached: "scsidebug"): basically, after "START SCSI_LoadUnload",
messages like

---snip---

##### START SenseHandler
Ident = [ULTRIUM-TD2], function = [generic_changer]
##### START GenericSenseHandler
##### START DecodeSense
GenericSenseHandler : Sense Keys
Extended Sense ASC 04
        ASCQ                          00
        Sense key                     02
                Not Ready
Sense2Action START : type(1), ignsense(0), sense(02), asc(04),
ascq(00) Sense2Action generic start :
Sense2Action generic END : no match for generic   return ->
2/Default for SENSE_ NOT_READY
##### STOP GenericSenseHandler
#### STOP SenseHandler

---snip---

keep filling up my chg-scsi.*.debug log; most of the times the file
itself grows up to 500k just filled with those message. There
doesn't seem to be a way to reproduce why exactly this is happening,
as it doesn't seem to depend on whether the right tape is in drive
or no tape is loaded at all. It also doesn't happen all the time;
sometimes am[check|dump] just run through smoothly without any
trouble at all. I'm currently thinking about several different ways
of dealing with this, including usage of chg-multi instead of
chg-scsi but I wanted to check whether anyone actually has
experiences already running amanda using the following device:

backer:~# loaderinfo -f /dev/sg4
Product Type: Medium Changer
Vendor ID: 'NEC     '
Product ID: 'LL0101H-0A      '
Revision: '0002'
Attached Changer: No
EAAP: Yes
Number of Medium Transport Elements: 1
Number of Storage Elements: 10
Number of Import/Export Element Elements: 0
Number of Data Transfer Elements: 1
Transport Geometry Descriptor Page: Yes
Invertable: No
Device Configuration Page: Yes
Can Transfer: No

I attached a scsidebug output and my changer.conf just in case
anyone might wants to look at it...
TIA and have a calm friday / weekend anyone...
Cheers,
Kris


<Prev in Thread] Current Thread [Next in Thread>