|
Re: [Mtx-general] mtx error, used with amanda
2007-11-19 15:57:00
On Mon, 19 Nov 2007, Brian Cuttler wrote:
Platform, Solaris 9, Amanda 2.4.4, MTX 1.3.9
Nope, not fixed, here are updated logs, any help would be apreciated.
Oddly... we seem ok on the C2 jukebox but not the L9, same software
on the same machine. Not even matching machines, two amanda configs
on the same box (one runs weekdays and the other saturdays).
Actually, it seems that mtx is doing its job, and a subsequent
command to access the tape or robot shows that things completed
successfully, however there is something wrong with amanda or
the amanda/mtx interface. (but its the same amanda and glue for
both configs...)
Could it be some sort of unidentified race condition with the amanda
library functions and mtx attempting to access some resource at the same
time without any way of preventing overlapping access attempts from
colliding into each other along the way?
My tape host system is a similar config (Sol 9 on Ultra 2, Amanda 2.5.2,
mtx 1.3.11 talking to a Sun-badged L9 array) and it's been operating fine
since all the issues with devices, and configuration were resolved.
Craig.
PS. I'm getting a new cleaning tape for my L9 this week and wondered if
anyone has thought more about how to get cleaning tape usage incorporated
into the Amanda schema somehow? I have left a spare slot in the array
cartridge for the cleaning tape to sit in.
--
Post by Craig Dewick (tm). Web @ "http://lios.apana.org.au/~cdewick";.
Email 2 "cdewick AT lios.apana.org DOT au". SunShack @ "http://www.sunshack.org";
Galleries @ "http://www.sunshack.org/gallery2";. Also lots of tech data, etc.
Sun Microsystems webring at "http://n.webring.com/hub?ring=sunmicrosystemsu";.
|
|
|
