On Fri, Mar 08, 2002 at 11:36:11AM +0100, Paul Bijnens wrote:
> You broke the report off just too early :-)

Ah, well...

> First Amanda said that it was writing md1 from host "cat" to tape
> while it ran out of tape.
> In the NOTES section of the Amanda report you'll find something
> like: (cut from my last archive report)
> 
> > NOTES:
> >   taper: tape ARCHIVE-040 kb 34115424 fm 111 writing file: short write
> >   driver: going into degraded mode because of tape error.

The report I forwarded yesterday had:

NOTES:
  planner: Adding new disk peterbilt://server/f$.
  planner: Last full dump of bradley:md10 on tape Daily08 overwritten on this run.
  planner: Adding new disk nova:hda1.
  taper: tape Daily08 kb 0 fm 0 writing filemark: Input/output error
  driver: going into degraded mode because of tape error.

Which I'd say confirms that it didn't really run out of tape.

> Or your capacity is for not-hardware compression, and you use both
> hard- and software compression...  resulting in LESS native capicity on tape.

Is hardware compression controlled by the host or the drive?  I'm
currently using a Linux host and a Quantum DLT4000 drive; the drive's
compression LED stays on (my predecessor set this up and I'm only now
having the need to look closely at it), so I assume it's using hardware
compression.  So far, I've only been able to turn it off by using the
drive's "Density Override" button; is there another/a better way?  (Yeah,
I would RTFdriveM, but I have no idea where TFM is.  Disappeared before
I got here.)

> Or you ran into an error somewhere halfway the tape (bad tape? drive? computer?)
> Look into the system logs to find out (dmesg, syslog, etc).

It gets worse...  This morning, I didn't have an amanda report waiting
in my mailbox, so I tried to ssh to the tape host.  No response.
Go into the server room, punch it up on the KVM, and see a kernel panic.
Looks like swap got hosed and the machine locked up while running the
backup last night.  *sigh*

Reply via email to