Alexandre Oliva wrote:
>
> On Jun 5, 2001, Eric Sproul <[EMAIL PROTECTED]> wrote:
>
> > I have continued to experience problems with writing filemarks and
> > getting 'bad file descriptor' errors on nightly Amanda runs as well as
> > with amflush.
>
> Sometimes, all this means is that you're reaching EOT. Does tape
> usage match this assumption?
No, and that's why I'm concerned. For example, last night's run:
These dumps were to tape daily-36.
*** A TAPE ERROR OCCURRED: [[writing filemark: Input/output error]].
Some dumps may have been left in the holding disk.
Run amflush to flush them to tape.
The next tape Amanda expects to use is: daily-37.
FAILURE AND STRANGE DUMP SUMMARY:
(numerous disks) [write_tapeheader: No space left on device]
(more disks) [no more holding disk space]
Only about half my disks dumped because I ran out of holding disk
space. Total output size after compression is 66MB. These are 8mm 20GB
Mammoth tapes. Tape usage is reported at 0.0% so it looks like nothing
was written to tape. I have a total of 8.8GB of holding disk, but since
nothing was taped it filled up rather quickly.
I am having a replacement Mammoth drive shipped to me so I am going to
try swapping the drive. I have changed just about everything else.
These problems began on 5/23, and that night I heard that the office may
have had a "power event" so perhaps the drive was damaged in that. It's
just a guess. :)
Thanks,
Eric
--
Eric Sproul, Unix Systems Engineering
Cornerstone Networks Inc.
An NTELOS Communications Company
http://www.cstone.net
