On Tue, Sep 18, 2001 at 12:39:07PM +0200, Andrea Cerrito wrote:
>  
> -----BEGIN PGP SIGNED MESSAGE-----
> 
> > On Mon, Sep 17, 2001 at 12:57:19PM +0200, Andrea Cerrito wrote:
> > > 12:32:24 Dumping 7f000005.17000002 to
> > > /extra0/17Sep2001/fs-server-b-b-7f000005.17000002 ...
> > > 12:33:13 WriteDump: Error RPC2_SEFAIL3 (F) in CheckSideEffect
> > 
> > SEFAIL3 is 'disk error', a write must have failed.
> 
> Ok.
> 
> > > 12:33:13 Transmitted bytes 385024 != requested bytes 509918!
> > ...
> > > 12:33:13 Transmitted bytes 0 != requested bytes 510493!
> > 
> > is the disk /extra0 where the dumps are going to full?
> 
> Nope. My backup command is: /usr/local/sbin/backup -t 135
> /vice/db/dumplist /codafs/backups/, so coda backups fill
> /codafs/backups, and this is a dir really empty...
> Then, I'll tar/gzip it in /extra0. All is done locally, no tape, no
> remote (at the moment).

The backup program only places links in /codafs/backups, which point to
the actual dumps which are stored in the partitions that are tagged as
"backup" type partitions in /vice/db/vicetab.

Why else would there be a message that the dump is going into extra0?

> > > 12:32:24 Dumping 7f000005.17000002 to /extra0/17Sep2001/fs-server-b-b-7f000005.17000002 ...

And the combined facts that one dump was truncated and all subsequent
ones are 0 bytes kind of gives the strong impression that the disk is
full.

Jan