[Dirvish] dirvish write error, filesystem probably full

Paul Slootman paul at debian.org
Wed Jul 27 11:57:50 PDT 2005


On Wed 27 Jul 2005, Jason Boxman wrote:
> On Wednesday 27 July 2005 11:57, Paul Slootman wrote:
> <snip>
> > I started getting these too on one of my backup systems at work, even
> > though the file system had 12% free (that's 126GB ;-). It's also not a
> > question of the filesystem being temporarily full, the system is
> > monitored by munin and it doesn't get above 92% full.
> 
> Obvious question, but are you running ext2/3?  Maybe you ran out of i-nodes?
> <snip>
> > I have had a reiserfs that reported "full" even though there was plenty
> > of space, a reiserfsck fixed that. However I'm not looking forward to
> > fscking 1TB :-(
> 
> Of course if you're running reiserfs, nevermind. :)

:-)

I've been doing some tests on a box with 24 400GB SATA disks... ext3
doesn't even create filesystems larger than 8TB :-) reiserfs 3 oopses
the kernel, reiser4 works fine, but I'm probably going with jfs as that
gave the best performance.

> Another remote possibility is a process that's holding onto a large file or 
> many files that have been unlinked, but not yet freed because the process is 
> still alive.  (My favorite instance is when X fills up ~/.xsession-errors and 
> of course unlinking it doesn't help you, you have to zero the file instead.)

Could be the case for the original poster; not for me, as (as I said) I
monitor the filesystem usage (via munin) and the max. is 92%.

Tomorrow more news, after tonight's run...


Paul Slootman


More information about the Dirvish mailing list