On Fri, 2002-10-04 at 03:41, Sidik Isani wrote:
> On Fri, Oct 04, 2002 at 09:04:33AM +0200, Seth Mos wrote:
> > At 17:04 3-10-2002 -1000, Sidik Isani wrote:
> > >Oct 3 16:32:10 pelmo kernel: Out of Memory: Killed process 806
> > >(xfs_repair).
> > >
> > >Hello again -
> > >
> > > We have a bit of a problem: we have needs for very large filesystems
> > > (more than 1 TB) but no need to scale the memory in the machines
> > > the same way ... except if you ever need to run xfs_repair on the
> > > filesystem, it seems!
> > That isn't right. Can you run it through strace and get the top and bottom
> > part?
> Yes. I put the whole thing at:
> Memory usage actually jumped to almost 800 MB right away and stayed
> there for a while. Then it went back under 10 MB, and then to 30
> for a while before exiting.
OK, I went and had a look, and we are actually as efficient as we can
be in use of the memory itself. Only problem is that we allocate 4 times
what we use. I fixed this in cvs.
Patch against previous repair code is attached.
Steve Lord voice: +1-651-683-3511
Principal Engineer, Filesystem Software email: lord@xxxxxxx
Description: Text document