The system:
SMP
IDE
2x750 MHz IA32
1GB RAM
ServerWorks IDE (osb4)
IBM 60GB ATA100
One drive. Here's my partition map
major minor #blocks name
3 0 60051600 hda
3 1 56196 hda1 ext2
3 2 1052257 hda2 ext2
3 3 58942485 hda3 LVM
58 0 2097152 lvma swap
58 1 18874368 lvmb xfs
58 2 4194304 lvmc ext2
58 3 2097152 lvmd ext2
58 4 4194304 lvme xfs
58 5 27484160 lvmf xfs
On Thu, May 24, 2001 at 01:26:17PM -0500, Austin Gonyou wrote:
> A question to ask here is does type of hardware fit into this equation? As
> in :
>
> SMP vs Uni
> SCSI vs IDE
> # of volumes/drives etc.
>
> --
> Austin Gonyou
> Systems Architect, CCNA
> Coremetrics, Inc.
> Phone: 512-796-9023
> email: austin@xxxxxxxxxxxxxxx
>
> On Thu, 24 May 2001, Tom Carroll wrote:
>
> > It is not just a dump. This behavior is evident within most
> > operations that touch the filesystem.
> >
> > Examples include compiling, untarring, samba, etc.
> >
> > -Tom Carroll
> >
> > On Thu, May 24, 2001 at 12:41:52PM -0500, Steve Lord wrote:
> > > > Hey,
> > > >
> > > > I just compiled linux-2.4-xfs 24/05/2001 16:30 GMT.
> > > >
> > > > The large memory consumption issue still exists. The previous
> > > > snapshot of /proc/slabinfo is accurate.
> > > >
> > > > Any suggestions?
> > > > Thanks in advance.
> > > >
> > > > -Tom Carroll
> > >
> > > OK, if the machine survives the failing dump process does the filesystem
> > > unmount afterwards? I will go an run some quick tests, but it smells of
> > > a reference count leak - dump uses a backdoor open function rather
> > > than pathnames, this is probably where the problem is.
> > >
> > > Steve
> > >
> > >
> > >
> >
>
>
|