> hi,
>
> as i wrote before i also experienced this problem.
>
> > Yup, I've been looking into this.
> > What are you doing to blow the 128k cache?
> i just tried a ~90gb large xfs ;-) (my box has 128mb ram and 512mb swap -
> this should be plenty...)
yes it should be plenty, could you run mkfs again on your disk and
send me the output - I want to see how it sizes things, that should
give us some clues. Also cat /proc/slabinfo and /proc/meminfo before and
after the mkfs and send that too.
Generally a clean mount should not have large memory requirements so this
is puzzling - unless the failure is not actually an out of memory thing
but some other issue.
Thanks
Steve
>
> > A "clean" file-system shouldn't be in reovery code, you must
> > have crashed the system at some point.
> i already sent a oops trace i got trying to mount the fresh 90gb
> xfs-partition.
Any chance of at least the ksymoops output from this? The raw numbers
do not do any good without symbol table info attachecd to them.
>
> bernd
|