please enable XFS in anaconda
Eric Sandeen
esandeen at redhat.com
Tue Aug 7 14:57:58 UTC 2007
Dave Airlie wrote:
> On Mon, 2007-08-06 at 23:38 -0500, Eric Sandeen wrote:
>> Jesse Keating wrote:
>>> On Thu, 26 Jul 2007 16:32:30 -0500
>>> Dan Yocum <yocum at fnal.gov> wrote:
>>>
>>>> It's stable, widely tested, widely deployed, and
>>>> it's being actively developed and maintained (which is more than can
>>>> be said of some other filesystems that remain in the default list).
>>>> It's in the kernel, it shouldn't be "hidden" in the depths of
>>>> anaconda anymore.
>>> How's the SELinux support these days? And why can't I boot from xfs
>>> yet?
>> Not addressing either of these questions directly, but with regard to
>> overall quality of xfs in F8... I ran the xfsqa test suite on
>> 2.6.23-0.71.rc2.fc8 over the weekend. Of the "auto" test group (those
>> expected to pass reliably), 90 of 93 tests passed (this after fixing a
>> quota bug I found & fixed over the weekend). Of the 3 failures, 2 are
>> "harmless" - i.e. no data corruption, security issues, or anything like
>> that - one is a bleeding-edge allocator feature not working quite 100%,
>> another has to do with slightly different log traffic pattern due to a
>> recent change w.r.t. the expected output.
>>
>> The "real" failure has to do with mmap writes into preallocated space;
>> I'll look into that as I have time.
>
> Don't we ship Fedora with 4K stacks, so I've heard XFS + RAID stuff can
> overflow the stack..
>
> so it may not be the stable..
Yeah, I talked about 4k stacks earlier in this thread - 4k stacks on x86
can cause xfs to run into trouble with more "interesting" io stacks.
(though there are non-trivial cases where ext3 can blow through 4k as
well; see recent LKML threads on 4k stacks...)
The testing above, though, was on x86_64 with a larger stack size.
-Eric
More information about the fedora-devel-list
mailing list