[ale] Ext4 adoption anyone?

[Jim Kinney]

Yeah, but Fedora default use almost guarantees it is unstable :-)

<written on a Fedora 10 machine...>

2009/1/23 Dennis Ruzeski <denniruz at gmail.com>:
> Here's a nice nugget from Slashdot-- Fedora 11 is using it as the default-
> http://linux.slashdot.org/article.pl?sid=09/01/23/1341237
>
>
>
> On Fri, Jan 23, 2009 at 10:14 AM, Jim Kinney <jim.kinney at gmail.com> wrote:
>>
>> Adding to this: Recent XFS seems to have destabilised somehow. Unless
>> you need single file sizes of  4+GB (i.e. pushing digital video files
>> around), XFS is not a good idea.
>>
>> JFS actually won a filesystem test in the past year or so for overall
>> usefulness, speed and reliability. It is a good general purpose
>> filesystem with a solid journalling system.
>>
>> EXT3 is quite stable except for a few, odd corner cases. The inability
>> to recover data inodes from a file deletion is perceived as a bad
>> thing.
>>
>> IF (!!) ZFS ever become available as a GPL addition to the kernel, we
>> will see some useful things happen in filesystems.  ZFS in Solaris is
>> pretty rock-solid.
>>
>> FUSE adds it's own layer of 'funk' to the mix. My experience has been
>> that fuse is mostly reliable. But an unravelling fuse stream can
>> destabalise mount-point end of the fuse'ed system. The source end
>> seems to be unaffected. Half-mounted, locked, unable to remount or
>> remove when it fizzles from 2.6.26 through 2.6.28. I have not decided
>> if the issue is with fuse or with the stupid gui mounting tools from
>> gnome (I'd put my money on gnome hosing something first!). Of course
>> it's not reliably repeatable breakage either.
>>
>> 2009/1/23 Michael B. Trausch <mike at trausch.us>:
>> > On Fri, 23 Jan 2009 09:04:39 -0500
>> > Jeff Hubbs <hbbs at comcast.net> wrote:
>> >
>> >> These newer filesystems, including the ones used through FUSE - are
>> >> any of them considered as reliable as ext3, jfs, xfs?
>> >
>> > I'd imagine that would depend on who you ask.  I've had so-called
>> > stable filesystems totally nuke my data, and filesystems in
>> > development that never so much as sneezed for me.
>> >
>> > I trust ext4 because of its common code with ext3.  And in fact, ext4
>> > exposed ext3 to new eyes because there are neophytes out there, and
>> > bugs were found in ext4 that were inherited from ext3 and then were
>> > fixed in both.  I trust XFS less than I trust FAT32 because I have
>> > never had a filesystem nuke my data as completely and efficiently as it
>> > did back in 2.6.25 days.  I still see people reporting major bugs
>> > (data loss ranging from small to extreme) in XFS to the kernel mailing
>> > list, and I personally wouldn't touch it with a ten foot pole.
>> >
>> > Looking at my LKML folder today, I see 22 emails on problems with XFS,
>> > compared to 6 for ext4.  Ted Tso jumped on that thread immediately (the
>> > ext4 problem, which reported ENOSPC with ~500 MB left on the
>> > filesystem); turns out the problem is that the user is unable to use
>> > the very last 1% of their root filesystem, and it's being worked on
>> > now.
>> >
>> > I see nothing on JFS; that could mean that there is nobody using that
>> > filesystem or that it is virtually bug-free.  Don't know which.  I also
>> > see nothing for vfat, and 7 messages in 4 threads for btrfs, though
>> > that is probably because only a few people are attempting to use btrfs
>> > at this point.
>> >
>> > As far as FUSE filesystems go, I haven't used many.  But I do know that
>> > of the ones I have used (sshfs, WikipediaFS, NTFS-3G, and Captive NTFS)
>> > I have never had any problems with the FUSE driver itself.  In the case
>> > of sshfs, the only problem that I have ever had was with the remote
>> > host dropping the connection after a timeout, which is easily fixed by
>> > keeping the filesystem active or by just remounting the filesystem.  I
>> > would like to play more with FUSE since the variety of file systems
>> > available there is kind of impressive.  The idea of being able to mount
>> > just about anything is kind of nice, and being in userspace, there may
>> > be lossage, but there won't be kernel crashes, and that's a big plus.
>> > That said, I haven't played enough with any single FUSE filesystem in
>> > order to be able to make any substantial positive claims on it, but it
>> > hasn't broke yet for me.
>> >
>> >        --- Mike
>> >
>> > --
>> > My sigfile ran away and is on hiatus.
>> > http://www.trausch.us/
>> >
>> > _______________________________________________
>> > Ale mailing list
>> > Ale at ale.org
>> > http://mail.ale.org/mailman/listinfo/ale
>> >
>> >
>>
>>
>>
>> --
>> --
>> James P. Kinney III
>> _______________________________________________
>> Ale mailing list
>> Ale at ale.org
>> http://mail.ale.org/mailman/listinfo/ale
>
>
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
>
>



-- 
-- 
James P. Kinney III