Re: Big filesystems.
On Tue, 2005-02-15 at 15:43 -0500, Adam Skutt wrote:
> Jeffrey W. Baker wrote:
> > Unfortunately XFS also repeatedly swallowed a number of my volumes. I
> > found it to be more unstable than any filesystem I have used (save
> > VxFS). When using XFS, one must not read from the underlying device, or
> > one risks corruption.
> In Linux, doing this on *any* filesystem will potentially cause corruption.
>
> This is why e2dump/e2restore are *unsafe* and not to be used.
>
> Raw I/O operations bypass the buffer cache, so of course it'd be corrupted.
The latter part of this sentence is not supported by the former,
> If you're dicking with device access while a partition is mounted and
> you lose data, you deserve what you get. This behavior has been a no-no
> forever.
Reading from the device never causes corruption of any other filesystem.
You can dump an ext3 filesystem all day long. You'll get a corrupt
dump, but you won't get a corrupt volume.
> This leads one to believe that using XFS on LVM,
> > md, or enbd would be somewhat risky.
> I dunno what 'enbd'
enbd is the enhanced network block device.
> is, but XFS on LVM or md is perfectly safe.
No, actually, it isn't. Look at the XFS mailing list. There's a number
of reports of boinked volumes just in the last month.
> They're
> kernel drivers for starters, so they can coordinate block I/O
> operations. They also sit below XFS in the I/O layer, XFS just sees
> them as another partition.
This does not explain why XFS is frequently reported to become corrupted
when exported over NFS.
> I dunno how you came across this conclusion at all.
I came across this conclusion by losing numerous large filesystems in
the course of only 6 month before abandoning XFS.
-jwb
Reply to: