Fwd: [Gluster-devel] Bug in unify rename

Raghavendra G raghavendra.hg at gmail.com
Wed Jan 23 03:23:40 UTC 2008


resending to list.

---------- Forwarded message ----------
From: Sam Douglas <sam.douglas32 at gmail.com>
Date: Jan 23, 2008 7:06 AM
Subject: Re: [Gluster-devel] Bug in unify rename
To: Raghavendra G <raghavendra.hg at gmail.com>


On 1/23/08, Raghavendra G <raghavendra.hg at gmail.com> wrote:
> Its the intended behaviour since 'mv oldfile newfile' has 'failed'. But a
> small doubt, does 'ls' in the directory rename was performed (here
cluster),
> show newfile or oldfile or both?

While the volume is offline, an ls will show newfile (to me this seems
wrong) -- the
file gets renamed in the namespace volume.

When the volume is brought back online, ls will show both newfile and
oldfile, however
as newfile probably doesn't exist on any volumes, you will get an I/O
error if you try to
read it.

To me it seems that calling rename on a file that is unavailable
should fail immediately
without modifying the namespace, or the file should be `magically'
renamed later when the
volume comes back online.

-- Sam Douglas



-- 
Raghavendra G

A centipede was happy quite, until a toad in fun,
Said, "Prey, which leg comes after which?",
This raised his doubts to such a pitch,
He fell flat into the ditch,
Not knowing how to run.
-Anonymous



More information about the Gluster-devel mailing list