On Wed, Mar 22, 2017 at 11:08:27AM -0400, Amir Goldstein wrote:
> On Wed, Mar 22, 2017 at 10:06 AM, Zorro Lang <zlang@xxxxxxxxxx> wrote:
> > On Wed, Mar 22, 2017 at 06:34:00AM -0400, Amir Goldstein wrote:
> >> On Wed, Mar 22, 2017 at 2:46 AM, Zorro Lang <zlang@xxxxxxxxxx> wrote:
> >> > The format of glusterfs' TEST_DEV or SCRATCH_DEV is XXX:XXX or
> >> > XXX:/XXX, but xfstests can't accept the latter now. So change
> >>
> >> Why can't xfstest accept the latter?
> >
> > Because I use "\w:\w" in commit "4cbc0a0 fstests: add GlusterFS support".
> > More details please see below.
> >
> >>
> >> > the regular expression from "\w:\w" to ":/?", to accept more
> >> > glusterfs device format.
> >> >
> >> > Signed-off-by: Zorro Lang <zlang@xxxxxxxxxx>
> >> > ---
> >> >
> >> > Hi,
> >> >
> >> > According to the feedback from glusterfs developer:
> >> >
> >> > "Yes, it would be good to have an optional "/" after the ":". It is not
> >> > required, but would probably help when someone runs the tests with the
> >> > "hostname:/volume" device format."
> >> >
> >>
> >> The developer also said
> >> "We mostly use the format of 192.168.1.1:/testvol, matching the way NFS"
> >> If we can enforce the more common format which xfstest already expects
> >> for nfs*|ceph)
> >> Why would we want to support another optional format and have a special
> >> test case for it?
> >
> > Niels is the developer of glusterfs' NFS module (sorry, I forgot its project
> > name), he test it likes test NFS:
> > mount -t nfs 192.168.1.1:/XXXX /mnt
> >
> > So he only can use XXXX:/XXX format, because xfstests limit that format
> > to mount NFS.
> >
> >>
> >> > Maybe there's a slash before the gluster volume name. This patch
> >> > try to support this format.
> >> >
> >> > Thanks,
> >> > Zorro
> >> >
> >> > common/rc | 4 ++--
> >> > 1 file changed, 2 insertions(+), 2 deletions(-)
> >> >
> >> > diff --git a/common/rc b/common/rc
> >> > index edfba5d..e1ab2c6 100644
> >> > --- a/common/rc
> >> > +++ b/common/rc
> >> > @@ -1496,7 +1496,7 @@ _require_scratch_nocheck()
> >> > {
> >> > case "$FSTYP" in
> >> > glusterfs)
> >> > - echo $SCRATCH_DEV | grep -q "\w:\w" > /dev/null 2>&1
> >> > + echo $SCRATCH_DEV | egrep -q ":/?" > /dev/null 2>&1
> >
> > As I know, mount a glusrerfs always do like this:
> > mount -t glusterfs 192.169.1.1:testvol /mnt
> > or
> > mount -t glusterfs 192.169.1.1:/testvol /mnt
> >
> > So glusterfs can use XXXX:XXX or XXXX:/XXX format, but if mount a nfs,
> > we only can use XXXX:/XXX format.
> >
> > So that's why NFS use `grep -q ":/"`, but glusterfs use `egrep -q ":/?"`
> >
>
> I understand.
> What I am saying is that if glusterfs *can* use XXXX:/XXX
> and a file system tester will dedicate a specific mount for xfstests,
> so tester can use the NFS format for that mount and we can require that.
>
> Is that really a big deal for the tester?
>
> If we do that we will have less special cases for format of *_DEV
> and I think that is a gain.
Hah, I already added many things for glusterfs in commit "4cbc0a0
fstests: add GlusterFS support", so I think it doesn't matter to
accept both formats for glusterfs too :)
Thanks,
Zorro
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
