Gluster users - Dec 2014 - Folder disappeared on volume, but exists on bricks.

Hi,
  In case the missing directory path is known, a fresh lookup on that path will
heal the directory entry across the cluster and it will be shown on the mount
point.

e.g on the mount point: ls <COMPLETE PATH of the directory>.

* The directory may not get a fresh lookup on the existing mount. In that case,
the best thing to do would be to create a new fuse mount and do ls on the
directory path.

Thanks,
Susant







----- Forwarded Message -----
> From: "Juan Jos? Pavlik Salles" <jjpavlik at gmail.com>
> To: "Peter B." <pb at das-werkstatt.com>, gluster-users at
gluster.org
> Sent: Monday, December 1, 2014 8:58:09 PM
> Subject: Re: [Gluster-users] Folder disappeared on volume,	but exists on
bricks.
> 
> Indeed, heal only works on replicated volumes, you are correct. Touching
> production storages it's usually a scary thing I know, but restarting
> gluster service shouldn't make the situation worse, it actually could
help.
> My recommendation is: look for a proper time and unmount the volumes, then
> restart gluster services, try to mount the volumes again and look for the
> missing files. If the files aren't visible from the mounted volumes
then, as
> far as I know, you will need to copy your data again.
> 
> 2014-12-01 12:21 GMT-03:00 Peter B. < pb at das-werkstatt.com > :
> 
> 
> 
> 
> Am Mo, 1.12.2014, 16:15 schrieb Juan Jos? Pavlik Salles:
> > Have you tried running the heal operation?
> 
> No, I have not tried to play around unless I know what I'm doing ;)
> 
> I'm trying to find something in the online documentation about such a
> case, but can't find anything :(
> 
> I thought "heal" is for replicated glusterfs, but we're only
using
> distributed [1]. This is a production storage, so I'm very very careful
> before I just try something ;)
> 
> 
> I haven't even dared to stop/restart the glusterd, since I wanted to
ask
> here first.
> 
> 
> Thanks again,
> Pb
> 
> 
> == References:
> [1]
>
https://access.redhat.com/documentation/en-US/Red_Hat_Storage/2.0/html/Administration_Guide/sect-User_Guide-Managing_Volumes-Self_heal.html
> 
> 
> 
> 
> --
> Pavlik Salles Juan Jos?
> Blog - http://viviendolared.blogspot.com
> 
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users
-- 
Thanks,
Anuradha.

Hi Susant,

Am Di, 2.12.2014, 11:49 schrieb Susant Palai:
>   In case the missing directory path is known, a fresh lookup on that path
> will heal the directory entry across the cluster and it will be shown on
> the mount point.
>
> e.g on the mount point: ls <COMPLETE PATH of the directory>.
>
> * The directory may not get a fresh lookup on the existing mount. In that
> case, the best thing to do would be to create a new fuse mount and do ls
> on the directory path.
Thanks for the tip!

Sorry, but obviously I overlooked your reply before.
When you say "existing mount", do you mean I should unmount/remount it
-
or should I rather create a new mount point, like e.g. "/mnt/test"?


Thanks again,
Pb