[Gluster-users] volume create failed

[William Kwan]

Thanks.  Still new to gluster.  Just found a bunch of stuff under /var/lib/glusterd and this webpage
http://joejulian.name/blog/glusterfs-path-or-a-prefix-of-it-is-already-part-of-a-volume/

Cleaned up the bricks according to the URL and I was able to recreate the gluster volumes

[root at onode1 glusterd]# df -k|grep kvm
/dev/sdb1                     10475504   32980  10442524   1% /data/glusterfs/kvm1/brick1
/dev/sdc1                     10475504   32980  10442524   1% /data/glusterfs/kvm2/brick1

[root at onode2 glusterd]# df -k|grep kvm
/dev/sdb1                     10475504   32980  10442524   1% /data/glusterfs/kvm1/brick2
/dev/sdc1                     10475504   32980  10442524   1% /data/glusterfs/kvm2/brick2

[root at onode1 glusterd]# gluster volume create kvm1 replica 2 transport tcp onode1:/data/glusterfs/kvm1/brick1 onode2:/data/glusterfs/kvm1/brick2
volume create: kvm1: success: please start the volume to access data

[root at onode1 glusterd]# gluster volume create kvm2 replica 2 transport tcp onode1:/data/glusterfs/kvm2/brick1 onode2:/data/glusterfs/kvm2/brick2
volume create: kvm2: success: please start the volume to access data

[root at onode1 glusterd]# gluster volume start kvm1
volume start: kvm1: success
[root at onode1 glusterd]# gluster volume start kvm2
volume start: kvm2: success

[root at onode1 glusterd]# gluster volume info

Volume Name: kvm1
Type: Replicate
Volume ID: d8f26ca7-23bc-41df-a299-d9567fcebd2e
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: onode1:/data/glusterfs/kvm1/brick1
Brick2: onode2:/data/glusterfs/kvm1/brick2

Volume Name: kvm2
Type: Replicate
Volume ID: 8f3dad35-ae18-4f0f-8cb9-4b722adc20ed
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: onode1:/data/glusterfs/kvm2/brick1
Brick2: onode2:/data/glusterfs/kvm2/brick2






On Friday, December 27, 2013 11:21 AM, BGM <bernhard.glomm at ecologic.eu> wrote:
 
with kvm2 u only got onode1 listed as the host?
sure there is no typo nowhere?

next: AFAIK with the path you give rather
kvm1 and kvm2 should be the montpoints for your xfs
while you use the directory brick1/2 to create the gluservolume (or you have to use --force, right?)

next: if volume creation fails, make sure you whip out the paths on BOTH sides
(i.e. unset the attr flags and remove the .glusterfs directory from inside the brick path... though know what you are doing!)

I found the later normaly solves this error.
Gluster seems to set at least one of the attr on at least one side already and don't roll it back although not the whole operation succeeded.
Unsetting and redoing without typos or other errors (partner daemon reachable? both partner peered with same adress/name?)

hth
Bernhard

On 27.12.2013, at 15:27, William Kwan <potatok at yahoo.com> wrote:


May be it is confusing or I'm using the wrong naming conventions. what I'm actually testing is
>
>
>Each node has one brick for each of glusterfs kvm1 and kvm2
>So glusterfs kvm1 will have
>
>    onode1:/data/glusterfs/kvm1/brick1
>
>    onode2:/data/glusterfs/kvm1/brick2
>
>while kvm2 will have two bricks
>
>    onode1:/data/glusterfs/kvm2/brick1
>    onode1:/data/glusterfs/kvm2/brick2
>On both systems, "brick1" and "brick2" are mount points of xfs file systems
>
>Now create volume kvm1 was successful while kvm2 is not.
>volume create: kvm2: failed: /data/glusterfs/kvm2/brick1 or a prefix of it is already part of a volume
>
>Thanks
>W
>
>
>
>
>
>
>On Thursday, December 26, 2013 5:24 PM, William Kwan <potatok at yahoo.com> wrote:
> 
>Hi all,
>
>Running 3.4.1 on Centos6.
>
>I have this issue, I created two xfs filesystem on each of two hosts
>onode1:/data/glusterfs/kvm1/brick1
>
>onode2:/data/glusterfs/kvm1/brick2
>
>
>onode1:/data/glusterfs/kvm2/brick1
>onode1:/data/glusterfs/kvm2/brick2
>
>
>The first volume was created successfully
># gluster volume create kvm1 replica 2 transport tcp onode1:/data/glusterfs/kvm1/brick1 onode2:/data/glusterfs/kvm1/brick2
>
>
>When I attempt to create the second volume, I got 
># gluster volume create kvm2 replica 2 transport tcp onode1:/data/glusterfs/kvm2/brick1 onode2:/data/glusterfs/kvm2/brick2
>volume create: kvm2: failed: /data/glusterfs/kvm2/brick1 or a prefix of it is already part of a volume
>
>
>I'm not sure I follow what the error means.  The 2nd set of filesystems have to mount under a totally different directory structure?
>
>
>Thanks in advance
>
>Will
>
>
>
_______________________________________________
>Gluster-users mailing list
>Gluster-users at gluster.org
>http://supercolony.gluster.org/mailman/listinfo/gluster-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20131227/d47537ea/attachment.html>