Hi Eco,
I believe you are asking that I run
find /mount/glusterfs >/dev/null
only? That should take care of the issue?
Thanks for your time,
Steve
________________________________
From: gluster-users-bounces at gluster.org [gluster-users-bounces at gluster.org] on behalf of Eco Willson [ewillson at redhat.com]
Sent: Tuesday, November 20, 2012 5:39 PM
To: gluster-users at gluster.org
Subject: Re: FW: cant mount gluster volume
Steve,
On 11/20/2012 01:32 PM, Steve Postma wrote:
[root at mseas-data gdata]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 18G 6.6G 9.7G 41% /
/dev/sda6 77G 49G 25G 67% /scratch
/dev/sda3 18G 3.8G 13G 24% /var
/dev/sda2 18G 173M 16G 2% /tmp
tmpfs 3.9G 0 3.9G 0% /dev/shm
/dev/mapper/the_raid-lv_home
3.0T 2.2T 628G 79% /home
glusterfs#mseas-data:/gdata
15T 14T 606G 96% /gdata
[root at nas-0-0 ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 137G 33G 97G 26% /
/dev/sda1 190M 24M 157M 14% /boot
tmpfs 2.0G 0 2.0G 0% /dev/shm
/dev/sdb1 21T 19T 1.5T 93% /mseas-data-0-0
[root at nas-0-1 ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 137G 34G 97G 26% /
/dev/sda1 190M 24M 157M 14% /boot
tmpfs 2.0G 0 2.0G 0% /dev/shm
/dev/sdb1 21T 19T 1.3T 94% /mseas-data-0-1
Thanks for confirming.
cat of /etc/glusterfs/glusterd.vol from backup
[root at mseas-data glusterd]# cat /root/mseas_backup/etc/glusterfs/glusterd.vol
volume management
type mgmt/glusterd
option working-directory /etc/glusterd
option transport-type socket,rdma
option transport.socket.keepalive-time 10
option transport.socket.keepalive-interval 2
end-volume
The vol file for 2.x would be in /etc/glusterfs/<volume name>.vol I believe. It should contain an entry similar to this output for each of the servers toward the top of the file.
Article you referenced is looking for the words "glusterfs-volgen" in a vol file. I have used locate and grep, but can find no such entry in any .vol files.
This would not appear if the glusterfs-volgen command wasn't used during creation. The main consideration is to ensure that you have the command in step 5:
find /mount/glusterfs >/dev/null
- Eco
Thanks
________________________________
From: gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org> [gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org>] on behalf of Eco Willson [ewillson at redhat.com<mailto:ewillson at redhat.com>]
Sent: Tuesday, November 20, 2012 4:03 PM
To: gluster-users at gluster.org<mailto:gluster-users at gluster.org>
Subject: Re: FW: cant mount gluster volume
Steve,
On 11/20/2012 12:03 PM, Steve Postma wrote:
The do show expected size. I have a backup of /etc/glusterd and /etc/glusterfs from before upgrade.
Can we see the vol file from the 2.x install and the output of df -h for
each of the bricks?
Its interesting that "gluster volume info" shows the correct path for each machine.
These are the correct mountpoints on each machine, and from each machine I can see the files and structure.
If the volume was created in a different order than before, then it is
expected you would be able to see the files only from the backend
directories and not from the client mount.
If this is the case, recreating the volume in the correct order should
show the files from the mount.
If the volume was recreated properly, make sure you have followed the
upgrade steps to go from versions prior to 3.1:
http://www.gluster.org/community/documentation/index.php/Gluster_3.0_to_3.2_Upgrade_Guide
This would explain why the files can't be viewed from the client, but
the size discrepancy isn't expected if we see the expected output from
df for the bricks.
[root at mseas-data<mailto:root at mseas-data><mailto:root at mseas-data> data]# gluster volume info
Volume Name: gdata
Type: Distribute
Volume ID: eccc3a90-212d-4563-ae8d-10a77758738d
Status: Started
Number of Bricks: 3
Transport-type: tcp
Bricks:
Brick1: gluster-0-0:/mseas-data-0-0
Brick2: gluster-0-1:/mseas-data-0-1
Brick3: gluster-data:/data
________________________________
From: gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org> [gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org>] on behalf of Eco Willson [ewillson at redhat.com<mailto:ewillson at redhat.com><mailto:ewillson at redhat.com><mailto:ewillson at redhat.com>]
Sent: Tuesday, November 20, 2012 3:02 PM
To: gluster-users at gluster.org<mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org>
Subject: Re: FW: cant mount gluster volume
Steve,
Does df -h show the expected directories on each server, and do they
show the expected size?
If the file
On 11/20/2012 11:09 AM, Steve Postma wrote:
Hi Eco, thanks for your help.
If I run on brick 1:
mount -t glusterfs gluster-data:/gdata /gdata
it mounts but appears as a 18 GB partition with nothing in it
To confirm, are the export directories mounted properly on all three
servers?
Does df -h show the expected directories on each server, and do they
show the expected size?
Does gluster volume info show the same output on all three servers?
I can mount it from the client, but again, there is nothing in it.
Before upgrade this was a 50 TB gluster volume. Was that volume information lost with upgrade?
Do you have the old vol files from before the upgrade? It would be good
to see them to make sure the volume got recreated properly.
The file structure appears intact on each brick.
As long as the file structure is intact, you will be able to recreate
the volume although it may require a potentially painful rsync in the
worst case.
- Eco
Steve
________________________________
From: gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org> [gluster-users-bounces at gluster.org<mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org><mailto:gluster-users-bounces at gluster.org>] on behalf of Eco Willson [ewillson at redhat.com<mailto:ewillson at redhat.com><mailto:ewillson!
@redhat.co
m><mailto:ewillson at redhat.com><mailto:ewillson at redhat.com><mailto:ewillson at redhat.com>]
Sent: Tuesday, November 20, 2012 1:29 PM
To: gluster-users at gluster.org<mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org><mailto:gluster-users at gluster.org>
Subject: Re: FW: cant mount gluster volume
Steve,
The volume is a pure distribute:
Type: Distribute
In order to have files replicate, you need
1) to have a number of bricks that is a multiple of the replica count,
e.g., for your three node configuration, you would need two bricks per
node to set up replica two. You could set up replica 3, but you will
take a performance hit in doing so.
2) to add a replica count during the volume creation, e.g.
`gluster volume create <vol name> replica 2 server1:/export server2:/export
