Re: Problems repairing volume

[Kevin Bridges <kevin@xxxxxxxxxxxxx>]

Thank you for the reply.  I've attempted the volume heal command and ran the stat command from a client that has this volume mounted.  It does not appear that the bricks are replicating all the files.

[root@gluster01 nmd]# find . -type f | wc -l
824371
[root@gluster02 nmd]# find . -type f | wc -l
741043

[root@gluster01 sdb1]# gluster volume heal nmd info
Brick gluster01.newmediadenver.com:/srv/sdb1/nmd/
Number of entries: 0

Brick gluster02.newmediadenver.com:/srv/sdb1/nmd/
Number of entries: 0

[root@gluster01 nmd]# gluster volume status nmd
Status of volume: nmd
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick gluster01.newmediadenver.com:/srv/sdb
1/nmd                                       49152     0          Y       27680
Brick gluster02.newmediadenver.com:/srv/sdb
1/nmd                                       49152     0          Y       20921
NFS Server on localhost                     N/A       N/A        N       N/A
Self-heal Daemon on localhost               N/A       N/A        Y       27673
NFS Server on 10.0.2.85                     N/A       N/A        N       N/A
Self-heal Daemon on 10.0.2.85               N/A       N/A        Y       20914

Task Status of Volume nmd
------------------------------------------------------------------------------
There are no active volume tasks

On Fri, Jul 17, 2015 at 8:25 AM, Curro Rodriguez <curro@xxxxxxxx> wrote:

Hello Kevin,

I think you can't rebalance because you are not using a distributed gluster, you are using replicated gluster instead, I had a similar problem some time ago, and It is supposed that gluster self-heal will sync the replicas.

You can trigger the heal with:

gluster volume heal nmd full

Anyway after reading and googling a lot one member told me that we could resync fast if you do

find -exec stat {} \;

on a mounted client. This solution wasn't the ideal but started to resync faster than heal-self. Still slow 5MB/s but better than nothing.

I am starting with glusterfs too, but I am sure someone could help you better than me.

Kind regards.

Curro Rodríguez.

On Fri, Jul 17, 2015 at 4:15 PM, Kevin Bridges <kevin@xxxxxxxxxxxxx> wrote:

I lost one of my bricks and attempted to rebuild it.  I did not understand what I was doing and created a mess.  I'm looking for guidance so that I don't create a bigger mess.

I believe that the gluster mount is relatively simple.  It's a 2 brick replicated volume (gluster01 & gluster02).  I lost gluster02 and attempted to replace it.  Now that it is replaced, the files do not seem to match what is on the brick that I did not loose.  I would like to repair these bricks and then add more storage capacity to the volume.

Below is the output of the `df -h`, `gluster peer status`, and `gluster volume info` for each of the servers.  I'm concerned by the `gluster peer status` and `gluster volume rebalance nmd status` commands.

Any help is vastly appreciated.

glusterfs 3.7.2 built on Jun 23 2015 12:13:13

[root@gluster01 /]# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/xvde       7.9G  2.3G  5.3G  30% /
tmpfs           3.7G     0  3.7G   0% /dev/shm
/dev/xvdf1       63G   38G   23G  62% /srv/sdb1
[root@gluster01 /]# gluster peer status
Number of Peers: 1

Hostname: 10.0.2.85
Uuid: 5f75bd77-0faf-4fb8-9819-83326c4f77f7
State: Peer in Cluster (Connected)
Other names:
gluster02.newmediadenver.com
[root@gluster01 /]# gluster volume info all

Volume Name: nmd
Type: Replicate
Volume ID: 62bec597-b479-4bfd-88dc-44f5bb88d737
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: gluster01.newmediadenver.com:/srv/sdb1/nmd
Brick2: gluster02.newmediadenver.com:/srv/sdb1/nmd
Options Reconfigured:
performance.readdir-ahead: on
[root@gluster01 /]# gluster volume info nmd

Volume Name: nmd
Type: Replicate
Volume ID: 62bec597-b479-4bfd-88dc-44f5bb88d737
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: gluster01.newmediadenver.com:/srv/sdb1/nmd
Brick2: gluster02.newmediadenver.com:/srv/sdb1/nmd
Options Reconfigured:
performance.readdir-ahead: on
[root@gluster01 /]# gluster volume rebalance nmd status
volume rebalance: nmd: failed: Volume nmd is not a distribute volume or contains only 1 brick.
Not performing rebalance

[root@gluster02 /]# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/xvde       7.9G  2.5G  5.1G  33% /
tmpfs           3.7G     0  3.7G   0% /dev/shm
/dev/xvdh1       63G   30G   31G  50% /srv/sdb1
[root@gluster02 /]# gluster peer status
Number of Peers: 1

Hostname: gluster01.newmediadenver.com
Uuid: afb3e1c3-de9e-4c06-ba5c-5551b1d7030e
State: Peer in Cluster (Connected)
[root@gluster02 /]# gluster volume info all

Volume Name: nmd
Type: Replicate
Volume ID: 62bec597-b479-4bfd-88dc-44f5bb88d737
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: gluster01.newmediadenver.com:/srv/sdb1/nmd
Brick2: gluster02.newmediadenver.com:/srv/sdb1/nmd
Options Reconfigured:
performance.readdir-ahead: on
[root@gluster02 /]# gluster volume info nmd

Volume Name: nmd
Type: Replicate
Volume ID: 62bec597-b479-4bfd-88dc-44f5bb88d737
Status: Started
Number of Bricks: 1 x 2 = 2
Transport-type: tcp
Bricks:
Brick1: gluster01.newmediadenver.com:/srv/sdb1/nmd
Brick2: gluster02.newmediadenver.com:/srv/sdb1/nmd
Options Reconfigured:
performance.readdir-ahead: on
[root@gluster02 /]# gluster volume rebalance nmd status
volume rebalance: nmd: failed: Volume nmd is not a distribute volume or contains only 1 brick.
Not performing rebalance

Thanks,

Kevin Bridges

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users