[Gluster-users] Replacing a third data node with an arbiter one

[Hoggins!]

Thank you, for that, however I have a problem.

Le 26/01/2018 à 02:35, Ravishankar N a écrit :
> Yes, you would need to reduce it to replica 2 and then convert it to
> arbiter.
> 1. Ensure there are no pending heals, i.e. heal info shows zero entries.
> 2. gluster volume remove-brick thedude replica 2
> ngluster-3.network.hoggins.fr:/export/brick/thedude force
> 3. gluster volume add-brick thedude replica 3 arbiter 1 <IP:brick path
> of the new arbiter brick>

Removing the current third brick was OK with
    volume remove-brick thedude replica 2
ngluster-3.network.hoggins.fr:/export/brick/thedude force

But while adding the arbiter brick with
    volume add-brick thedude replica 3 arbiter 1
arbiter-1.network.hoggins.fr:/gluster/thedude force

... I got this :
    "volume add-brick: failed: Commit failed on
arbiter-1.network.hoggins.fr. Please check log file for details."

On arbiter-1 brick, the log says :

    [2018-01-29 15:15:52.999698] I [run.c:190:runner_log]
    (-->/usr/lib64/glusterfs/3.12.5/xlator/mgmt/glusterd.so(+0x3744b)
    [0x7fcd49fef44b]
    -->/usr/lib64/glusterfs/3.12.5/xlator/mgmt/glusterd.so(+0xd252c)
    [0x7fcd4a08a52c] -->/lib64/libglusterfs.so.0(runner_log+0x105)
    [0x7fcd4f48d0b5] ) 0-management: Ran script:
    /var/lib/glusterd/hooks/1/add-brick/pre/S28Quota-enable-root-xattr-heal.sh
    --volname=thedude --version=1 --volume-op=add-brick
    --gd-workdir=/var/lib/glusterd
    [2018-01-29 15:15:52.999816] I [MSGID: 106578]
    [glusterd-brick-ops.c:1354:glusterd_op_perform_add_bricks]
    0-management: replica-count is set 3
    [2018-01-29 15:15:52.999849] I [MSGID: 106578]
    [glusterd-brick-ops.c:1359:glusterd_op_perform_add_bricks]
    0-management: arbiter-count is set 1
    [2018-01-29 15:15:52.999862] I [MSGID: 106578]
    [glusterd-brick-ops.c:1364:glusterd_op_perform_add_bricks]
    0-management: type is set 0, need to change it
    [2018-01-29 15:15:55.140751] I
    [glusterd-utils.c:5941:glusterd_brick_start] 0-management: starting
    a fresh brick process for brick /gluster/thedude
    [2018-01-29 15:15:55.194678] E [MSGID: 106005]
    [glusterd-utils.c:5947:glusterd_brick_start] 0-management: Unable to
    start brick arbiter-1.network.hoggins.fr:/gluster/thedude
    [2018-01-29 15:15:55.194823] E [MSGID: 106074]
    [glusterd-brick-ops.c:2590:glusterd_op_add_brick] 0-glusterd: Unable
    to add bricks
    [2018-01-29 15:15:55.194854] E [MSGID: 106123]
    [glusterd-mgmt.c:312:gd_mgmt_v3_commit_fn] 0-management: Add-brick
    commit failed.
    [2018-01-29 15:15:55.194868] E [MSGID: 106123]
    [glusterd-mgmt-handler.c:603:glusterd_handle_commit_fn]
    0-management: commit failed on operation Add brick


However, when I query again :

    gluster> volume info thedude

    Volume Name: thedude
    Type: Replicate
    Volume ID: bc68dfd3-94e2-4126-b04d-77b51ec6f27e
    Status: Started
    Snapshot Count: 0
    Number of Bricks: 1 x (2 + 1) = 3
    Transport-type: tcp
    Bricks:
    Brick1: ngluster-1.network.hoggins.fr:/export/brick/thedude
    Brick2: ngluster-2.network.hoggins.fr:/export/brick/thedude
    Brick3: arbiter-1.network.hoggins.fr:/gluster/thedude (arbiter)
    Options Reconfigured:
    cluster.server-quorum-type: server
    transport.address-family: inet
    nfs.disable: on
    performance.readdir-ahead: on
    client.event-threads: 8
    server.event-threads: 15


... I can see that the arbiter has been taken into account.

So is it, or is it not ? How to ensure that ?

Thanks !

    Hoggins!


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: OpenPGP digital signature
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20180129/f8bfbd51/attachment.sig>