On Fri, Sep 21, 2018 at 12:44 AM Chaloulos, Klearchos (Nokia - GR/Athens) <klearchos.chaloulos@xxxxxxxxx> wrote:
Hello,We are using glusterfs version 3.7.14, and have deployed the glusterfs servers in containers. We are trying to use the “gluster volume add-brick” command to extend a volume, but it fails:gluster volume add-brick oam replica 3 172.01.01.01:/mnt/bricks/oam forcevolume add-brick: failed: Commit failed on 172.01.01.01. Please check log file for details.The logs show that it fails to mount /dev/fuse:[mount.c:341:gf_fuse_mount] 0-glusterfs-fuse: cannot open \/dev\/fuse (No such file or directory)[glusterd-utils.c:11294:glusterd_handle_replicate_brick_ops] 0-management: mount command failed.[MSGID: 106074] [glusterd-brick-ops.c:2372:glusterd_op_add_brick] 0-glusterd: Unable to add bricks[MSGID: 106123] [glusterd-mgmt.c:294:gd_mgmt_v3_commit_fn] 0-management: Add-brick commit failed.Questions:
- Why does it need /dev/fuse in order to add a brick?
/dev/fuse or the fuse kernel module is needed on a server that contains a brick as several features in gluster rely on mounting a native(fuse) client locally on the server.
- Is there a way to bypass the need for /dev/fuse in order to add the extra brick?
I don't think there is an easy way at the moment. Making the fuse kernel module available to the glusterfs installation in the container would be the preferred route as of now.
- Is this related to this issue: https://bugzilla.redhat.com/show_bug.cgi?id=1420027 ?
Yes, this is related to the issue reported in the bug. There is a fair amount of work needed to avoid dependency of the glusterfs server stack on a fuse mount.
Best Regards,
Vijay
_______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx https://lists.gluster.org/mailman/listinfo/gluster-devel