Re: gluster peer probe newnode fails for new node after upgrade 3.2.7 > 3.4 on Ubuntu

[Xset Mode <xsetmode@xxxxxxxxx>]

Someone has actually just done an upgrade on an existing cluster, from glusterfs-3.4.0-3.el6.x86_64 to glusterfs-3.4.0-8.el6.x86_64
that caused the operating version to change on /var/lib/glusterd/glusterd.info operatiog-version=1 to operating-version=2

But that was RHEL based distribution

I'm running Ubuntu 13.04 using the packages from

deb http://ppa.launchpad.net/semiosis/ubuntu-glusterfs-3.4/ubuntu raring main


On my old cluster nodes I have same version of packages installed, but operating-version=1 on /var/lib/glusterd/glusterd.info file

Can I manually edit  operating-version=2 on /var/lib/glusterd/glusterd.info file & restart glusterd on all old servers and then try
gluster peer probe storage8 (storage8 is the newest node without /var/lib/glusterd/glusterd.info file)

---

From: Xset Mode <xsetmode@xxxxxxxxx>
To: "gluster-devel@xxxxxxxxxx" <gluster-devel@xxxxxxxxxx>
Sent: Monday, August 19, 2013 5:35 PM
Subject: Re: gluster peer probe newnode fails for new node after upgrade 3.2.7 > 3.4 on Ubuntu

Also in old cluster node logs I see this:

[2013-08-19 13:54:24.560339] E [glusterd-handshake.c:900:__glusterd_mgmt_hndsk_version_cbk] 0-management: failed to validate the operating version of peer (storage8)

---

From: Xset Mode <xsetmode@xxxxxxxxx>
To: "gluster-devel@xxxxxxxxxx" <gluster-devel@xxxxxxxxxx>
Sent: Monday, August 19, 2013 5:02 PM
Subject: gluster peer probe newnode fails for new node after upgrade 3.2.7 > 3.4 on Ubuntu

I followed following instructions to upgrade existing gluster cluster 3.2.7 to 3.4 gluster on Ubuntu 13.04

http://vbellur.wordpress.com/2012/05/31/upgrading-to-glusterfs-3-3/

I stopped all glusterd and client processes first.

I had to copy files from /etc/glusterd/ to /var/lib/glusterd as the location has changed in 3.4

I also had to do the following to be able to start the volume after upgrade:

vol=datavol; brick=/home/glusterfs; setfattr -n trusted.glusterfs.volume-id -v 0x$(grep volume-id /var/lib/glusterd/vols/$vol/info | cut -d= -f2 | sed 's/-//g') $brick

All old nodes were upgraded to packages from the repository

ii  glusterfs-client                    3.4.0final-ubuntu1~raring1       amd64        clustered file-system (client package)
ii  glusterfs-common                    3.4.0final-ubuntu1~raring1       amd64        GlusterFS common libraries and translator modules
ii  glusterfs-server                    3.4.0final-ubuntu1~raring1       amd64        clustered file-system (server package)

Now when I try to add a new fresh Ubuntu 13.04 node with same packages

ii  glusterfs-client                    3.4.0final-ubuntu1~raring1       amd64        clustered file-system (client package)
ii  glusterfs-common                    3.4.0final-ubuntu1~raring1       amd64        GlusterFS common libraries and translator modules
ii  glusterfs-server                    3.4.0final-ubuntu1~raring1       amd64        clustered file-system (server package)

From existing nodes trying to do:

# gluster peer probe storage8
peer probe: failed: Peer storage8 is already at a higher op-version

How can I get the op-version of existing cluster nodes to the same op-version as the new fresh node so that gluster peer probe storage8 works from

existing nodes?

From existing nodes I see operating-version=1

Old nodes:

# /var/lib/glusterd/glusterd.info
UUID=ed2b70f4-b0b1-4929-a55a-2f49f4f9580f
operating-version=1

Any help would be appreciated?

_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
https://lists.nongnu.org/mailman/listinfo/gluster-devel

_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
https://lists.nongnu.org/mailman/listinfo/gluster-devel