It would be a useful feature if "gluster volume status $vol clients" could report the version number of the clients and "gluster volume status version" could report the version of the servers. This would help ensure that the clients and servers are all at the expected version for change audits after upgrading.
Feature requests make most sense against the 'mainline' release, there is no ETA for an implementation and requests might get forgotten when filed against a particular version.
The first part of the request is now addressed through http://review.gluster.org/16303 and will be available as part of 3.10. On gluster v status version, we already have a cluster.op-version which indicates what op-version cluster is running with. On top of it we also introduced cluster.max-op-version (http://review.gluster.org/16283 ) which will indicate the version at which cluster can get bumped up to. IMHO, individual server options do not make much sense as GlusterD does all the client compatibility checks based on cluster.op-version. Based on the above I am moving this bug to MODIFIED.
The sense of per server versions is that on several occasions I've encountered users that have installed multiple servers thinking they had installed them all from the ppa when, in fact, they had missed adding the ppa to one server. Even with cluster.max-op-version all we would know is that it's lower than expected. Showing the version for each server would easily show why that is.
(In reply to Joe Julian from comment #3) > The sense of per server versions is that on several occasions I've > encountered users that have installed multiple servers thinking they had > installed them all from the ppa when, in fact, they had missed adding the > ppa to one server. Even with cluster.max-op-version all we would know is > that it's lower than expected. Showing the version for each server would > easily show why that is. One can always use `gluster --version` in that case, no?
Yes, but that requires Nx operations vs one operation. Additionally, adding this feature would show servers that haven't restarted since upgrading and would just generally expose version problems faster and easier.