gluster
Upgrade guides: include guidance for gfapi based applications running on remote systems
The upgrade guide does not specify how to handle gfapi based applications running on systems. I think more clarity on ordering in this case would remove some guess work.
upgrading gfapi clients and op-version
- trusted pool is updated
- gfapi clients are updated
- op-version is updated
Is the gfapi client aware of the new op-version? Does the client need to be restarted to make use of the new op-version features? What's the mechanism is in play?
upgrading within major version
General guidance here would be appreciated, even if it's as simple as "review the release notes"
3.9 (applying to 3.8 in my specific case)
If you are using gfapi based applications (qemu, NFS-Ganesha, Samba etc.) on the servers, please stop those applications too.
(Where "servers" is 'every storage server in your trusted storage pool'?)
I read this as stop gfapi on the current trusted pool member. Is this correct? Should this be more specific?
After the upgrade is complete on all servers, run the following command: #gluster volume set all cluster.op-version 30900
(where "servers" is 'every storage server in your trusted storage pool'
After reading op-version I think I know what the right thing to do is. A link might be helpful. (like 3.10)
3.10
Stop all applications that run on this server and access the volumes via gfapi (qemu, NFS-Ganesha, Samba, etc.)
Does this apply to the 3.9 guide? Should the other improvements b
NIT: number of steps is not correct.
General structure
Perhaps it would make more sense to have a general guide? There could be a links or callouts to considerations for specific versions.
