The current sysvinit/upstart/systemd script method of restarting (or not restarting) glusterfsd instances (bricks) produces very high load if successful, or doesn't restart bricks at all leaving them running older versions.
I propose a new "system" command:
gluster system restart [target_load=2] [nfs=false]
It would gracefully kill single bricks, and re-spawn them, waiting to proceed to the next one until the system load fell below target_load. It would also kill and restart the self-heal daemon and if nfs=true, the nfs daemon. Once it's finished all of those, can it reload itself?
because of the large number of bugs filed against mainline version\ is ambiguous and about to be removed as a choice.
If you believe this is still a bug, please change the status back to NEW and choose the appropriate, applicable version for it.
Feature requests are only valid against the 'mainline' version. Once patches have been posted, it might be possible to get them backported, and at that time you can request a backport for a certain current version (clone this bug, or ask on irc/mailinglist).
Migrated to github:\n\nhttps://github.com/gluster/glusterfs/issues/586\n\nPlease follow the github issue for further updates on this bug.
Migrated to github:
Please follow the github issue for further updates on this bug.