[Bugs] [Bug 1511318] output of command "glsuter volume status <volume>" show sth wrong with shd status

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1511318



--- Comment #1 from zhou lin <zz.sh.cynthia at gmail.com> ---
after my investigation, i found that the when sn node execute restart, the
glusterd_restart_bricks will invode glusterd_svcs_manager,
the shd will be firstly stopped then started,
normally the stop will cause rpc disconnect event and start will cause rpc
connect event, however, the hanlder for events (glusterd_conn_common_notify)
involved the glusterd big lock, it will block the disconnect event, and only
when shd get started, it will release the big lock for a while and very often,
that release cause the "connect event" notify function get a change to execute
before the "disconect event" notify function.

there is no good way to handle this issue,
one way is to release the glusterd "big lock" for a short time, before
glusterd_shdsvc_manager trigger shd start and after glusterd_shdsvc_manager
trigger shd stop. so the "disconnect event" notify function  got a chance to
run.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.