[Bugs] [Bug 1462447] brick maintenance - no client reconnect

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1462447



--- Comment #1 from Markus Stockhausen <mst at collogia.de> ---
Status on storage side is:

gluster volume status gluster1 clients| grep -i -E "brick|conn"
Client connections for volume gluster1
Brick : cfilers101:/var/data/brick1/brick
Clients connected : 10
Brick : cfilers201:/var/data/brick2/brick <- was in maintenance
Clients connected : 7
Brick : cfilers201:/var/data/brick1/brick <- was in maintenance
Clients connected : 9
Brick : cfilers301:/var/data/brick2/brick
Clients connected : 10
Brick : cfilers301:/var/data/brick1/brick
Clients connected : 10
Brick : cfilers101:/var/data/brick2/brick
Clients connected : 10

We observed two symptoms:

First: Clients try to reconnect, flood the gluster logs, but cannot reconnect
to the restarted storage node. More see here:
http://lists.gluster.org/pipermail/gluster-users/2017-June/031525.html

Second: Clients show no signs of reconnect. No new entries in their logs. 

Log of one client attached.

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