[Bugs] [Bug 1422781] Transport endpoint not connected error seen on client when glusterd is restarted

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1422781



--- Comment #12 from Jeff Darcy <jdarcy at redhat.com> ---
The key is probably "Exhausted all volfile servers" from comment 10.  This code
is reused on clients, bricks, and (I think) auxiliary daemons - i.e. everywhere
but glusterd and pure gfapi clients.  Perhaps that's a bad idea, because those
things do have different needs.  In this case, while it's semi-appropriate to
have a brick daemon give up at this point, it's just as clearly the wrong thing
to do for clients.  I've reverted the change to this behavior, which was made
as part of the multiplexing patch.  This should fix the immediate problem,
though it's hard to verify whether there might still be others because I still
haven't been able to reproduce the problem on a normal configuration.  Also,
there's a pretty high chance that the revert will re-introduce whatever problem
the change was originally made to fix.  I doubt that it was changed for no
reason, but I don't recall which specific test was failing before.  I guess
we'll find out.  ;)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=OBgtSob9AV&a=cc_unsubscribe