RabbitMQ handles exit signals improperly
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
High
|
MOS Maintenance |
Bug Description
--- Environment ---
MOS Version: 7.0 + MU-3
RabitMQ: 3.5.4-1~u14.04+mos3
--- Description ---
RabbitMQ cluster broke down and Pacemaker was not able to bring it up automatically
Master/Slave Set: master_
p_rabbitmq-server (ocf::fuel:
Masters: [ cic-1.domain.tld ]
Slaves: [ cic-3.domain.tld ]
As the result, OpenStack environment became unoperable:
OpenStack CLI started to return "Service Unavailable (HTTP 503)" error
There is the following crash report in the RabbitMQ log files:
=CRASH REPORT==== 19-Apr-
crasher:
initial call: rabbit_
pid: <0.19859.1062>
registered_
exception exit: channel_
in function rabbit_
in call from rabbit_
in call from rabbit_
in call from rabbit_
in call from rabbit_
in call from rabbit_reader:run/1 (src/rabbit_
in call from rabbit_
ancestors: [<0.12202.
messages: [{'EXIT',# Port<0.
links: []
dictionary: [{{channel,1},
trap_exit: true
status: running
heap_size: 1598
stack_size: 27
reductions: 25262
neighbours:
There is the corresponding bug in RabbitMQ-server bugtracker at Github:
https:/
According to the bug status, it was fixed in RabbitMQ-server 3.6.1 version
Is it possible to get RabbitMQ 3.6.1 or this certain fix in MOS 7.0 updates?
If some additional log files are required, I can provide them directly
Changed in fuel: | |
milestone: | none → 7.0-updates |
assignee: | nobody → MOS Linux (mos-linux) |
importance: | Undecided → High |
status: | New → Confirmed |
tags: | added: area-linux |
Changed in fuel: | |
milestone: | 7.0-updates → 7.0-mu-4 |
Changed in fuel: | |
milestone: | 7.0-mu-4 → 7.0-updates |
I guess, about updates (in MOS 7.0 - updates ) it's better to ask mos-maintenaince team ( https:/ /launchpad. net/~mos- maintenance ).