Bug 1034913 - Clustered broker fails to start on host reboot; logs critical cluster error re: exclusive queue
Summary: Clustered broker fails to start on host reboot; logs critical cluster error r...
Keywords:
Status: NEW
Alias: None
Product: Red Hat Enterprise MRG
Classification: Red Hat
Component: qpid-cpp
Version: 2.3
Hardware: Unspecified
OS: Unspecified
high
unspecified
Target Milestone: ---
: ---
Assignee: messaging-bugs
QA Contact: MRG Quality Engineering
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-11-26 17:10 UTC by Ken Giusti
Modified: 2020-11-04 22:32 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed:
Target Upstream Version:


Attachments (Terms of Use)
logs obtained by user (107.27 KB, application/zip)
2013-11-26 17:10 UTC, Ken Giusti
no flags Details

Description Ken Giusti 2013-11-26 17:10:19 UTC
Created attachment 829390 [details]
logs obtained by user

Description of problem:

qpidd is being used as part of a cluster, used with openstack, deployed with an haproxy.  One of the hosts running a qpidd cluster member is rebooted - qpidd fails to restart.

Version-Release number of selected component (if applicable):

[root@Controller01 .qpidd]# rpm -qa | grep qpid
qpid-qmf-0.18-18.el6_4.x86_64
qpid-cpp-client-0.18-17.el6_4.x86_64
qpid-cpp-server-0.18-17.el6_4.x86_64
python-qpid-qmf-0.18-18.el6_4.x86_64
qpid-cpp-server-cluster-0.18-17.el6_4.x86_64
python-qpid-0.18-5.el6_4.noarch
qpid-tools-0.18-8.el6.noarch

How reproducible:  Occasional.


Steps to Reproduce:
1.
2.
3.

Actual results:
qpidd fails to start.

Expected results:
qpidd should start and join the cluster.


Additional info:

See attached cluster logs - cluster is issuing a critical log message regarding an exclusive queue.  Openstack uses these exclusive queues as temporary reply queues.


[root@Controller01 .qpidd]# cat /etc/qpidd.conf
# GENERATED BY PUPPET
#
# Configuration file for qpidd. Entries are of the form:
#   name=value
#
# (Note: no spaces on either side of '='). Using default settings:
# "qpidd --help" or "man qpidd" for more details.
cluster-name="openstack"
#cluster-size=3
port=5672
max-connections=500
worker-threads=17
connection-backlog=10
auth=no
realm=QPID
log-enable=debug+:cluster
log-to-file=/tmp/qpidd.log

Comment 1 pgsousa 2013-11-28 12:57:35 UTC
I have 3 brokers, whenever I kill one of them, rebooting the server, It fails to start:

v 28 12:52:27 Controller03 qpidd[13540]: 2013-11-28 12:52:27 [System] error Connection 192.168.30.202:5672-192.168.30.200:60498 No protocol received closing
Nov 28 12:52:27 Controller03 qpidd[13540]: 2013-11-28 12:52:27 [System] error Connection 192.168.30.202:5672-192.168.30.200:60500 No protocol received closing
Nov 28 12:52:33 Controller03 qpidd[13540]: 2013-11-28 12:52:33 [System] error Connection 192.168.30.202:5672-192.168.30.200:60616 No protocol received closing
Nov 28 12:52:33 Controller03 qpidd[13540]: 2013-11-28 12:52:33 [System] error Connection 192.168.30.202:5672-192.168.30.200:60617 No protocol received closing

Comment 2 pgsousa 2013-11-28 16:24:39 UTC
Another note: to start the node that's not working, I have to stop the remaining nodes and start them again.

Comment 3 pgsousa 2013-11-28 17:32:52 UTC
MEssage after starting the node I rebooted:

qpidd[3334]: 2013-11-28 17:32:00 [Protocol] error Execution exception: resource-locked: Cannot subscribe to exclusive queue reply_d0f7bd5f86c8451ab219e23d242cb9f2 (qpid/broker/SessionAdapter.cpp:414)
Nov 28 17:32:00 Controller03 qpidd[3334]: 2013-11-28 17:32:00 [System] error not-found: Unknown destination 0 session=anonymous.45b850d8-2042-4d5d-b155-804f9a038ce4:0 (qpid/broker/SemanticState.cpp:580)
Nov 28 17:32:00 Controller03 qpidd[3334]: 2013-11-28 17:32:00 [HA] critical cluster(192.168.30.202:3334 UPDATEE) catch-up connection closed prematurely 192.168.30.200:5672-192.168.30.200:41163(192.168.30.202:3334-30 local,catchup)
Nov 28 17:32:00 Controller03 qpidd[3334]: 2013-11-28 17:32:00 [HA] notice cluster(192.168.30.202:3334 LEFT) leaving cluster openstack


Note You need to log in before you can comment on or make changes to this bug.