[rabbitmq-discuss] Stale/Stuck queue on cluster
hkraemer at anynines.com
Thu Mar 6 10:21:55 GMT 2014
I've a strange problem in case HA queues.
* RabbitMQ 3.2.3-1 (Ubuntu 12.04)
* 4 node cluster (2x RAM, 2x Disc)
* HA-Queues, durable false
* Management plugins activated
I've more then 200 queues in my vhost and all works fine expect for one.
I don't see this queue via
but i see on the webinterface (management plugin) this queue is stored
on one Ram node. I can't sync this queue, I can't get informations about
this queue i see only on the panel '? ? ?'.
If i try to delete this queue via CLI => stuck
If i try to sync this queue via CLI => stuck
if i try to get messages via CLI => stuck
In addition on this node where the queue is present (what the web
interface shows) can't be stopped (service rabbitmq-server stop)
What i tried:
* via CLI the commands above
* kill erlang processes on this node, kill rabbitmq-server delete,
delete this node from cluster, delete mnesia home dir, start clean, add
to cluster again and resync without success
I don't see anything in the logfiles:
* No errors in sasl log
* No errors in shutdown logs
The only one what i see is:
=ERROR REPORT==== 5-Mar-2014::16:07:29 ===
webmachine error: path="/api/queues/%2F/lbaas_process_on_host_agent"
=ERROR REPORT==== 6-Mar-2014::10:17:27 ===
<0.786.0> to <0.309.0> in an old incarnation (1) of this node (2)
but that's impossible, the user has admin privileges and all permissions:
xxxx .* .* .*
on the right vhost.
Do anyone have a hint ? I need to delete this queue without to build a
complete new cluster. Is that possible ?
Cheers and thanks
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 555 bytes
Desc: OpenPGP digital signature
More information about the rabbitmq-discuss