[rabbitmq-discuss] Upgrading...

Wed Oct 19 21:09:33 BST 2011

James:  You should be quite safe doing that...

Jerry

----- Original Message -----
From: "James Carr" <james.r.carr at gmail.com>
To: "Jerry Kuch" <jerryk at vmware.com>
Cc: "rabbitmq-discuss" <rabbitmq-discuss at lists.rabbitmq.com>
Sent: Wednesday, October 19, 2011 12:29:41 PM
Subject: Re: [rabbitmq-discuss] Upgrading...

nevermind... I think we're just going to redeclare all the queues in
production anyway since the apps using them will be stopped for a
brief outage.Seems to be the easiest route :)

Thanks,
James

On Wed, Oct 19, 2011 at 2:09 PM, James Carr <james.r.carr at gmail.com> wrote:
> Thanks for the quick reply.
>
> All of the queues on the master node are durable.
>
> However I noticed when creating new queues from the management console
> I'm given the option to mirror the queue across all nodes or not. I'm
> assuming that the defult for existing queues from 2.5.0 is to not
> mirror them. Is there anyway to change this besides deleting and
> re-declaring the queue?
>
> Thanks,
> James
>
>
> On Wed, Oct 19, 2011 at 2:00 PM, Jerry Kuch <jerryk at vmware.com> wrote:
>> Hi, James:
>>
>> Are your queues declared durable?
>>
>> See the section "Starting and stopping nodes" for a
>> discussion of the behavior you should expect:
>>
>> http://www.rabbitmq.com/ha.html
>>
>> Jerry
>>
>> ----- Original Message -----
>> From: "James Carr" <james.r.carr at gmail.com>
>> To: "Jerry Kuch" <jerryk at vmware.com>
>> Cc: "rabbitmq-discuss" <rabbitmq-discuss at lists.rabbitmq.com>
>> Sent: Wednesday, October 19, 2011 11:57:16 AM
>> Subject: Re: [rabbitmq-discuss] Upgrading...
>>
>> So I upgraded to 2.6.1 and clustered it by adding another node to it.
>> I was experimenting with 2.6.1 because we're looking to have
>> replicated queues in our cluster since it's one thing that currently
>> kills us. :)
>>
>> I noticed that when I go to node2 and view the queues, I see the
>> queues that were created on node1. However, if I bring node1 down the
>> queues just go away.
>>
>> Is there something special I need to setup for replicated queues?
>>
>> Thanks,
>> James
>>
>> On Wed, Oct 19, 2011 at 1:39 PM, Jerry Kuch <jerryk at vmware.com> wrote:
>>> Hi, James...
>>>
>>> Ah, good to hear that you didn't lose any data!
>>>
>>> Best regards,
>>> Jerry
>>>
>>> ----- Original Message -----
>>> From: "James Carr" <james.r.carr at gmail.com>
>>> To: "Jerry Kuch" <jerryk at vmware.com>
>>> Cc: "rabbitmq-discuss" <rabbitmq-discuss at lists.rabbitmq.com>
>>> Sent: Wednesday, October 19, 2011 11:38:01 AM
>>> Subject: Re: [rabbitmq-discuss] Upgrading...
>>>
>>> I'm stupid... as part of upgrading our test environment we added a new
>>> node to the test cluster. I was viewing that node's management panel
>>> and hadn't joined it yet. :)
>>>
>>>
>>> On Wed, Oct 19, 2011 at 1:24 PM, Jerry Kuch <jerryk at vmware.com> wrote:
>>>> Rabbit's generally pretty smart (at least with upgrades between modern
>>>> versions) about migrating over and updating Mnesia contents and the
>>>> like.  What probably happened with the RPM upgrade is that the RPM
>>>> install process blew away the directories where that stuff lives.
>>>>
>>>> I think one could make note of where all that stuff is, move it aside
>>>> for safekeeping before doing the RPM upgrade/install, and then move it
>>>> back, and allow the broker to up-version the stuff at its next start up.
>>>> Rabbit's logic for doing the right thing is fairly sophisticated and I'd
>>>> think it should work in that scenario, although I've not tried it
>>>> myself.
>>>>
>>>> Somebody else on the Rabbit team with more day to day experience on that
>>>> stuff is likely able to comment more authoritatively, or at least refute
>>>> what I just said if it's obviously wrongheaded...?
>>>>
>>>> Best regards,
>>>> Jerry
>>>>
>>>> ----- Original Message -----
>>>> From: "James Carr" <james.r.carr at gmail.com>
>>>> To: "rabbitmq-discuss" <rabbitmq-discuss at lists.rabbitmq.com>
>>>> Sent: Wednesday, October 19, 2011 11:02:03 AM
>>>> Subject: [rabbitmq-discuss] Upgrading...
>>>>
>>>> Is there a nice simple way to upgrade rabbitMQ and have all of the
>>>> queues. exchanges, and users ported over?
>>>>
>>>> I just upgraded on our test servers using the noarch RPM on Oracle
>>>> Enterprise Linux and noticed that all of the users, queues, and
>>>> exchanges  are now gone.
>>>>
>>>> Thanks,
>>>> James
>>>> _______________________________________________
>>>> rabbitmq-discuss mailing list
>>>> rabbitmq-discuss at lists.rabbitmq.com
>>>> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>>>>
>>>
>>
>