Queues, replication, failover and disaster recovery

[an...@uk.bnpparibas.com]

Hello,

I wonder if anyone on this ML has used ActiveMQ in an environment where 
the software must failover to a disaster recovery (DR) site. It's quite a 
different thing from failing over because the primary machine has died.

I realise that AMQ does provide failover but it looks to me like it is 
designed to work soley within a site, to gaurd against machine failure. 
This is very good and for many envs this is probably all that is needed. 
However, some High Availability (HA) envs have to cope with the data 
centre being knocked out. When that happens all the machines in that 
building become unavailable and the organisation is forced to switch to 
machines at their DR site. DR sites are usually connected to the primary 
site via a WAN. Unlike the network used at the primary, the WAN will be 
operating over far greater distances and will thus have much higher 
latency. Connections may also be unreliable, e.g dropped from time to 
time. Ha anyone any experience of developing HA systems in such an 
environment? Whilst using AMQ?

Regards,

Andrew Marlow

___________________________________________________________
This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited.

Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-845C5H  for additional disclosures.

Re: Queues, replication, failover and disaster recovery

[wjmcdonald <wi...@transcentra.com>]

The only thing I've seen is 
http://activemq.apache.org/replicated-message-store.html
<http://activemq.apache.org/replicated-message-store.html>  .  By protecting
the message store using conventional DR techniques, you should have DR for
the broker network.  Bringing up brokers on the DR system may be a process
in itself - you can't use slaves because they would compete with normal
'local' fault-tolerance (M/S pair) fail-over operations - you probably don't
want your DR system to be a part of your prod system.  

Now I've been asked to design DR if I only have file-level replication and
not block-level replication... anyone have any ideas?



--
View this message in context: http://activemq.2283324.n4.nabble.com/Queues-replication-failover-and-disaster-recovery-tp2364243p4656968.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.