[rabbitmq-discuss] Question about HA configuration

• Previous message: [rabbitmq-discuss] Message acknowledgment scenarios
• Next message: [rabbitmq-discuss] Question about HA configuration
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi 

Please consider this scenario:
- Active/Passive HA configuration for MQ servers (Pacemaker+corosync+DRBD)
- NOT using persistent queues: so every time there is a failure all messages in the queues are lost

Does it make sense to have a shared backend storage, for example using DRDB? Which is the value of the shared storage if messages are not stored on the file system?
Maybe it's useful because in the shared folder are stored information about queues?!

Regards
--
Andrea Rosa

• Previous message: [rabbitmq-discuss] Message acknowledgment scenarios
• Next message: [rabbitmq-discuss] Question about HA configuration
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]