NATS streaming cluster and queue resilience

[siraj....@icloud.com]

Hi,

I've been reading through the documentation of NATS and NATS streaming, it seems like NATS streaming sits on top of NATS server (gnastd) which can be clustered.
However the documentation does not mention anything about queue resilience if one is using NATS streaming cluster with persistence to achieve HA. Are the queues durable to node failure ? are they mirrored or replicated ?. 

What happens to messages on a certain node in case of network partitions or node failure ? Will another node pick the workload maybe in a master/slave manner ?.

If queues are persisted to disk but not replicated or NATS does not support other nodes picking up the workload of a failed node, what happens when a dead node gets back online again and joins the cluster ? does it pick up work from where it left ?.

If anyone could help clear out that part it would be great. And give an example of a highly available setup they've done or seen across WAN or multiple availability zones.

Thanks,

Siraj

[Henrik Johansson]

I believe it is a work in progress to make streaming clustered and replicated as well.

--
You received this message because you are subscribed to the Google Groups "nats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to natsio+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Siraj....@icloud.com]

Yea, i figured that out after looking the github issues, but i couldn't find a delivery date or a roadmap, do you have any clue ?.

Thanks,

[Henrik Johansson]

I seem to remember Q1 which is very soon but no updates on progress in a while.

[Brian Flannery]

Henrik/Siraj:

Thanks for raising this question, and can understand the question on status/timing. Streaming Clustering is a very high priority deliverable currently, and design and PoC work on this is actively underway.

The Q1 delivery was something we revised in approximately Nov 2016 to Q2 17' - but I can appreciate that at the moment finding this information is a bit less centralized than we'd like.

We are in the process of updating the roadmap format and milestones, and I hope we will have this available for you in March to help you plan accordingly.

Siraj -- if you would like to discuss any high level aspects of what you are trying to do, I'd be curious to learn more and we could sync on (NATS Slack or short Skype call) - br...@nats.io is my email if you want to discuss.

B

[Siraj....@icloud.com]

Thanks for the update mate. I will get in touch and hop on the slack channel when i get the chance.

Cheers,

Siraj