Heartbit doesn't delete resources

[netDalek]

I probably just can't understand the rigth behaviour, but I don't know the better way to ask. May be this question will help smb.

I start two nodes with resource_discovery application (with heartbit). Run on the first:

resource_discovery:add_local_resource_tuple({worker, a}).
resource_discovery:sync_resources().

And run on the second:

resource_discovery:add_target_resource_types([worker]).
resource_discovery:sync_resources().
resource_discovery:get_resources(worker). #[a]

And get the resource. It's OK. Then I kill first node. But I still get registered resource, calling resource_discovery:get_resources(worker). Calling sync_resources and contract_nodes doesn't help. Should it disappeared? Where is my mistake?

[ericbmerritt]

@martinjlogan thoughts

[martinjlogan]

It should not. Heartbeat is not a one sized fits all fit for understanding
if a resource is there or not. All the heartbeat does is rejoin a cluster
if it gets separated for some reason. Honestly, at this point I question
whether hearbeat is even a good feature or something that should be
implemented at the application level or at least a higher level than RD. I
am open to dissenting ideas though.

On Wed, Apr 1, 2015 at 6:17 PM, Eric Merritt notifications@github.com
wrote:

@martinjlogan https://github.com/martinjlogan thoughts

—
Reply to this email directly or view it on GitHub
#12 (comment)
.

Martin Logan
Erlang & OTP in Action (Manning) http://manning.com/logan
http://twitter.com/martinjlogan
http://erlware.org