Any way to disable the rabbitmq river while ES is starting up?

Elastic Stack Elasticsearch
1

I have a single node, 1 shard/1 replica setup, and I'm using 0.17.8 with
the RabbitMQ river plugin. The thing is, when I restart ES, the river seems
to start consuming right away, even though the indices are still getting
allocated (I'm not sure what this means yet...). This seems to sometimes
corrupt the cluster, thought this could be due to other issues that I'm
trying to debug right now. In general, it'd be nice if I could validate
that the cluster started up properly before the river starts dumping docs
into it for indexing. Is there such a way right now?

Cheers!
Matt

2

No, there isn't a way to tell the river to wait and only start indexing
when the index it is indexing to is allocated (or at least have enough
shards started). It should not cause any corruption though..., thats
strange... . You can open an issue for it.

On Tue, Nov 8, 2011 at 9:37 AM, Matt matt.chu@gmail.com wrote:

I have a single node, 1 shard/1 replica setup, and I'm using 0.17.8 with
the RabbitMQ river plugin. The thing is, when I restart ES, the river seems
to start consuming right away, even though the indices are still getting
allocated (I'm not sure what this means yet...). This seems to sometimes
corrupt the cluster, thought this could be due to other issues that I'm
trying to debug right now. In general, it'd be nice if I could validate
that the cluster started up properly before the river starts dumping docs
into it for indexing. Is there such a way right now?

Cheers!
Matt

3

BTW, I opened a feature request to suspend a river :

David

Le 9 nov. 2011 à 11:39, Shay Banon kimchy@gmail.com a écrit :

No, there isn't a way to tell the river to wait and only start indexing when the index it is indexing to is allocated (or at least have enough shards started). It should not cause any corruption though..., thats strange... . You can open an issue for it.

On Tue, Nov 8, 2011 at 9:37 AM, Matt matt.chu@gmail.com wrote:
I have a single node, 1 shard/1 replica setup, and I'm using 0.17.8 with the RabbitMQ river plugin. The thing is, when I restart ES, the river seems to start consuming right away, even though the indices are still getting allocated (I'm not sure what this means yet...). This seems to sometimes corrupt the cluster, thought this could be due to other issues that I'm trying to debug right now. In general, it'd be nice if I could validate that the cluster started up properly before the river starts dumping docs into it for indexing. Is there such a way right now?

Cheers!
Matt