We rely on curl to query the elasticsearch api for cluster state,
deleting indexes and so on. Ensure it is installed on the elasticsearch
nodes.
Change-Id: I843c6be65e5750856cb0b68f68eb7fac47e04e29
The package maintainer scripts start elasticsearch when it gets
installed, but it seems to return control before the pidfile is
created which allows puppet to race for it and ultimately launch a
second daemon resulting in an OOM condition. It's configured to
start at boot anyway, so the puppet ensure running isn't buying us
much.
Change-Id: I1b374f90c198b9cb73d412fe63b439623f434d9a
* modules/elasticsearch/manifests/init.pp: Brown bag fix... the
package installation used to make this directory, but now we want to
put files in it prior to package installation.
Change-Id: Id3069999d59f965226bcae8f21ef7a93e4d44c72
* modules/elasticsearch/manifests/init.pp: The config files need to
be in place before package installation starts the service, and they
don't depend on the package itself anyway.
Change-Id: Iabc57e3091533eb5168c670f5c29d0d2dac9a1a2
wiki.o.o is much smaller than our typical elasticsearch nodes. For it to
use elasticsearch it needs to allocate a much smaller heap. Make the
heap size configurable (keep the 16g default) and set the heap size to
1g for wiki.o.o.
Change-Id: I0c5fb4865eb07c0fa5bbe4cf698a9f0e9ea9f2fc
To allow for reuse of elasticsearch I'm splitting it away from
logstash. Also, I'm doing a hash check of the elasticsearch wget
for better security.
Change-Id: Iff42d538cd941abd50b000879ea4a237ea48d40e
