logsearch/201.0.2
You can find the source of this version on GitHub at logsearch/logsearch-boshrelease. It was created based on the commit 270bfa5e
.
Release Notes¶
v201.0.2 RELEASE NOTES
ElasticSearch: 2.2.0
LogStash: 2.2.0
Kibana: 4.4.0
Notable job spec changes
Haproxy
haproxy.inbound_port => haproxy.ingestor.inbound_port
haproxy.backend_servers => haproxy.ingestor.backend_servers
haproxy.backend_port => haproxy.ingestor.backend_port
ElasticSearch
deleted elasticsearch.drain
If you wish to deploy without executing the drain script you should do the following:
$ bosh deploy --skip-drain elasticsearch_data
New features
New cluster update process
The wrapping scripts around stopping and starting ElasticSearch data nodes during a deployment upgrade have been significantly changed. There are some caveats to this strategy, see Known limitations for more infomation.
NOTE: While we are waiting for a new feature to be added to BOSH a manual step is required to finish a deployment:
$ bosh deployment logsearch-manifest.yml
$ bosh run errand enable_shard_allocation
To take advantage of this errand, make sure to regenerate your manifest with the scripts/generate_deployment_manifest
script.
Reverse proxying to kibana
The haproxy job can now be configured to act as a reverse proxy to kibana allowing basic auth to be enabled.
jobs:
- name: haproxy
instances: 1
properties:
haproxy:
kibana:
inbound_port: 5601
backend_servers: [<kibana_internal_ips>]
backend_port: 5601
auth:
username: admin
password: password
Kibana plugins
It is possible to specify plugins to be installed for kibana
jobs:
- name: kibana
properties:
kibana:
plugins:
someplugin: /var/vcap/packages/co-located-kibana-plugin/plugin.zip
someotherplugin: elasticsearch/marvel/latest
anotherplugin: https://some.sample.url/directory
Kibana environment variables
jobs:
- name: kibana
properties:
kibana:
env:
SOMEVARIABLE: AWESOME_VALUE
Known limitations
Cluster update process
In order to guarantee a safe upgrade without risk of losing data, we disable shard allocation for the duration of the deployment process. New indices cannot be created during this time. It can affect log ingestion during deploy if more elaborate index patterns are used in the deployment such as index per log type or hourly indexes. Logs are not lost though, they will wait in the queue until the cluster resumes normal operation.
WARNING: If you are running with more than 1 replica / index, your cluster might also become red during updates.
Usage¶
You can reference this release in your deployment manifest from the releases
section:
- name: "logsearch" version: "201.0.2" url: "https://bosh.io/d/github.com/logsearch/logsearch-boshrelease?v=201.0.2" sha1: "713058d7e38804c6fc76f48e34ea64a611916f50"
Or upload it to your director with the upload-release
command:
bosh upload-release --sha1 713058d7e38804c6fc76f48e34ea64a611916f50 \ "https://bosh.io/d/github.com/logsearch/logsearch-boshrelease?v=201.0.2"
Jobs¶
- archiver
- curator
- elasticsearch
- elasticsearch_config
- enable_shard_allocation
- haproxy
- ingestor_archiver
- ingestor_relp
- ingestor_syslog
- kibana
- nats_to_syslog
- parser
- queue