Skip to content

Instantly share code, notes, and snippets.

@dbenson
Created February 24, 2011 18:44
Show Gist options
  • Star 2 You must be signed in to star a gist
  • Fork 3 You must be signed in to fork a gist
  • Save dbenson/842638 to your computer and use it in GitHub Desktop.
Save dbenson/842638 to your computer and use it in GitHub Desktop.
elasticsearch.yml
name: ${HOSTNAME}
# The cluster should be named with the name of the AD domain
cluster:
name: bldrprod-0.14.0
routing:
allocation:
# As small as possible, very large performance hits during recovery
concurrent_recoveries: 1
# Force all memory to be prealloced and locked, forcing JVM to never swap
bootstrap:
mlockall: true
gateway:
type: local
# We will only start the cluster with n-1 nodes available. We wait 5 mins for all N to be available
# If the number of nodes in the cluster changes, we MUST update this value
recover_after_nodes: 3
recover_after_time: 5m
expected_nodes: 4
discovery:
zen:
ping:
unicast:
hosts: ['dm-essearchp101.bldrprod.local:9300', 'dm-essearchp102.bldrprod.local:9300', 'dm-essearchp103.bldrprod.local:9300', 'dm-essearchp104.bldrprod.local:9300']
index:
# This combined with concurrent_recoveries throttles recovery to have the least impact on performance
# This does mean that recovery may be quite slow, but that is the sacrifice we're making to avoid bad perf
shard:
recovery:
concurrent_streams: 1
# All of our snapshotting occurs via esindexer
gateway:
snapshot_interval : -1
# We have so many shards that we need to keep translogs relatively small for recovery
# This does result in more frequent flushes. The default is 5000 and this is per index
translog:
flush_threshold: 1000
# Defaults to 1024, but that is to small for very large portfolio pages
query:
bool:
max_clause_count: 10240
analysis :
analyzer :
# This analyzer mimics verity behavior
verity_tokenizer :
type: pattern
lowercase: true
pattern: '(?:(?!\w).)+'
stopwords: _none_
flags: DOTALL
# Analyzer is used for sorting multi-term fields
lowercase_keyword :
type : custom
filter : [lowercase]
tokenizer : keyword
# IDOL TextMatch, PipeDelimited
pipe_tokenizer :
type: pattern
lowercase: true
pattern: '\|'
stopwords: _none_
flags: DOTALL
# IDOL FullText
fulltext_tokenizer :
type : custom
filter : [standard, lowercase, stop, porterStem]
tokenizer : standard
path:
logs: /opt/wsod/var/log/elasticsearch/
work: /opt/wsod/var/work/elasticsearch/
data: /opt/wsod/var/data/elasticsearch/
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment