Configuration properties for Presto (Java) - coordinator and worker nodes
You can customize the coordinator and worker configuration properties through an API for Presto (Java).
| Property name | Type | Validation added |
|---|---|---|
experimental.optimized-repartitioning |
Boolean | True or False |
experimental.reserved-pool-enabled |
Boolean | True or False |
fragment-result-cache.enabled |
Boolean | True or False |
fragment-result-cache.max-cached-entries |
Integer | 1000000 |
fragment-result-cache.base-directory |
String | file:///mnt/tmpfs/fragment |
fragment-result-cache.cache-ttl |
String | 24h |
heap_dump_on_exceeded_memory_limit.enabled |
Boolean | True or False |
heap_dump_on_exceeded_memory_limit.file_directory |
String | Any string |
heap_dump_on_exceeded_memory_limit.max.number |
Integer | Limit {1, 1000} |
heap_dump_on_exceeded_memory_limit.max.size |
Integer | Limit {1, 1000} |
memory.heap-headroom-per-node |
String | Limit {1, 1e9}; supported values are numbers with or without units TB, MB, GB, B, KB |
node-scheduler.include-coordinator |
Boolean | True or False |
query.execution-policy |
String | Any string |
query.low-memory-killer.policy |
String | Any string |
query.max-memory |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
query.max-memory-per-node |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
query.max-stage-count |
Integer | Limit {1, 1000} |
query.max-total-memory-per-node |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
query.min-expire-age |
String | Limit {1, 1e13}; supported values are numbers with or without units m, s, ms, h |
query.stage-count-warning-threshold |
Integer | Limit {1, 1000} |
task.concurrency |
Integer | Limit {1, 1000} |
task.max-drivers-per-task |
Integer | Limit {1, 100} |
join-distribution-type |
String | Value should be automatic or broadcast or partitioned |
exchange.client-threads |
Integer | Limit {1, 1000} |
exchange.http-client.max-connections |
Integer | Limit {1, 10000} |
exchange.http-client.max-connections-per-server |
Integer | Limit {1, 100000} |
http-server.log.max-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
http-server.log.max-history |
Integer | Limit {1, 100} |
http-server.threads.max |
Integer | Limit {1, 1000} |
join-max-broadcast-table-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
log.max-history |
Integer | Limit {1, 100} |
log.max-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
node-scheduler.max-pending-splits-per-task |
Integer | Limit {1, 3000} |
node-scheduler.max-splits-per-node |
Integer | Limit {1, 3000} |
optimize-nulls-in-join |
Boolean | True or False |
optimizer.default-filter-factor-enabled |
Boolean | True or False |
optimizer.exploit-constraints |
Boolean | True or False |
optimizer.prefer-partial-aggregation |
Boolean | True or False |
query.client.timeout |
String | Limit {1, 1e13}; supported values are numbers with or without units m, s, ms, h |
query.max-execution-time |
String | Limit {1, 1e13}; supported values are numbers with or without units m, s, ms, h |
query.max-history |
Integer | Limit {1, 100} |
query.max-total-memory |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB |
query.min-schedule-split-batch-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
query.max-length |
Integer | Limit {1, 1000000} |
scale-writers |
Boolean | True or False |
shutdown.grace-period |
String | Limit {1, 1e13}; supported values are numbers with or without units m, s, ms, h |
sink.max-buffer-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
experimental.max-revocable-memory-per-node |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
experimental.max-spill-per-node |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
experimental.pushdown-dereference-enabled |
Boolean | True or False |
experimental.pushdown-subfields-enabled |
Boolean | True or False |
experimental.query-max-spill-per-node |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
experimental.spiller-max-used-space-threshold |
Float | Float 64 |
experimental.spiller-spill-path |
String | Any string |
http-server.max-request-header-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
experimental.internal-communication.max-task-update-size |
String | Limit {1, 1e13}; supported values are numbers with or without units TB, MB, GB, B, KB |
To enable fragment cache, you must set up a cache-enabled environment. To set up, patch the following properties into /opt/presto/etc/config.properties.
fragment-result-cache.enabled: true
fragment-result-cache.base-directory: file:///mnt/tmpfs/fragment
fragment-result-cache.max-cached-entries: 1000000
fragment-result-cache.cache-ttl: 24h
The mount path /mnt/tmpfs/fragment is by default available on a cache-enabled environment with the system's volume mounted on it.