Updated on 2022-08-16 GMT+08:00

HBase Dual-Read Configuration Items

This section provides the details of all the configurations required for the HBase dual-read feature.

HBase Dual-Read Operations

Table 1 Configuration items in hbase-dual.xml

Configuration Item

Description

Default Value

Level

hbase.dualclient.active.cluster.configuration.path

HBase client configuration directory of the active cluster

None

Mandatory

hbase.dualclient.standby.cluster.configuration.path

HBase client configuration directory of the standby cluster

None

Mandatory

dual.client.schedule.update.table.delay.second

DR table update interval

5

Optional

hbase.dualclient.glitchtimeout.ms

Maximum glitch time can be tolerated in the active cluster

50

Optional

hbase.dualclient.slow.query.timeout.ms

Slow query alarm log

180000

Optional

hbase.dualclient.active.cluster.id

Active cluster ID

ACTIVE

Optional

hbase.dualclient.standby.cluster.id

Standby cluster ID

STANDBY

Optional

hbase.dualclient.active.executor.thread.max

Maximum size of the thread pool for processing requests to the active cluster

100

Optional

hbase.dualclient.active.executor.thread.core

Core size of the thread pool for processing requests to the active cluster

100

Optional

hbase.dualclient.active.executor.queue

Queue size of the thread pool for processing requests to the active cluster

256

Optional

hbase.dualclient.standby.executor.thread.max

Maximum size of the thread pool for processing requests to the standby cluster

100

Optional

hbase.dualclient.standby.executor.thread.core

Core size of the thread pool for processing requests to the standby cluster

100

Optional

hbase.dualclient.standby.executor.queue

Queue size of the thread pool for processing requests to the standby cluster

256

Optional

hbase.dualclient.clear.executor.thread.max

Maximum size of the thread pool for clearing resources

30

Optional

hbase.dualclient.clear.executor.thread.core

Core size of the thread pool for clearing resources

30

Optional

hbase.dualclient.clear.executor.queue

Queue size of the thread pool for clearing resources

Integer. MAX_VALUE

Optional

dual.client.metrics.enable

Whether to print client metric information

true

Optional

dual.client.schedule.metrics.second

Interval for printing client metric information

300

Optional

dual.client.asynchronous.enable

Whether to asynchronously request the active and standby clusters

false

Optional

Printing Metric Information

Table 2 Basic specifications

Metric Name

Description

Log level

total_request_count

Total number of queries in a period

INFO

active_success_count

Number of successful queries in the active cluster in a period

INFO

active_error_count

Number of failed queries in the active cluster in a period

INFO

active_timeout_count

Number of query timeouts in the active cluster in a period

INFO

standby_success_count

Number of successful queries in the standby cluster in a period

INFO

standby_error_count

Number of failed queries in the standby cluster in a period

INFO

Active Thread pool

Periodically printed information about the thread pool for processing requests to the active cluster

DEBUG

Standby Thread pool

Periodically printed information about the thread pool for processing requests to the standby cluster

DEBUG

Clear Thread pool

Periodically printed information about the thread pool for releasing resources

DEBUG

Table 3 Histogram indicators for GET, BatchGET, and SCAN requests

Metric Name

Description

Log level

averageLatency(ms)

Average latency

INFO

minLatency(ms)

Minimum latency

INFO

maxLatency(ms)

Maximum latency

INFO

95thPercentileLatency(ms)

Maximum latency of 95% requests

INFO

99thPercentileLatency(ms)

Maximum latency of 99% requests

INFO

99.9PercentileLatency(ms)

Maximum latency of 99.9% requests

INFO

99.99PercentileLatency(ms)

Maximum latency of 99.99% requests

INFO