Published on 2019-02-14 00:00:00

Humio 1.4.0

High Availability for Ingest and Digest

Version

Type

Release Date

End of Support

Upgrades From

Data Migration

Config. Changes

1.4.0

Archive

2019-02-14

2020-02-14

1.3.2

No

No

JAR Checksum

Value

MD5

fb0290d5203f178cfbbef8df7b89106a

SHA1

7e2f17d867734264c91c697849884bb530fbc450

SHA256

cce65b639ab277dd50cf29f2d53ff119d705c64d33a4a118b8e49b899d8dd27c

SHA512

7cfad54614ef63f35fe6d3cc50d2239650845b390e0aea63f1e2e199745a68854524fdb24cad5f59867339cbe0eb2432bd904cef7e7a90e9b115d5abe6f1a52a

TGZ Checksum

Value

MD5

fb0290d5203f178cfbbef8df7b89106a

SHA1

7e2f17d867734264c91c697849884bb530fbc450

SHA256

cce65b639ab277dd50cf29f2d53ff119d705c64d33a4a118b8e49b899d8dd27c

SHA512

7cfad54614ef63f35fe6d3cc50d2239650845b390e0aea63f1e2e199745a68854524fdb24cad5f59867339cbe0eb2432bd904cef7e7a90e9b115d5abe6f1a52a

Summary

High availability for ingest and digest.

Change Log

  • If rolling back, make sure to roll back to version 1.3.2+

  • Digest partitions can now be assigned to more than one host. Doing so enables the cluster to continue digesting incoming events if a single host is lost from the cluster.

  • Emphasis is on efficiency during normal operation over being efficient in the failure cases: After failure the cluster will need some time to recover during which ingested events will get delayed. The cluster needs to have ample cpu to catch up after such a fail-over. There are both new and reinterpreted configuration options in the config environment for controlling how the segments get build for this.

  • Segments are flushed after 30 minutes. This makes S3 archiving likely to be less than 40 minutes after the incoming stream.

  • Limit match/lookup functions to 20.000 rows or whatever “MAX_STATE_LIMIT” is set to.

  • Bug Fix: Clone existing dashboard when creating from the frontpage was broken.