[HIVE-21906] Blacklist limping nodes - ASF JIRA

XML

Word

Printable

JSON

Details

Type: New Feature
Status: Resolved
Priority: Major
Resolution: Fixed
Affects Version/s: None
Fix Version/s: None
Component/s: llap
Labels:
None

Description

If a specific LLAP node is limping it can degrade the performance of the whole cluster.

We should find a way to identify the limping nodes and act on them either by disabling the node altogether, or decreasing the load on this node

Attachments

Sub-Tasks

Add a new LlapDaemon Management API method to set the daemon capacity

Closed

Peter Vary

100%

LlapDaemon node status should be reflected in the metrics

Closed

Antal Sinkovits

100%

Publish the LLAP Daemon capacity through ZooKeeper and honor the capacity when scheduling new tasks

Closed

Antal Sinkovits

100%

Multiple target location generation in HostAffinitySplitLocationProvider

Resolved

Peter Vary

100%

Pluggable LlapMetricsListener on Tez side to disable / resize Daemons

Closed

Peter Vary

100%

Implement BlacklistingLlapMetricsListener

Closed

Peter Vary

100%

When LlapDaemon capacity is set to 0 and the waitqueue is not empty then the queries are stuck

Closed

Peter Vary

100%

Do not consider nodes with 0 capacity when calculating host affinity

Closed

Peter Vary

100%

Activity

People

Assignee:: Peter Vary

Reporter:: Peter Vary

Votes:: 0 Vote for this issue

Watchers:: 4 Start watching this issue

Dates

Created:: 21/Jun/19 15:53

Updated:: 06/Aug/19 12:00

Resolved:: 06/Aug/19 12:00

Time Tracking

Estimated:

Not Specified

Remaining:

Logged:

16h 40m

Include sub-tasks