0005422: LB should mark node as unavailable on 503 for a configurable period of time.

Viewing Issue Advanced Details [ Jump to Notes ]

[ View Simple ] [ Issue History ] [ Print ]

Category

Severity

Reproducibility

Date Submitted

Last Update

0005422

[Resin]

minor

always

04-16-13 16:43

08-05-13 10:59

Reporter

alex

View Status

public

Assigned To

ferg

Priority

normal

Resolution

fixed

Platform

Status

closed

Projection

none

OS Version

ETA

none

Fixed in Version

4.0.36

Product Version

4.0.35

Product Build

Summary

0005422: LB should mark node as unavailable on 503 for a configurable period of time.

Description

rep by Ethan Larson
To review, here's the behavior we would expect to see:

Context: starting node.

1. A node in the cluster goes down.
2. The load balancer routes a request to it, and it times out after 1 second. The LB then re-routes the request to another server in the cluster.
3. The LB marks the node as down for 1 minute and does not send any more traffic to it within that time period.
4. After 1 minute, the LB again attempts to route traffic. If the node is available, great. If not, another 1-second timeout occurs and the process repeats.
5. At no time does the LB route traffic to a node that is known to be unavailable, or respond to a request with a 500/503 when there are nodes available.

Steps To Reproduce

Additional Information

Attached Files

Relationships

Notes
(0006303) ferg 08-05-13 10:59	server/26t2

Issue History
Date Modified	Username	Field	Change
04-16-13 16:43	alex	New Issue
04-16-13 17:36	eplarson	Issue Monitored: eplarson
08-05-13 10:59	ferg	Note Added: 0006303
08-05-13 10:59	ferg	Assigned To	=> ferg
08-05-13 10:59	ferg	Status	new => closed
08-05-13 10:59	ferg	Resolution	open => fixed
08-05-13 10:59	ferg	Fixed in Version	=> 4.0.36

Mantis 1.0.0rc3[^]

31 total queries executed.
27 unique queries executed.