0001766: Load-Balancer detection of frozen server

Viewing Issue Simple Details [ Jump to Notes ]

[ View Advanced ] [ Issue History ] [ Print ]

Category

Severity

Reproducibility

Date Submitted

Last Update

0001766

[Resin]

minor

always

05-29-07 09:43

06-21-07 13:03

Reporter

ferg

View Status

public

Assigned To

ferg

Priority

normal

Resolution

fixed

Status

closed

Product Version

3.1.1

Summary

0001766: Load-Balancer detection of frozen server

Description

(rep by Kevin MacClay)

We are experiencing a problem where the Resin front end load balancer is failing to detect that one of the back-end Resin instances is not processing requests. The problem is that the back-end Resin instance is accepting connections but does not actually process the request. Since the front-end LB does not mark the back-end instance down, it continues directing traffic there. Users who are directed to the ?bad? instance are frozen up.

We are currently using Resin 3.0.14 for this application. I noticed Resin 3.0.20 has a ?client-connect-timeout? enhancement to the LB. But I?m not sure that this would help either since a connection can be established to the back-end instance, it just hangs after that.

In addition, we would also like to create a Nagios monitoring script that would be able to notify us if a back-end instance enters this state. Do you have any recommendations on how to test whether a back-end ?srun? instance is accepting _and handling_ requests?

Additional Information

Attached Files

Relationships

There are no notes attached to this issue.

Issue History
Date Modified	Username	Field	Change
05-29-07 09:43	ferg	New Issue
06-21-07 13:03	ferg	Assigned To	=> ferg
06-21-07 13:03	ferg	Status	new => closed
06-21-07 13:03	ferg	Resolution	open => fixed
06-21-07 13:03	ferg	Fixed in Version	=> 3.1.2
06-21-07 13:03	ferg	Description Updated

Mantis 1.0.0rc3[^]

27 total queries executed.
24 unique queries executed.