Pivotal Knowledge Base

Follow

NATS timeout issue in Elastic Runtime

Environment

Product       Version

Pivotal Cloud Foundry® (PCF) Elastic Runtime 

1.7.14 and Older

 

IPSec add-on                             

 Older than 1.5.37

Symptom

Pivotal has discovered a bug in PCF Elastic Runtime (ERT) version 1.7.14 and prior which will cause a 2 to 6 minute window where applications are not routable and instead return 404 messages if the active NATS server is killed without stopping cleanly. Additionally, the IPSec add-on has been found to sever network connections on shutdown, which triggers this bug. Any upgrades which restart NATS VMs on an environment with IPSec add-on versions prior to 1.5.37 installed are likely to experience unroutable applications for 2 to 6 minutes. During this downtime, applications will continue to run, and the routing tier will recover automatically. 

Resolution

ERT versions at or greater than 1.7.15 are patched for this issue. Additionally, IPSec 1.5.37 and greater do not sever network connections until all other jobs on the VM have stopped.

Upgrades to IPsec 1.5.37 or ERT 1.7.15 are expected to trigger ERT gorouter 404 messages for 2 to 6 minutes in affected environments.

If you have any concerns or would like to discuss ways to mitigate this issue, please contact Pivotal Support.

 

 

Comments

Powered by Zendesk