Timeout error message in CLI while uploading plugins inside HA cluster (the upload itself works)

Description

Steps to reproduce:
1. Start cluster on one manager
2. Upload plugin

You would expect the regular message of "Plugin uploaded. The plugin's id is 8f241af1-e2e1-43ca-9c56-4e96916e4596"

The message I got is "HTTPConnectionPool(host='10.239.1.205', port=80): Read timed out. (read timeout=5)"

Activity

Show:
Łukasz Maksymczuk
July 19, 2017, 9:36 AM

Fixed in https://github.com/cloudify-cosmo/cloudify-rest-client/pull/183

This was due to timeout only being used in a cluster http client. Uploading a plugin from the web, using `cfy plugins upload http://...` made the manager only send any response after the plugin was installed, which can take longer than the default timeout. We simply disabled read timeout (kept the connect timeout) for that one kind of request.

Further work might involve installing the plugin respond immediately with an Execution, much like creating a deployment environment, instead of just blocking for a long time.

Fixed

Assignee

Unassigned

Reporter

Uri Wygodny

Labels

Bug Type

None

Target Version

None

Severity

None

Epic Link

Fix versions

Affects versions

Configure