Bug #598: A 5 second timeout is to small on the proxy interface - Foreman

Actions

Copy link

Bug #598

closed

A 5 second timeout is to small on the proxy interface

Added by Paul Kelly over 13 years ago. Updated about 13 years ago.

Status:

Closed

Priority:

Normal

Assignee:

Paul Kelly

Category:

Host creation

Target version:

0.2

Difficulty:

Triaged:

Bugzilla link:

Pull request:

Fixed in Releases:

Found in Releases:

Red Hat JIRA:

Description

Some smart-proxy operations take considerable time. The free_ip call could take a very long time so 5 seconds is just not long enough

Actions

Copy link

Updated by Ohad Levy over 13 years ago

more than 5 seconds is not acceptable in terms of a web request.
in which scenarios does 5 seconds is too short? so far I didn't see such a problem....

Actions

Copy link

Updated by Paul Kelly over 13 years ago

Category set to Host creation
Status changed from New to Ready For Testing
Assignee set to Paul Kelly
Branch set to bug/598-short-proxy-timeout

^{[[4;36;1mSmartProxy Columns (0.7ms)}[[0m ^{[[0;1mSHOW FIELDS FROM `smart_proxies`}[[0m
^{[[4;35;1mSmartProxy Load (0.3ms)}[[0m ^[[0mSELECT * FROM `smart_proxies` WHERE (`smart_proxies`.`id` = 2) ^[[0m
failed to fetch a free IP from our proxy: Request Timeout

ActionView::MissingTemplate (Missing template subnets/freeip.erb in view path app/views):
/usr/lib/ruby/1.8/webrick/httpserver.rb:104:in `service'
/usr/lib/ruby/1.8/webrick/httpserver.rb:65:in `run'

Add a DHCP reservation for brslc025.brs.infineon.com/172.29.205.25
Failed to set the DHCP record: Request Timeout
Rolling back due to a problem: DHCP Settings for brslc025.brs.infineon.com 10 failed brslc025.brs.infineon.comsetDHCP

It may be possible that the DHCP timeout was due to my code being stopped in the debugger, (but I do not think that it was,) but the freeip was definitely a timeout.

What is the purpose of this timeout? It is so that the system does not lock up and need manual intervention. With this logic then a one minute timeout should be fine. Foreman should never timeout if there is any possibility whatsoever that the operation will complete. ("Slow" is not a bug or failure, though it does need some investigation.)

This slowness has been observed when running the native_ms DHCP proxy. Maybe the Windows box was being hammered when this operation took place. It is a virtual NT machine, and does not host the DHCP server itself, though they are both on the same site.

Actions

Copy link