Closed
Bug 1322199
Opened 8 years ago
Closed 8 years ago
Add `10.*.*.*` to noproxy settings for Linux and Mac nodes
Categories
(Mozilla QA Graveyard :: Infrastructure, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: whimboo, Assigned: whimboo)
References
Details
As seen on bug 1320642 we fail to load pages which are served through the public IP of the local host. Marionette tests will hang in navigate() for about 300s for each and every call to a locally hosted page. Given that we have a hard build timeout of 60 minutes, Jenkins kills the job after that time.
Assignee | ||
Comment 1•8 years ago
|
||
All Linux and OS X machines for staging have been updated and as it looks like we do not have a problem anymore. I will follow-up with production soon.
Assignee: nobody → hskupin
Status: NEW → ASSIGNED
Assignee | ||
Comment 2•8 years ago
|
||
Looks like there are still problems with Linux not being able to load some HTTPS URLs.
Assignee | ||
Comment 3•8 years ago
|
||
I updated the ESX pages for the no-proxy settings of 10.0.0.0/24. It at least make nearly all tests passing. But i still see failures for the safebrowsing tests when the following page gets loaded:
https://fr.phish-error.mozilla.com/?hl=fr&url=https%3A%2F%2Fwww.itisatrap.org%2Ffirefox%2Fits-a-trap.html
Here the browser hangs and doesn't load that page. I will investigate more tomorrow.
Assignee | ||
Comment 4•8 years ago
|
||
I was playing around with the wpad.dat for automatic proxy configuration and it seems to work fine except that two tests are failing which want to verify the ev certificate via the certificate's OSCP URL. This check fails because loading ocsp.digicert.com never ends. As it looks like this is an issue with the automatic proxy configuration. :/
I will check if setting all the proxy values for gnome similar to the env variables will fix it.
Assignee | ||
Comment 5•8 years ago
|
||
Ok, so with the following settings all is working fine now:
export DISPLAY=:0
gsettings reset-recursively org.gnome.system.proxy
gsettings set org.gnome.system.proxy use-same-proxy true
gsettings set org.gnome.system.proxy mode "'manual'"
gsettings set org.gnome.system.proxy ignore-hosts "['localhost', '127.0.0.0/8', '10.22.73.0/8', '*.scl3.mozilla.com', '*.phx1.mozilla.com', 'mm-ci-staging', 'mm-ci-production', '*.mozqa.com']"
gsettings set org.gnome.system.proxy.http host "proxy.dmz.scl3.mozilla.com"
gsettings set org.gnome.system.proxy.http port 3128
gsettings set org.gnome.system.proxy.http use-authentication false
gsettings set org.gnome.system.proxy.http enabled true
gsettings list-recursively org.gnome.system.proxy
I will let this run today and update production later or tomorrow.
Assignee | ||
Comment 6•8 years ago
|
||
So bug 1322347 has been fixed faster than expected. I will wait with the push to production and maybe enable automatic-proxy detection instead. So production systems have to wait.
Assignee | ||
Comment 7•8 years ago
|
||
All Linux and Mac slaves in staging are looking good. Going to update production today.
Assignee | ||
Comment 8•8 years ago
|
||
This has been finished for production now. Everything seems to be green again.
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Updated•6 years ago
|
Product: Mozilla QA → Mozilla QA Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•