c178ba3686
It turns out the get-pip.py download doesn't always fail with a straight "can't connect to host". It often will fail just by stalling out in the download. This is kind of hard to detect, however a reasonable heuristic is no data found after 10s and eliminate all the 100% complete downloads with 434k content (which is the size of get-pip.py). tripleo needs to be filtered out because it downloads way too much stuff which fails, but the scripts work around it (or possibly just fail late). Change-Id: I09dbbe7217be999fee525b8d1d9aad76456e414c
9 lines
207 B
YAML
9 lines
207 B
YAML
query: >
|
|
tags:console
|
|
AND NOT build_queue:check-tripleo
|
|
AND
|
|
(message:"couldn't connect to host"
|
|
OR
|
|
(message:"0 0 0 0 0 0 0 0 --:--:-- 0:00:10 --:--:-- 0"
|
|
AND NOT message:"100 434k 100 434k"))
|