Investigate why CI appears to be downloading Chromium on every job
The following message appeared in a failed job recently (see !149 (comment 132608953)), and I am wondering whether this means we're re-downloading chrome for every job. If so, this sounds pretty inefficient. We should either cache this binary to save time, or make sure it is baked into the docker image as it is for our other CI jobs.
Command: node install.js
Arguments:
Directory: /builds/gitlab-org/gitlab-ui/node_modules/puppeteer
Output:
ERROR: Failed to download Chromium r571375! Set "PUPPETEER_SKIP_CHROMIUM_DOWNLOAD" env variable to skip download.
{ Error: connect ETIMEDOUT 74.125.141.128:443
at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1117:14)
errno: 'ETIMEDOUT',
code: 'ETIMEDOUT',
syscall: 'connect',
address: '74.125.141.128',
port: 443 }
info Visit https://yarnpkg.com/en/docs/cli/install for documentation about this command.
ERROR: Job failed: exit code 1