SSL The handshake operation timed out to github

Has anyone been recently receiving increasing _ssl.c:574: The handshake operation timed out when attempting to open a url to through a travis-ci?

Our builds are failing when using rosdep update

Error caused by python3.6 urllib.urlopen(). Timeout is set to 15 seconds, increasing failures over the past few days.

ERROR: error loading sources list:
	<urlopen error <urlopen error ('_ssl.c:574: The handshake operation timed out',)> (>

I have experienced probably related thing - hanging or failing ‘pip install commands’ - for example or - also increased number of failures (20%/30% of our jobs) over the course of last few days.

It looks like a routing issues from the Travis build workers… I raised an issue with Travis support


I think I’m experiencing something similar

requests.packages.urllib3.exceptions.ReadTimeoutError: HTTPSConnectionPool(host='', port=443): Read timed out. (read timeout=300.0)

Seeing similar issues. It started 3 days back I guess randomly and now seen for all runs.
I see no updates on Travis status page. Has they responded to any of the support ticket?

1 Like

Not yet. Hey Travis support? Can you do anything about it? It affects us heavily.

1 Like

We are also experiencing timeouts when downloading Elixir dependencies since last week. It can only be reproduced on Travis.

Seems That Travis engineers are investigating it now and provided a possible workaround (not very helpful in our case) in the meantime:

I don’t think this is the same issue. I was seeing both the inexplicable timeouts/hangs and the failed ssl handshakes. The timeouts seems to have been fixed (as promised in the incident you linked), but the failed ssl handshakes are persisting.

It’s the same issue I think. It’s not fixed yet also for us . I am reopening the issue.

Seeing this too, but only when contacting a specific host. Can’t share a run because it’s on a private project, but I opened up a ticket through support and am awaiting follow-up info.

i just heard back from Travis support: they have found a fix and are currently testing it before rolling it out.

For the impatient among you: there is a workaround, which involves adding the following to your .travis.yml:
- |
sudo tee /etc/resolvconf/resolv.conf.d/tail <<EOF
options attempts:5
- sudo service resolvconf restart

1 Like