Comment 14 for bug 1908452

Note that, in our environment, I've observed in 2 occasions that https connections to images.maas.io sometimes fail initially w/ an SSL error, then succeed on subsequent attempts[*]. We've no proxy configured in our environment, and I'm told no transparent one either, so it's possible the issue is on the server side. I wonder if this is related to the regiond infinite read() hang - and if so - why doesn't it eventually fail like wget does? I wonder if there's a way to simulate this w/ a standalone script that approximates whatever regiond is doing at this point. Would this do the trick?

sstream-query --keyring /snap/maas/current/usr/share/keyrings/ubuntu-cloudimage-keyring.gpg https://images.maas.io/ephemeral-v3/daily/

Unfortunately we can't readily reproduce the SSL failure, we just have to wait for it to happen again naturally.

[*]
ubuntu@maas:~$ wget https://images.maas.io
--2021-01-14 15:22:05-- https://images.maas.io/
Resolving images.maas.io (images.maas.io)... 91.189.88.136, 2001:67c:1360:8001:1::4
Connecting to images.maas.io (images.maas.io)|91.189.88.136|:443... connected.
Unable to establish SSL connection.
ubuntu@maas:~$ wget https://images.maas.io
--2021-01-14 15:23:25-- https://images.maas.io/
Resolving images.maas.io (images.maas.io)... 91.189.88.136, 2001:67c:1360:8001:1::4
Connecting to images.maas.io (images.maas.io)|91.189.88.136|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1460 (1.4K) [text/html]
Saving to: ‘index.html’

index.html 100%[===================>] 1.43K --.-KB/s in 0s

2021-01-14 15:23:25 (37.7 MB/s) - ‘index.html’ saved [1460/1460]