Http server intermittently unavailable

Since a few days I’m getting a lot of http timeout from lets encrypt servers
Even with a high read timeout (30 seconds) I’m still getting http errors.
Just a simple http request to https://acme-v01.api.letsencrypt.org/ with wget sometime fails

# wget https://acme-v01.api.letsencrypt.org -O /dev/null
--2016-05-10 15:46:50--  https://acme-v01.api.letsencrypt.org/
Résolution de acme-v01.api.letsencrypt.org (acme-v01.api.letsencrypt.org)… 2a02:26f0:e1:2b0::3d5, 2a02:26f0:e1:2ba::3d5, 23.201.217.118
Connexion à acme-v01.api.letsencrypt.org (acme-v01.api.letsencrypt.org)|2a02:26f0:e1:2b0::3d5|:443… connecté.
requête HTTP transmise, en attente de la réponse… 500 Internal Server Error
2016-05-10 15:47:26 erreur 500 : Internal Server Error.

Eventually it says in your log “500 Internal Server Error”. So I suppose if this is really coming from Let’s Encrypt’s Boulder server then LE people can respond to this thread to say why.

Meanwhile though I see your client connects over IPv6. This should work (of course, otherwise it would be wrong for Let’s Encrypt to advertise an AAAA record with an IPv6 address in it) but it might mean that you have other IPv6 problems and only noticed the effect on Let’s Encrypt because many things are still IPv4-only. So it can be worth trying wget’s IPv4-only mode as a diagnostic, to learn if that helps.

I’m now also getting some urn:acme:error:serverInternal from server

Just saw the https://letsencrypt.status.io/pages/incident/55957a99e800baa4470002da/57321180cd71dab819000a0c

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.