EOF Error

I am running the cloudflared DNS over HTTPS proxy with pi-hole, every few hours i get the following error:

time="2018-04-19T20:46:26+02:00" level=error msg="failed to connect to an HTTPS backend \"https://1.1.1.1/dns-query\"" error="failed to perform an HTTPS request: Post https://1.1.1.1/dns-query: EOF"
time="2018-04-19T20:46:34+02:00" level=error msg="failed to connect to an HTTPS backend \"https://1.0.0.1/dns-query\"" error="failed to perform an HTTPS request: Post https://1.0.0.1/dns-query: EOF"

Restarting the daemon fixes the problem for a while. This error also occurs right after rebooting, again restarting the daemon seems to work for a few hours.

[gnossen]

I've also been seeing this off and on. Oftentimes, I'll see it right after startup. Recently, we had a good run of about a week with no issues. Today, we saw the following error:

Apr 22 22:08:16 raspberrypi cloudflared[24544]: time="2018-04-22T22:08:16Z" level=error msg="failed to connect to an HTTPS backend \"https://1.1.1.1/dns-query\"" error="failed to perform an HTTPS request: Post https://1.1.1.1/dns-query: write tcp 192.168.1.5:59494->1.1.1.1:443: write: broken pipe"

From that point on, cloudflared stopped returning responses. Periodically, it logged this message:

Apr 22 22:08:41 raspberrypi cloudflared[24544]: time="2018-04-22T22:08:41Z" level=error msg="failed to connect to an HTTPS backend \"https://1.1.1.1/dns-query\"" error="failed to perform an HTTPS request: Post https://1.1.1.1/dns-query: EOF"

Restarting cloudflared did not alleviate the issue. Things only went back to normal once a full reboot of the system was performed.

This was observed on

cloudflared version 2018.4.6 (built 2018-04-11-1516 UTC)

[yon2004]

Pull request #24 fixes this and gives more useful errors

[andreagrandi]

@yon2004 hi, I'm still getting this problem with this more recent version of cloudflared:

cloudflared version 2018.5.3 (built 2018-05-11-1704 UTC)

[wheelq]

same here, but I see this less often once moved to different platform