Home assistant will not start after crash

can someone help me out with this? my server crashed and now I can’t get home assistant to start…

I’m running it in a proxmox Alpine linux vm using the official docker container…version 0.70.1 (haven’t had time to update recently).

2019-02-02 19:49:25 ERROR (MainThread) [homeassistant.core] Error d
oing job: Task exception was never retrieved
Traceback (most recent call last):
File “uvloop/loop.pyx”, line 1795, in create_connection
File “uvloop/loop.pyx”, line 1774, in uvloop.loop.Loop.create_con
nection
ConnectionResetError: [Errno 104] Connection reset by peer

Like since May 2018 :slight_smile:

ugh…I had faith in the internet that someone would actually try to help and not pull this crap.

how long it has been since I updated is irrelevant. it was working just fine until the server crashed last night. this isn’t a “my install is outdated” issue… something happened somewhere to cause this.

Then again new releases are introduced to fix bugs, and this is one.

Anyway, those lines look familiar, have you searched the forum for that error, or put them into google?

PS do you use asuswrt? That component seems to feature in similar error message threads.

if this were a bug, it would have affected me from the jump. this has been working just fine for months, including moving it from a pi3 to my new server.

I’ve searched for hours and found nothing, I finally got frustrated to the point where I had to post. I’m stumped.

nope. I’ve tried dropping the database and recreating, thought maybe it was a db issue from my searches…nope, didn’t work.

completely random what caused this…i had my other computer connected to the same UPS (which is what caused the crash to begin with, my UPS battery died and shut everything off), and that computer had an old docker image running on it that came back up when i booted it (without me realizing it was still there).

there was a conflict with something on the network - i’m guessing my alarm system, which can only connect to one other thing on the network at a time. since the old docker image came up first, it was connected to the alarm system and preventing my actual install from connecting. that’s my guess, anyway. as soon as i took that old image down and restarted the docker container on my actual install, it came right up.

not that anyone else would have this particular issue, but i wanted to throw it out there in case it might help someone else…

That is pretty random!

I really do recommend an update, as HA has progressed remarkably over the time we are talking. But I’d also be wary of the changes in 16 versions.

1 Like

It’s difficult when you are so far behind as most people have no idea because they are on a more recent version… My memory for instance just isn’t that good… This means when you have an error or problem, there’s limited people able to help which is why you need to work real hard with Google to get to the answer sometimes.

Good luck…

i do plan on updating soon, i just need to make the time to sit down and do it. the reason i generally take so long (other than life getting in the way) is that i have several custom components that i end up needing to update sometimes as well depending on what was changed…including not just HA custom components, but also changes to the python ecobee and envisalink APIs as well.

the nice thing about having it dockerized and inside a VM though is that i can snapshot it and have a backup if something goes wrong. i’ve also got all of my configuration stuff and my custom components in git, so i can create a new branch and go back easily that way as well.