Not able to download Add-ons

Good morning,

I have first posted about the issue here https://community.home-assistant.io/t/addons-not-accessible-or-starting-after-move-to-ssd-boot-and-restore-from-backup/614369/6 which I have not been able to resolve.

Since I wanted to move my system to a NUC anyway, I took the chance and did it now. HA OS installed on the NUC (directly not in a docker or so), trying to download Add-ons and voila, same problem as described in my other Post on the PI.

Upon digging in my logs, it appears to me that there is some sort of Network error which I really can’t explain. My Firewall is not blocking anything, at least this is what my Unifi logs are telling me and I have not changed anything. I also can’t narrow it down to one specific issue as there are several error messages. Strangely enough, I have been able to download SSH Add-on and tried a few random other ones (ESPHome for example) which have been working. Mosquitto broker however is not working. I am posting my logs and anything I’ve been able to gather so far below.

HA Core Errors:

Logger: homeassistant.components.hassio
Source: components/hassio/websocket_api.py:138
Integration: Home Assistant Supervisor (documentation, issues)
First occurred: 22. September 2023 um 17:00:33 (27 occurrences)
Last logged: 10:06:45

Failed to to call /addons/a0d7b954_ssh/install - 'AddonManager.install' blocked from execution, no host internet connection
Failed to to call /addons/core_mosquitto/install - Can't install homeassistant/amd64-addon-mosquitto:6.3.1: 500 Server Error for http+docker://localhost/v1.42/images/create?tag=6.3.1&fromImage=homeassistant%2Famd64-addon-mosquitto&platform=linux%2Famd64: Internal Server Error ("error parsing HTTP 408 response body: invalid character '<' looking for beginning of value: "<html><body><h1>408 Request Time-out</h1>\nYour browser didn't send a complete request in time.\n</body></html>\n"")
Failed to to call /addons/core_rpc_shutdown/install - Can't install homeassistant/amd64-addon-rpc_shutdown:2.4: 500 Server Error for http+docker://localhost/v1.42/images/create?tag=2.4&fromImage=homeassistant%2Famd64-addon-rpc_shutdown&platform=linux%2Famd64: Internal Server Error ("error parsing HTTP 408 response body: invalid character '<' looking for beginning of value: "<html><body><h1>408 Request Time-out</h1>\nYour browser didn't send a complete request in time.\n</body></html>\n"")
Failed to to call /addons/core_samba/install - Can't install homeassistant/amd64-addon-samba:10.0.2: 500 Server Error for http+docker://localhost/v1.42/images/create?tag=10.0.2&fromImage=homeassistant%2Famd64-addon-samba&platform=linux%2Famd64: Internal Server Error ("error parsing HTTP 408 response body: invalid character '<' looking for beginning of value: "<html><body><h1>408 Request Time-out</h1>\nYour browser didn't send a complete request in time.\n</body></html>\n"")
Failed to to call /network/interface/wlp0s20f3/update -
Logger: async_upnp_client.ssdp
Source: components/ssdp/__init__.py:365
First occurred: 22. September 2023 um 17:07:35 (516 occurrences)
Last logged: 10:17:36

Received error: [Errno 101] Network unreachable, transport: <_SelectorDatagramTransport fd=26 read=polling write=<idle, bufsize=0>>, socket: <asyncio.TransportSocket fd=26, family=2, type=2, proto=0, laddr=('0.0.0.0', 37432)>
Logger: async_upnp_client.ssdp
Source: runner.py:186
First occurred: 22. September 2023 um 17:06:18 (2066 occurrences)
Last logged: 10:18:24

Received error: [Errno 101] Network unreachable, transport: <_SelectorDatagramTransport fd=36 read=polling write=<idle, bufsize=0>>, socket: <asyncio.TransportSocket fd=36, family=2, type=2, proto=0, laddr=('0.0.0.0', 37057)>
Received error: [Errno 99] Address not available, transport: <_SelectorDatagramTransport fd=35 read=polling write=<idle, bufsize=0>>, socket: <asyncio.TransportSocket fd=35, family=10, type=2, proto=0, laddr=('::', 39268, 0, 0)>

These are the most re-occuring errors on the Core logger.

The supervisor logs are basically all green, except the one below when I am trying to download the Mosquitto broker.

23-09-23 09:52:20 ERROR (MainThread) [supervisor.docker.interface] Can't install homeassistant/amd64-addon-mosquitto:6.3.1: 500 Server Error for http+docker://localhost/v1.42/images/create?tag=6.3.1&fromImage=homeassistant%2Famd64-addon-mosquitto&platform=linux%2Famd64: Internal Server Error ("error parsing HTTP 408 response body: invalid character '<' looking for beginning of value: "<html><body><h1>408 Request Time-out</h1>\nYour browser didn't send a complete request in time.\n</body></html>\n"")

NMCLI status is “full” and “connected”.
My host dns (172.30.32.3, which I am not able to ping) is another one than the rest of my network (1.1.1.1)
Nslookup also gives me correct feedback.
Pinging Google.com is working as well.

I am really at a loss and have no clue what to check else. If I missed any information or logs that are needed, please let me know.

I really hope someone is able to tell me what to check to find out where the issue is.

Thanks very much!

EDIT: This is only happening for official add-ons. Community Add-ons are working.

Error in the 500 range is generally errors on the server, so GitHub probably have issues and that is the reason for your problems.

Hey, but wouldn’t then all other people here complain about this issue as well?
Is there a way I can test this?

Their CDN (Content Distribution Network) sometimes have nodes that are totally out of sync and your DNS system have now cached a record to one of those nodes.
The problem here is that you might have several caching DNS servers that have the wrong record, like your machine, which might have asked your router, which might asked your ISP, which might have asked Github.
You can either wait for the record to get stale and flushed or you can try to change DNS to another server, like Cloudflare (1.1.1.1) or Google (8.8.8.8) and then shutdown your router followed by your host, where after you can turn them on again.

This is not guaranteed to work though, because you might get the same server again, because the CDN is location-oriented and since you are coming from the same place, then it might hand you the same information again.
Then it is a question about the record getting updated or the server getting refreshed.

2 Likes

Hi WallyR,

thanks for explaining this. That gave me the idea to also reboot the router from my ISP which sits before my Unifi equipment (not ideal but there is no workaround for now). After changing the DNS and rebooting the ISP router it worked. Strangely enough that also solved two other issues I had on my network. So thanks very much for pointing me in the right direction!