Closed Nisaba closed 4 years ago
What version number. is your BTCPay node running & your deployment method ?? also please check this issue, from a few days ago, I had mentioned it to @NicolasDorier & the RTL team https://github.com/Ride-The-Lightning/RTL/issues/391
I just tested on a lunanode BTCPayServer v1.0.5.4 and Server settings / Services / Ride the Lightning server link working ok. for me.
but I did spot this error on a Azure build...and bad gateway.. on my Azure server. was told this fix after by RTL team.
RTL config path should be set to ‘/data/’ only, NOT ‘/data/RTL.conf/‘.
but still not had time to test yet.
Nisaba can you check your RTL container logs ?? I spotted this error. from v1.0.5.3
OK, Thanks for the information. I've the v1.0.5.3 version.
I will ask to my admin and keep you informed
I just tested on a lunanode BTCPayServer v1.0.5.4 and Server settings / Services / Ride the Lightning server link working ok. for me.
but I did spot this error on a Azure build...and bad gateway.. on my Azure server. was told this fix after by RTL team. RTL config path should be set to ‘/data/’ only, NOT ‘/data/RTL.conf/‘. but still not had time to test yet.
Nisaba can you check your RTL container logs ?? I spotted this error. from v1.0.5.3
Hey @Eskyee please note a correct the config path should be /data
@Nisaba are you fully synched?
@Nisaba also have you updated to latest version with btcpay-update.sh
?
Hi all, i m on the same team with @Nisaba, we did the update with btcpay-update.sh and the docker is updated to v1.0.5.4. We check also on the config path is /data.
those are the log
docker logs --tail 100 generated_lnd_bitcoin_rtl_1
Error: No such container: generated_lnd_bitcoin_rtl_1
docker logs --tail 10 btcpayserver_lnd_bitcoin
2020-08-06 09:26:56.361 [INF] CRTR: Processed channels=0 updates=31 nodes=1 in last 59.999876734s
2020-08-06 09:27:42.753 [INF] DISC: Broadcasting 140 new announcements in 14 sub batches
But still get the 502 Bad Gateway.
@mankwes, @NicolasDorier, @Eskyee
Are you running it on localhost
or remote? localhost is the default value for HOST
setting but if you are running it remotely then you need to provide HOST environment value.
https://github.com/Ride-The-Lightning/RTL/issues/391#issuecomment-666374818
@ShahanaFarooqui i'm running it on remote domain name.
I'm running on remote
@mankwes wrong command, should be
docker logs generated_bitcoin_rtl_1
Sorry here are the log
docker logs generated_bitcoin_rtl_1
Server is up and running, please open the UI at http://localhost:3000
I can confirm problem on another instance. Nothing useful in logs. On nginx getting:
[error] 22#22: *14866 connect() failed (111: Connection refused) while connecting to upstream, client: [my ip] ... upstream: "http://172.18.0.9:3000/rtl/api/authenticate/cookie?access-key=38b46ac12d5xxxxx..."
I can confirm problem on another instance. Nothing useful in logs. On nginx getting:
[error] 22#22: *14866 connect() failed (111: Connection refused) while connecting to upstream, client: [my ip] ... upstream: "http://172.18.0.9:3000/rtl/api/authenticate/cookie?access-key=38b46ac12d5xxxxx..."
Hi @rockstardev has there any change been made recently for the Nginx setting? Ideally, Nginx need not query rtl with the IP address in the url, since Nginx and RTL are on the same host.
There should be two possible ways to resolve this:
http://localhost:3000/rtl/api/authenticate/cookie?access-key=38b46ac12d5xxxxx...
There hasn't been any change in Nginx setting. The change seems to have come from adding this new config setting.
@saubyk deleting new setting "host": "localhost"
from RTL-Config.json resolves the issue.
And Nginx and RTL are on the same host, but they are in different containers. So when ngxin container is hitting RTL container, of course it'll be different IP in the docker's internal network.
If you set nginx container to query http://localhost:3000 it's going to query itself.
All BTCPayServer instances that are launched since this config change was introduced are impacted and RTL won't work for them. Operators can manually fix this problem by:
cd /var/lib/docker/volumes/generated_lnd_bitcoin_rtl_datadir/_data
vim RTL-Config.json
delete "host":"localhost" line (or set it to container's IP for connection not to be refused)
remove container, start a new one that'll pick up volume and restart service
Alternative is that we tweak nginx somehow... @NicolasDorier let me know if you have any idea.
@rockstardev missed the container nuance (docker noob still) :-) Deleting the localhost setting for RTL-Config.json, results in host(container) IP address being chosen as the default I guess, which is how it should be serving.
@saubyk yeah... I'm not sure we can fix it some other way. You do expose 3000 but you are not binding port to outside Docker network (with ports
setting). So we would need some proxy process within your container in order to be able to hit it with localhost:3000.
I'm not sure if @NicolasDorier has some nginx trick up his sleeve... where we could set host... but it's question how your process would return response to outside if host is "localhost".
@rockstardev I don't think RTL should serve on the localhost in this case.
We can ensure that the default RTL-Config.json
which is generated by RTL, does not contain the host
config. That should address the problem.
That would be great. If you publish the update on your Docker image let us know so we can bump version asap to stop people from deploying instances that'll need to be manually fixed.
@rockstardev Docker images tagged with v0.8.4: https://hub.docker.com/r/shahanafarooqui/rtl/tags
GJ... @NicolasDorier merged with Emperor speed.
ugh thanks @rockstardev I think we should mention that on the doc? Or maybe let people know about the workaround so if noobs get stucks we can help.
ping @Kukks @pavlenex if noobs on channel see problems
https://github.com/btcpayserver/btcpayserver-docker/issues/346#issuecomment-670230046
The command line though is not good, as somebody not knowing how to use vim will not be able to fix the sissue.
From what I understood this only happened to newly deployed instanced on v1.0.5.3 so probably a small number of people are affected. I haven't seen many people complaining tbh. But they can easily re-deploy if cli is problem, this only affects those who want to use RTL.
Hello,
When I want to got to RTL : Server settings / Services / Ride the Lightning server / See information / Browser connection
-> I get a 502 Bad Gateway error on the browser.
Thank you for your help