Open koelslaw opened 3 years ago
This could be an issue with Elastic 7.11 it looks like. I will keep this open but want to try another elastic version to see if we get the same errors with the dashboards before changing them.
As discussed I believe a good place to look for this issue is in the lighttpd proxy config. I confirmed this in 7.11 but didn't test 7.12. Here is a snippet of the lighttpd error log:
2021-03-11 16:10:38: (http-header-glue.c.1032) proxy backend sent invalid response header (Transfer-Encoding) to HTTP/1.0 request
2021-03-11 16:10:38: (gw_backend.c.2125) response not received, request sent: 927 on socket: tcp:127.0.0.1:5601 for /internal?, closing connection
Ended up removing the lighttpd which was buggy and installed nginx.
You can use this simple configuration as it works, but if your rock web access is in a closed network only. If you feel like switching to a SSL certificate and auth you will have a little bit more work to do.
location / {
proxy_pass http://127.0.0.1:5601;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-for $remote_addr;
port_in_redirect off;
proxy_connect_timeout 3600;
}
Also in kibana.yml add this so it doesn't buzz all the time with the publicbaseurl notice:
server.publicBaseUrl: "http://127.0.0.1:5601"
server.host: 127.0.0.1
server.port: 5601
Hope it helps.
Indexes are present but the dashboards error out.
Conditions:
4 Cores 2 NICs 16 GB RAM Automated Install Single Node rocknsm-20210514-14_14_30.iso