I am having some trouble to start 2 inputs on the graylog interface. The error message says “Address already in use. This means that you are unable to receive any messages from this input”.
So i read previous post with the same error message, my understanding is that this happens when there’s another process already using the network socket (IP address and port). The ports are different but we can see that the same foreign adress is use. How can i correct this mistake? I dont see how i can change this from graylog server, this has to be corrected on the sending side of the logs right?
Also we can see that for the working port (5503) we have “LISTEN”. And for the not working port (5502, 5501) we have “ESTABLISHED”.
How can i correct this?
Sorry if it is not clear enought. Let me know if you need more information to understand the probleme.
Thanks
So by chaging the inputs type in Syslog TCP and GELF TCP instead of Syslog UDP and GLEF UDP i manage to get the inputs running. So they are both running now that’s good, when doing a netstat peanut | grep ":5501" and netstat peanut | grep ":5502" i have a “LISTEN” for both.
The probleme is that i still dont have any messages, and when i do a tcp dump their is no traffic. Im looking in the firewall configuration as you adviced but i dont see anythign wrong, the ports mentionned are opened both on TCP and UDP.
I checked also on the Nxlog config file on the servers supposed to send the logs and the ports numbers are good.
hi
I guess it’s a routing problem، So zero-address 0.0. 0.0 change to < IP Graylog Server >
please share config NXLOG and Explain exactly what you want to do .
Maybe you forgot to check your graylog’s logs. It has to full with port in use, or something simmilar messages.
Check your pictures…
5003 - nginx on TCP, java, on UDP, so it’s NOT the same protocoll+port.
5001,5002 - nginx is on udp, and you wonder the graylog can’t listen on the same UDP port???
I suggest to config the GL to lo address, and nginx for your ip.
But I can’t understand why you want to use nginx. In your case the simplest solution is the best.
Hello @macko003 and @bahram
Thank you very much for your concern and your answers. It really helps me a lot.
It seems like you guys need more information to understand my configuration so i am going to give all the config file that seems relevant to me:
To answer the question of macko003 i use nginx as proxy for the web interface.
Basicaly, i have 2 Graylog server (1 master: 172.X.X.31, 1 slave: 172.X.X.32), i installed nginx in command line on the master graylog server for the web interface. And there are some servers on the other side sending logs to 172.X.X.31 on port 5501 and 5502 with Nxlog. The goal would be to receive thoses logs on the web interface.
/etc/graylog/server/server.conf:
# Enable the embedded Graylog web interface.
# Default: true
web_enable = true
http_bind_address = 0.0.0.0:9000
http_publish_uri = http://172.X.X.31:9000/
http_external_uri = http://172.X.X.31:9000/
# Web interface listen URI.
I dont have the Nxlog config file yet baharam sorry, i’ve asked for it, i may have it later. But i had the confirmation that the Nxlog services are running and sending to 172.X.X.31 on port 5502 or to 172.X.X.31 on port 5501.