I have raised topics previously which were vaguely linked to this, and have been trying to debug them (unsuccessfully). Perhaps I explained badly in those, so I will try a different approach/explanation.
I get worker timeouts on various items from time to time. I have trolled all the logs, and the one that looks off-colour is the node-socketio.log
- My erpnext is set up as ERPNext.domain.co.za
- My DNS server does not have an IP for that FQDN
- yet (… perhaps that would be the quickest fix - to be confirmed)
- When I get a timeout, the log file contains something like this…
- xxx.xxx.xxx.xxx is the IP address if the *.domain.co.za from the DNS server
- :8120 is the () port on which the ERPNext system is being served
{ Error: connect EHOSTUNREACH xxx.xxx.xxx.xxx:8120
at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1191:14)
errno: 'EHOSTUNREACH',
code: 'EHOSTUNREACH',
syscall: 'connect',
address: 'xxx.xxx.xxx.xxx',
port: 8120,
response: undefined }
It seems to me that node is sending the request out to the IP address of the looked up machine
ERPNext.domain.co.za, instead of the internal/local machine.
I tried adding an entry in /etc/hosts for
127.0.0.1 ERPNext.domain.co.za
but this makes no difference to the result.
Does anyone know if there is a way to tell/trick node and/or any of the node modules (eg. the net.js as above) to use some internal equivalent of a hosts file?
Much appreciated