Nxlog gets a TCP timeout, then the next line it sends is garbage
I am using nxlog to send logs to another nxlog server set up as a concentrator. The TCP connection is going via an Amazon Load Balancer. The nxlog sender gets tcp connection error if it hasn't sent for a while. When it goes to send another message, it reconnects ok. The problem is that the line it sends ends up appearing as garbage on the concentrator.
It's not really an option to remove the load balancer, we need to have the concentrator server autorestart if it crashes so this is the easiest way.
Any ideas?
Hi,
The problem has been solved and - just in case anyone will experience similar issue - here's what helped:
If stickiness enabled in the AWS load balancer settings, this apparently may cause the corruption. Changing it removes problems.
Best regards,
Rafał
