Server Failure - Multi-side environment

Answered

Bert van de Ven - Securicon

October 27, 2022 11:21 Edited

We have placed a few servers with NX at a multi-side location. Every day we receive a few "Server Failure, Reason: Connection to server is lost" messages from almost all servers at the same times, mostly in the evening. The information below can be read in the logbook. However, it is unclear to us what the cause could be.

We don't miss any recordings in any location, so it seems that something is restarting in the network between these locations. How can we clarify or resolve this message.

2022-10-26 19:05:07.845 146204 WARNING nx::network::cloud::udp::TunnelAcceptor(0xffff84461120): connectionAck error: networkError

2022-10-26 19:05:07.845 146204    INFO nx::network::cloud::CloudServerSocket(0xffff743ff6c0): Cloud connection (session dcd21734-ffb8-4b1b-8a80-9d516140408d) from ms_2485e1b8-3708-d21c-db34-bec5e02a4a7d_1685295486 has not been accepted with error Software caused connection abort. Info UDP hole punching acceptor. remote endpoints { xx.xx.xx.xx:43639, xx.xx.xx.xxx:43639 }

Which connection is lost? The connection with the cloud or with one "Master" NX Server?

 

0

• 

  Norman

  • Network Optix team

  October 27, 2022 11:57

  Hello Bert van de Ven - Securicon,

  It seems like there was a brief issue between the Nx Cloud and the applicable server (the one you collected the logs from).

  For more details, more VERBOSE logging would be required, but most likely the warning is just FYI. It can happen that, for some reason, a request wasn't answered in time, and such warning is provided, but the request will be made again, and most likely will be replied to in time.

  If operational issues occur, it makes sense to dive deeper, but at this stage with the given information, I would consider it just FYI.

  0
• 

  Bert van de Ven - Securicon

  October 27, 2022 12:53

  Hello Norman, thanks for your answer. We've had this notification 228 times in the past 30 days, so I'd love to see what happened anyway. Which logs and levels you suggest to activate? EC2_TRAN, HTTP, MAIN?

  0
• 

  Norman

  • Network Optix team

  October 28, 2022 07:10

  Hi Bert van de Ven - Securicon,

  You could elevate the MAIN loglevel to VERBOSE.

  Please note, we use logs to find the cause of an issue, we don't investigate logs to find an issue.
  Especially, since this is most likely an internet connection related phenomenon. 

  Moreover, for such issues, Wireshark is probably the better tool to investigate network issues.

  0

Please sign in to leave a comment.