Redundancy errors

[diemon]

Hi all!
Today i was editing my project, trasfering it to primary server and then press "Reload project" via remote transport. Then i noticed that my clients (2 computers) start to switch between primary/secondary server. And again and again....cyclically.

I started diagnosis Viewer and save 2 files(1 for primary, 1 for secondary). Can you look at them and tell me something, in what direction I must look to solve the problem.

Before this redundancy worked normally, i've specially tested it before.

I have filtered diagnostic log only with "Zen" name of log connection
Zenon 7.11 EE, Build: 12493

[ursulak]

in zenon network the client always "follows" the process server. So when the client changes to connect to another Sever when the current becomes standby (or disappears - not processing watchdog in network timeout).

Please check which server - 1 or 2 - is the current process server and which is standby. Are they switching too?

Make sure the Network Timeout (default 30s, access via Startup Tool - Application - Network configuration') is set to time long enough for data load exchange in your system. Make sure this timeout is the same on Server 1 and Server 2. Also client uses its timeout to detect if the process server is still there. So when a client access huge data (example big historical data for a trend) the server could be too busy to process the network watchdogs in time.

Have you change something in project network settings before transport and reload? Have you transport the project to current process server or to server 1 (or 2) - is not always the same...

imho the Forum is not the right platform to deliver log files to the Support for analysis. When you see in log some error/warning you do not understand you can ask here, but i suppose it would be hard to find here someone having time to check the files. But maybe i'm just pessimistic today .

[diemon]

Yeah, they are switching too (look at CEL screen)

[ursulak]

when the process Server writes in CEL that SB and Client are breaking down, but you know that they were there, then it means that your network timeout is too short. You have now probably bigger data exchange in the zenon network and is then too busy to exchange and process network watchdog on time (in 1/3 of network timeout).

[diemon]

Thanks, ursulap. I have changed network timeout to 120 sec. Seems that it helped, now all ok.