Ran into an issue with a Windows server running WAS. After a server restart, a certain WAS process couldn't start up because "oneagentos" was using a port required by it (9402).
I asked about this via chat, and the response was that this IP for the oneagentos process is assigned by the OS, and there's no way to configure that manually (like we can for the Watchdog processes in the 50000+ range).
Does anyone have any tips on how avoid these conflicts? My first thought was to set all OneAgent services to "Delayed Start". I wonder if that setting persists after a OneAgent update? And on the other hand, with a risk like this being there, why not just have the OneAgent service run as "Delayed Start" always by default?
From the port number its look like its the OA watch dog range of ports
Check out if setting the OA WD ports with onagentctl can help you as described in What can I do when OneAgent blocks a port I need?
We did not change default OneAgent WatchDog default port range : 50000 - 50100.
I mentioned port number 49791. It may have been consumed by either one of : oneagentloganalytics - oneagentnetwork - oneagentos - oneagentplugin . Not by the oneagentwatchdog.
In my case : the customer application did not start anymore. My Win Admins had no choice but to delay windows OneAgent service. Which is a no go because it will brake full-stack deep monitoring on processes started before the OA.
Anyway one could specify port range for this (not oneagentwatchdog) sub-processes ?