Applications using UDP fragmentation crash when repeatedly deployed and underplayed.
Proposal
I noticed this issue randomly while continuously deploying and undeploying the pipeline described here
After deploying the application, stressing it, and undeploying it, the communication starts behaving oddly, causing the application to crash. The solution is to restart the network manager. This is why this looks more like a communication problem than anything else.
Most likely, the application is not very fault-tolerant. Nevertheless, this issue underlines a problem in the forwarded packets.
This issue is currently under investigation. Next time this happens, I'll post here the trace routes, logs, and stack traces.
Short
Applications using UDP fragmentation crash when repeatedly deployed and underplayed.
Proposal
I noticed this issue randomly while continuously deploying and undeploying the pipeline described here After deploying the application, stressing it, and undeploying it, the communication starts behaving oddly, causing the application to crash. The solution is to restart the network manager. This is why this looks more like a communication problem than anything else.
Most likely, the application is not very fault-tolerant. Nevertheless, this issue underlines a problem in the forwarded packets.
This issue is currently under investigation. Next time this happens, I'll post here the trace routes, logs, and stack traces.
Solution
None yet
Status
Under investigation
Checklist