Open pintify opened 3 years ago
It has just been reduced significantly the number of data entering Kapua in order to solve the issue.
Reducing the streaming of data has solved the issue apparently. What would be required to increase again the quantity of data. The resources used in the system doesn't seem to be saturated during the issue.
The service failed soon after the previous comment. Besides, it has been detected that some new devices are not detected correctly. Even after a restart, the new devices are only classified as Device Connections. In order to get them detected, the platform must be restarted again.
After some weeks with a periodic failure each day and daily reboots of the platform, we made a clean reinstallation removing all database data (from both H2 and ElasticSearch) and the issue seems to be solved now.
Regarding the new deployment, the pending data from the devices seems to be uploading slowly and some devices required some hours to be properly connected, possibly because of the messages accumulation. Could this be the cause of the original problem? In that case, how could it be solved?
Describe the bug Kura devices are unable to open new connections with the broker (some of the connections remains while they are not restarted).
The performance of the machine is on average values. Restarting Kapua restores the functionality and wipes the issue.
To Reproduce It is produced every several hours after restarting the running instance of Kapua but it has not been reproduced in a different device.
Version of Kapua 1.2.4
Type of deployment [ ] Local Vagrant deployment [x] Docker [ ] Openshift (in its variants) [ ] Others
Main component affected [ ] Console (in case of console please report info on which browser you encountered the problem) [ ] REST API [x] Message Broker [ ] - Others
Additional context
Already discussed in gitter.
Logs from broker:
Logs from Kura: