Description
When tests fail in the CI it is often impossible to debug it after the run, to solve this i suggest to always send the syslog as remote syslog
Store the logs per test and per DUT, to be able to be able to trace errors from CI.
- test-config always setup a syslog to the test PC (test container) over the mgmt interface (IPv6)
- Test runs
- Infamy "collect" and package the logs
- Repeat
Example content:
logs/
test1/
dut1/
kern.log
messages
syslog
dut2/
kern.log
messages
syslog
dut3/
kern.log
messages
syslog
test2/
dut1/
kern.log
messages
syslog
Another thing you could log, without overhead, is the operational state gotten (for each test) hook into get_data() (or what it is called) when getting operational, save it to the test directory. To ease debug.
then compress it to logs.tar.gz, if running in CI, publish it as an artifact.
Additional Information
No response
General Information
Anyone can help out by sponsoring development of new features or contributing pull requests.
Please use this issue for discussions related to the feature.
Description
When tests fail in the CI it is often impossible to debug it after the run, to solve this i suggest to always send the syslog as remote syslog
Store the logs per test and per DUT, to be able to be able to trace errors from CI.
Example content:
Another thing you could log, without overhead, is the operational state gotten (for each test) hook into get_data() (or what it is called) when getting operational, save it to the test directory. To ease debug.
then compress it to logs.tar.gz, if running in CI, publish it as an artifact.
Additional Information
No response
General Information
Anyone can help out by sponsoring development of new features or contributing pull requests.
Please use this issue for discussions related to the feature.