|
|
Every time cnxmq crashes, we have a script that restarts it (per your recommendation), and when it's restarted the new process's cnxmq_run.log overwrites the old process's, so when we come in later we have no idea what caused the crash because the error message in its log was lost.
Of course we fixed it here to copy the old log aside before restarting, but you should fix the issue at the root. It's accepted in the industry that logs are always appended to, never overwritten.
This also goes for cnxrun.log etc' etc'.
| Use Case | Replication |