We want to use mysql general query log to do real-time monitoring and auditing.

Currently our approach is:

  • set general_log=on;
  • sleep 15m;
  • set general_log=off;
  • scp & rm xxx.log;
  • set general_log=on;...

But the main problem is when turn on/off general log it'll cause a peak of slow query.

I also thought of another approach: turn on genlog; tail -f it and send out the log; periodically truncate the logfile (with "> xxx.log" or "cat /dev/null >xxx.log").

I'm wondering whether it's practical.

If only mysql would provide some built-in general log message queue stuff...

没有正确的解决方案

许可以下: CC-BY-SA归因
不隶属于 dba.stackexchange
scroll top