Hi,

I am encountering duplicate runs of gzip related to big log files that need rotating when calling logrotate i.e. hourly.

My use case:

• I am receiving 100-500GB of logs via rsyslog (daily) that I save to a local logfile.
• To ensure that logs are rotated in i.e. sizeable junks, I run logrotate hourly.
• To save on disk space, i enabled delaycompress.
• When the system is under load (CPU and/or IO), compression might take >1hr. Thus, logrotate is started again and ends up delaycompress-ing the same file again, contributing to the overload of the system. In this way, I sometimes see 4-5 concurrent instances of logrotate & gzip.

Is there a way to handle this (w/o implementing a locking mechanism myself), or would this require new functionality in logrotate?

Thanks,
Raoul