For DC/OS we do two pieces. We simplify Mesos' log output via a module so that it doesn't have any of it's internal logrotate logic, and just writes to a single straight output file. We also include a mesos module so that mesos task output goes to systemd journald, making it so every piece of logging inside DC/OS ends up at a single source which can then be piped wherever people like. The modules: https://github.com/dcos/dcos-mesos-modules
Systemd auto-rotates the journald logs and does cap sized logging with it. For the non-systemd logs we package logrotate ( https://github.com/dcos/dcos/tree/master/packages/logrotate) + a simple helper script to catch accidentally orphaned files. On Tue, Jan 24, 2017 at 9:40 AM Tomek Janiszewski <[email protected]> wrote: > We are using logrotate to rotate, compress and delete old data. To keep > logs easier to search we put them into Elastic/Kibana. > > wt., 24.01.2017, 18:20 użytkownik Charles Allen < > [email protected]> napisał: > > Anyone have good hints for best practices for log rotation? > > Our mesos master ended up with many gigabytes of logs once we started > running SPARK on it (approx 2GB of master INFO logs per day). > >
