Separated params for backup and tidy cronjobs - Fixes #3

[gmwubs]

In manifests/config.pp the module creates a cronjob with the following oneliner:
command => "/usr/bin/find ${influxdb::backup_directory} -mtime +${influxdb::backup_keep} -type f -delete";

If backup_keep is set to 2, this creates something like:
/usr/bin/find /var/lib/influxdb/backups -mtime +2 -type f -delete

This is caused by the cronjobs for backup and tidy being created to run at the exact same time. In certain circumstances, this results in the the backup location running low on disk space, triggering "false alerts".

The solution introduced here separates parameters for both cron jobs, avoiding a race condition between the two cron jobs.

[xandm]

Thanks - can you also update data/common.yaml (lines 11 and 12) to have new default settings?

[gmwubs]

Of course! Sorry I didn't think of that!

[xandm]

I don't mind merging it in the current state but wonder if the daily_backup_hour should be later than daily_tidy_hour so tidy happens first?

[gmwubs]

At first, I thought it was perhaps best to create a backup first. On second thought: if the space available is not enough, the backup will fail. I would suggest people make sure that they do off-line backups, for example by creating snapshots or syncing the locally created backups elsewhere :-)

[gmwubs]

Any indication when this PR will be accepted?