cleanup jobs logs

[kraj007]

Hello team,

We have configured below genie configurations as our genie instances were used to get full on storage frequently.

genie.tasks.diskCleanup.expression=0 0 */3 * * * genie.tasks.diskCleanup.retention=2

Even after this , we are facing disk space issues. I want to check the cleanup job logs to check what exactly happening in logs - like how many job files deleted , how much space got free. Where i can check the logs ?

Can anybody please answer this

[mprimi]

There are metrics that measure what you're asking for. Look for genie.tasks.diskCleanup.* in our metrics documentation. They include numberDeletedJobDirs, numberDirsUnableToDelete, unableToGetJobs and unableToDeleteJobsDir.

You can also observe the logs emitted by class com.netflix.genie.web.tasks.node.DiskCleanupTask.

All of this may help if the disk space is taken by job directories. But are you sure this is the case? It may be that something else is filling up your disk. Example:

• Genie server logs (Genie does not have built-in rotation and trimming)
• Jobs archive folder (unlike the jobs folder, this is for long-term storage, and it's never cleaned)
• Jobs creating files outside of their job folder (for example, your job may be writing data to /tmp/)
• etc.

Have you analyzed what is taking up disk space? If so, do you mind sharing some more detail that make you think job directories are not being cleaned up?

Can you also share the exact version of Genie you are running?