Panel: du -hsb causing massive server IO lag

[vipesz]

• Panel or Daemon: Daemon
• Version of Panel/Daemon: v0.5.x
• Server's OS: CentOS 7.4
• Your Computer's OS & Browser: N/A

---

Load Average (Post fix)

Alternative Approach

I am aware that this can be configured, however there should be an alternative approach to monitor the status of the du command rather then giving it a predetermined amount of time and spawning a new process to bog down a node.

The daemon has used upwards of 3 gigs of ram and has caused the load average to spike very high as indicated above. For now, this has been configured accordingly.

[dje4321]

why not just use ionice to manage disk access time? This will allow du to use the disk but the servers have higher priority so they do not get as affected by this

just set it to best effort at 6?
ionice -c 2 -n 6 du -hsb

[Espantonius]

I have the same problem. Having servers with millions of files collapses the machines.

Could it be modified so that it does not check on a node?

The code mentioned is here: https://github.com/pterodactyl/daemon/blob/develop/src/controllers/fs.js#L76

We hope some solution.

[Espantonius]

[Update]

I have modified the code so that it does not check the size of the folder. Sure it's not a good solution but not collapsing servers. I have changed it like this:

    size(next) {
        return next(null, 0);
    }

Until there is no other solution I will have to keep it that way.

[DaneEveritt]

Closing this as we do not use this method in the new daemon to calculate storage usage.