Why am I getting the 'Configured user limit on inotify instances has been reached' error in MDSS on Linux/Kubernetes?

Linux uses inotify to monitor file system events, but it enforces limits on how many files and directories a single user can watch. The default limits (often 128 instances and 8,192 watches) are too low for Kubernetes workloads. You will need to raise this value higher to run MDSS.

If using a cloud-managed cluster (e.g., GKE, EKS, AKS), nodes are preconfigured, and changes may not be necessary or possible. Ensure this setting is adjusted on all Kubernetes nodes if needed.

Sample Error Log:

Log
Copy

You can check your current inotify instances of your nodes by running these commands:

Bash
Copy

Solution:

  1. To immediately raise the limits without restarting, run:
Bash
Copy

This takes effect immediately but will reset after a reboot. You can modify the values to be higher based on your needs.

  1. To ensure these limits remain after a restart, add them to the system configuration file:
Bash
Copy
  1. Once updated, apply the new settings by running:
Bash
Copy

If Further Assistance is required, please proceed to log a support case or chat with our support engineer.

Type to search, ESC to discard
Type to search, ESC to discard
Type to search, ESC to discard