The readahead section in plugins/cassandra-expert/references/general/os-settings.md currently only shows how to check and set readahead immediately with blockdev. It should also include commands for making the setting persistent across reboots.
The previous udev rule example was removed because it was incorrect (udev admin trigger instead of udevadm trigger, and the rule itself may not be the best approach for all environments).
What to add:
- Correct udev rule for persistent readahead on modern Linux (if udev is the right approach)
- Alternative: systemd tmpfiles or a systemd service unit
- Tuned profile approach (common on RHEL/CentOS)
- Cloud-specific guidance (AWS EBS, GCP PD, etc.) where readahead defaults may differ
- Verification command to confirm the setting survived a reboot
Keep it practical — operators need copy-pasteable commands, not a survey of options.
The readahead section in
plugins/cassandra-expert/references/general/os-settings.mdcurrently only shows how to check and set readahead immediately withblockdev. It should also include commands for making the setting persistent across reboots.The previous udev rule example was removed because it was incorrect (
udev admin triggerinstead ofudevadm trigger, and the rule itself may not be the best approach for all environments).What to add:
Keep it practical — operators need copy-pasteable commands, not a survey of options.