Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

I have seen udev rules break when devices appeared during startup with my own custom scripts, specifically when there were multiple actions on the same trigger.

  • Do you see the appropriate symlinks in /dev/disk/by-path or /dev/disk/by-uuid before running udevadm trigger? (UUID obviously only if you have an FS on it)
  • Can you duplicate this? If so, can you provide instructions?
  • Are you booting these instances with disks attached, or attaching after booting?
  • Please provide distro and version of udev

I have seen udev rules break when devices appeared during startup with my own custom scripts, specifically when there were multiple actions on the same trigger.

  • Do you see the appropriate symlinks in /dev/disk/by-path or /dev/disk/by-uuid before running udevadm trigger? (UUID obviously only if you have an FS on it)
  • Can you duplicate this? If so, can you provide instructions?
  • Are you booting these instances with disks attached, or attaching after booting?
  • Please provide distro and version of udev

Update

I will test this tonight.

In the meantime, a few things to eliminate... Does the problem occur:

  • on a different distro (preferably one with stand alone udev)?
  • when performing the Heat steps manually?
  • when using a different cinder backend (ceph/lvm/qcow2)?

Also, I have not tried it myself with systemd, but you should be able to log all the udev events for debugging purposes by tweaking the appropriate /usr/lib/systemd/system/systemd-udev* files.

My suspicion currently is the initial udev-trigger rule is getting killed halfway through. Either way, debug logs should tell you all the events you are interested in.