Oct 11 – 13, 2010
Rutherford Appleton Laboratory
Europe/London timezone

Proposed changes to ncm-filesystems (and a future ncm-stab)

Oct 11, 2010, 11:00 AM
CR03 (Rutherford Appleton Laboratory)


Rutherford Appleton Laboratory

STFC R61 Rutherford Appleton Laboratory DIDCOT OX11 0QX UK


Luke Fernando Muñoz Mejías (CERN)


A few months ago, Ian requested some things that make perfect sense about ncm-filesystems. I more or less have the time to implement it now, so here it goes: 1) "preserve" and "format" don't make sense anymore. They do for the installer, but not on a live system. So, the component should just ignore them. It should remove a FS if it disappears from the profile, and create it if it appears. Great care should be taken so that nothing gets removed by mistake, so check your "protected_mounts" field!!! 2) Sometimes all users need is to modify the fstab for some devices that aren't handled by this component. Say, tmpfs, SMB mounts.... according to Ian's proposal, fstab manipulation should be MOVED to a different component, namely ncm-fstab. 3) As a consequence, ncm-filesystems will depend on ncm-fstab. 4) There will be NO modifications to the current schemas for block devices and filesystems. Maybe some relaxation of the checks, to cover more FSs and pseudos. In any case, it will all be backwards-compatible.

Primary author

Presentation materials