disk-monitor should activate fault/fail indicators
The disk-monitor FMA module presently tracks the relationship between disk topology nodes, the bay they are installed in, and the fail/etc indicator facility nodes attached to that bay. It also has a means for expressing (declaratively, through XML maps, etc) a set of state-change-to-indicator-presentation rules which seem only to be specified for Sun X4500 servers.
Most of the plumbing required to make this more general exists in this module already. We should add fall-back behaviour so that when no rules are provided (through XML, or otherwise) we fall back to a reasonable default case:
- If a failed (or repaired) topology node as reported in an opened FMA case is a disk node
- and it is a child of a bay node
- and that bay node has an indicator=fail
- then set the mode of that indicator to ON or OFF, as appropriate.