SSD wear on Saruman may need smartctl through hpsa: the iLO's wear columns are blank for both SM863a drives

[Gerrrt]

Conditional on the logical drive

docs/roadmap.md (the Saruman disk paragraph) records that the wear and endurance columns are blank on both SM863a drives as the iLO reports them today, and says: if they stay blank once the drives are in a logical drive, wear monitoring on the newest drives in the estate needs smartctl through the hpsa path, "which is an issue the runbook's closing paragraph already describes."

This is that issue, filed with the condition stated rather than after the fact. Do not start it until the SSD fit issue has created logical drive 2 and the iLO's cpqDaPhyDrv wear fields have been re-read.

If the condition holds

• smartctl -d cciss,N against the controller from the host, on a timer, into the textfile collector smart-state already uses on Saruman (SMART is watched on Saruman's array only — oracle, prometheus and morpheus have none #351's mechanism) — or a reason it cannot.
• A rule on the SSDs' wear indicator, whichever attribute the SM863a exposes through cciss, with a unit test.
• docs/observability.md's SMART paragraph names the two paths: iLO SNMP for the spindles, hpsa for the SSDs, and why.

Done when

Either the iLO reports wear for both SSDs and this closes as unneeded, or smart-state carries a wear series for both and a rule reads it.

Refs #418, #351.

[Gerrrt]

The condition holds — this issue is no longer gated.

Logical drive 2 was created on 2026-09-19 (#527, recorded in #539) and the iLO re-read with both SSDs configured since 13:40 UTC reads exactly what it read unconfigured:

Metric, cpqDaPhyDrvIndex 2 and 3	Value
cpqDaPhyDrvSSDWearStatus	1 other
cpqDaPhyDrvSSDPercntEndrnceUsed	4294967295
cpqDaPhyDrvPowerOnHours	4294967295
cpqDaPhyDrvSSDEstTimeRemainingHours	4294967295
cpqDaPhyDrvHasMonInfo	1 false (all four drives)
cpqDaPhyDrvSmartStatus	2 ok
cpqDaPhyDrvCurrentTemperature	29 / 31 °C

So "an unconfigured drive may simply not be polled" was not the explanation. The iLO does not report wear for these drives in any state, and smartctl -d cciss,N through hpsa is the path. One thing worth knowing before starting: ssacli is still not on the host (the array was made from the offline SSA), which does not block smartctl — it talks to the driver, not the vendor tool — but the hpsa device numbering N has to be found rather than assumed, and bays 3 and 4 are the SSDs.