hw-health reporting memory error on healthy system

[Michał Ajduk]

hw-health reports Nagios memory error on a health system.

Root cause:
/usr/sbin/ipmimonitoring tool used reports:
122 | Mem_Stat_C01S03 | Memory | Critical | N/A | N/A | 'Presence detected' 'Configuration error'
123 | Mem_Stat_C01S04 | Memory | Critical | N/A | N/A | 'Presence detected' 'Configuration error'
130 | Mem_Stat_C01S11 | Memory | Critical | N/A | N/A | 'Presence detected' 'Configuration error'
131 | Mem_Stat_C01S12 | Memory | Critical | N/A | N/A | 'Presence detected' 'Configuration error'
134 | Mem_Stat_C02S03 | Memory | Critical | N/A | N/A | 'Configuration error'
135 | Mem_Stat_C02S04 | Memory | Critical | N/A | N/A | 'Configuration error'

However the memory status in iLO (attached screenshot) is perfectly fine.

[Michał Ajduk (majduk)]

Details of ipmi-sensors debug:

=====================================================
Get Sensor Reading Request
=====================================================
[ 2Dh] = cmd[ 8b]
[ 13h] = sensor_number[ 8b]
=====================================================
Get Sensor Reading Response
=====================================================
[ 2Dh] = cmd[ 8b]
[ 0h] = comp_code[ 8b]
[ 0h] = sensor_reading[ 8b]
[ 0h] = reserved1[ 5b]
[ 0h] = reading_state[ 1b]
[ 1h] = sensor_scanning[ 1b]
[ 1h] = all_event_messages[ 1b]
[ 80h] = sensor_event_bitmask1[ 8b]
[ 0h] = sensor_event_bitmask2[ 7b]
[ 1h] = reserved2[ 1b]
135 | Mem_Stat_C02S04 | Memory | N/A | N/A | 'Configuration error'

[James Troup (elmo)]

What version of iLO is in use? It looks like this could be a firmware bug.

[Michał Ajduk (majduk)]

It's the latest Gen10 ILO, v5:

Product Name
ProLiant DL380 Gen10
Product ID
P19717-B21
System ROM
U30 v2.34 (04/08/2020)
System ROM Date
04/08/2020
Redundant System ROM
U30 v2.34 (04/08/2020)
iLO Advanced
iLO Firmware Version
2.30 Aug 24 2020

[Eric Chen (eric-chen)]

This issue was pending for a while and this charm is no longer being actively maintained. Please consider using the new hardware-observer-operator instead. (https://github.com/canonical/hardware-observer-operator)

This issue is not critical, therefore, I mark it as "won't fix"