GSM7228PS coldstarts

[lutz]

There are a couple of new GSM7228PS switches in a customers location. All switches are behind new USVs and within professional cooled service-rooms. Some of this switches have a bad behaviour. First they send messages that the temperature has reached 255 Celsius (which ist stupid, rooms and racks are cooled with exact 20 Celsius). A short time later these switches make a 'coldstart' by themselve. This equipment seems unusable in a professional manner.


UNITMGR[213901888]: hpc_local_sensors.c(524) 128184 %% CPU Maximum Temperature Exceeded (255 degrees Celsius).
UNITMGR[213900992]: hpc_local_sensors.c(524) 386 %% CPU Maximum Temperature Exceeded (255 degrees Celsius).
TRAPMGR[213902736]: traputil.c(625) 34 %% Temperature change alarm: Sensor ID: 0 Event: 1
TRAPMGR[213902736]: traputil.c(625) 35 %% Temperature change alarm: Sensor ID: 0 Event: 3

snmptrapd[9418]: xxxxx [UDP: [x.x.x.x]:49156->[x.x.x.x]]:#012DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (8825) 0:01:28.25#011SNMPv2-MIB::snmpTrapOID.0 = OID: SNMPv2-MIB::coldStart


Are these switches all broken hardware - or is there a firmware problem ?

[lutz]

A Netgear Representative said, there is a bug in the firmware code that gave false output in the heat data (255 Celsius problem).

This error and fix in the last firmware version is not documentend in the Netgear Fix History.

One have to upgrade to the latest code (NOT to the firmware version Netgear links at their support websites, this page is broken, search for version 8...25).

The problem with sporadic coldstarts is not fixed with this firmware :mad: