"Feature Request: Add Proactive Transceiver Temperature Monitoring and Kernel Logging for Thermal Alarms" #53
Description
when we use 200G optical modules, if the cage temperature or the module's internal temperature reaches 80 degrees Celsius , the module shuts down as a precautionary measure and kernel does not note it as a 'thermal initiated shutdown' , rather it treats it as link down.
if that's the case and if there is no ethernet card firmware component involved, i request an enhancement in ice driver to include temperature monitoring, or at least thermal warning from the module to be logged to dmesg/journalctl/some kernel logging mechanism.
this has serious impact as we go with higher speeds, high watt modules and temperature handling is not done appropriately (appropriate cooling measures are not in place).