Linux Patches Updated to Fix Latest Suspend/Resume Issues for Some AMD Laptops
phoronix.com
phoronix.com
Just last week I spend days troubleshooting a debian server issue that was shutting down unexpectedly. The server, located on prem, was registering "power key pressed" events so someone must have pressed the power button, right? But the events were on late afternoons, almost always on weekends, when no one was there!
Going through all the logs I could see the system entering suspend mode after the key press and never actually resuming.
The client would call their IT guy, who would iLO to the server and do a reboot - after which everything was fine. The logs even said as much, seeing the related virtual keyboard events and all.
After much googling, greping, zgreping and talking with their IT guy, we were able to find the root cause.
The client had placed the server in a wooden cabinet, with no ventilation whatsoever. After hours, when A/C was off for a while, the server would overheat (but not the CPUs, so `sensors`wasn't showing anything out of the ordinary). iLO would then trigger a critical temperature event and send an OS shutdown event that was registered by debian as "power key pressed". Oh man!
Their IT guy drilled a whole on the cabinet and installed a fan. It remains to be seen how that will work out.