As the title suggests, I have a T500 that often hangs during heavy I/O load. There is nothing at all left in the kernel logs nor do any LEDs flash indicating a kernel panic. The hard drive LED lights steady, the screen stops updating, the laptop does not respond to ping, and it stays this way indefinitely until I hard reset it. The fan may spin faster if the temperature increases, but I've had this happen where the fan stays on a low speed. If there is any audio playing, it will loop the last several seconds endlessly.
I see in Issues 289 and 493 there is mention that applying a microcode update can help this kind of stability. However I am skeptical that something else isn't causing this. I say this because I have also bought an identical laptop for someone else. The other laptop seems to run forever without this issue. That laptop's owner watches HD videos like crazy, browses heavy web content, etc., and the laptop just went 93 days until it had to be shut down (cleanly) due to a zombie process (not due to this issue). I am lucky if I can go two weeks on mine. The CPU's are P8400 and both laptops have 4 GB ram, identical brands/types in identical slots. Both run the exact same identical OS, block for block (BLFS compiled by me, kernel 4.4.14 from Slackware, recompiled and deblobbed). Both use the elevator I/O scheduler with significantly reduced dirty ratios.
The only difference is that use a SSD with LUKS but I also get this problem with a traditional rotating hard drive. It almost feels as though once a certain amount of I/O activity has been reached, this problem arises. If I use my laptop very lightly, it might go two weeks, but if I use it heavily, it only goes a couple of days. I've even tried slowing down the SATA connection, using the hard drive via a USB connector, and nothing has helped. I've also tried four different hard drives and nothing.
Are the others with this issue using SSD, LUKS, particular brands of RAM, particular sizes of RAM, etc? There must be something offensive here if one laptop can run normally and another can't.
Besides updating µcode, Sounds like classic issues with linux on the desktop and HDD would make it worse. Most likely cause (besides old µcode) is running out of ram and oom-killer not being as great as it was in the past. Linking some userspace alternatives below:
Also this bug tracker is not a support forum, please use irc channel instead (#libreboot @freenode).