Comment 54 for bug 569680

Hardware: Norhtec MicroClient JrMX (an i582 SoC based on Xcore86) with HD hooked up to SATA
Kernel: compiled from source 2.6.36 #7 Wed Dec 8 22:05:03 EST 2010 i586 GNU/Linux
Distro: Ubuntu 10.04
Symptoms: rare (once in tens of hours) freezes for arbitrary amounts of time during significant network and hard-drive activity

Suspecting this might be a HD problem, I switched to a spare HD (dd'ed the file system) only to find that the problem persists. Until today's miracle I didn't even have any output from my headless box, but today the box unfroze after almost *4 days* of being in the frozen state (I have logs that prove it, in addition to observations). First thing I did after I noticed that it came alive was dmesg and there was the error cited above in this bug report [ attached ].

I vaguely recollect that the freezes started with 10.04, although there were other problems before (network interface dying) [which prompted me to compile the kernels with hopes of newer network driver fixing things]. I will test with an earlier kernel and report back -- I apologize for not doing so before posting: wanted to get another vote out to this bug sooner rather than later.

lspci [useless? the only possibly relevant entries are these, but there exists an SATA interface... ]:
00:00.0 Host bridge: RDC Semiconductor, Inc. R6021 Host Bridge (rev 02)
00:07.0 ISA bridge: RDC Semiconductor, Inc. Device 6036
00:0c.0 IDE interface: RDC Semiconductor, Inc. Device 1011 (rev 01)