Hello,
I'm seeing some strange shutdowns on two Blade 2000 (kernel 2.6.21.27)
due to temp0 > 85 °C. I have cleaned these workstations and done some
stuff to investigate for some weeks. Problem persists, but it's for me
very strange.
Following lines are printed by bbc_envctlr when this workstation builds
a linux kernel with -j4 (loadavg < 4). Room temperature is about 28 °C.
Aug 10 15:18:59 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:05 rayleigh kernel: temp0: cpu(62 C) amb(30 C)
Aug 10 15:19:05 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:10 rayleigh kernel: temp0: cpu(62 C) amb(30 C)
Aug 10 15:19:10 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:15 rayleigh kernel: temp0: cpu(62 C) amb(30 C)
Aug 10 15:19:15 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:20 rayleigh kernel: temp0: cpu(62 C) amb(30 C)
Aug 10 15:19:20 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:25 rayleigh kernel: temp0: cpu(61 C) amb(30 C)
Aug 10 15:19:25 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:30 rayleigh kernel: temp0: cpu(61 C) amb(30 C)
Aug 10 15:19:30 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:35 rayleigh kernel: temp0: cpu(61 C) amb(30 C)
Aug 10 15:19:35 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:40 rayleigh kernel: temp0: cpu(61 C) amb(30 C)
Aug 10 15:19:40 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
Aug 10 15:19:45 rayleigh kernel: temp0: cpu(61 C) amb(30 C)
Aug 10 15:19:45 rayleigh kernel: temp1: cpu(63 C) amb(30 C)
I have checked fan speeds :
Aug 10 15:19:45 rayleigh kernel: fan speeds: cpu(0c) sys(3f)
These values seem to be good (3f is forced because amb temp >= 30 °C,
0c is good because cpu temp < 65 °C).
Sometimes, reported temp0 values can grow until 85 °C. When temp0 is
growing, system is idle (thus, I don't understand how temperature can be
greater than 85 °C...).
Aug 10 04:20:06 rayleigh kernel: temp0: Above safe CPU operating
temperature, 85 C.
Aug 10 04:20:06 rayleigh kernel: temp0: Outside of safe CPU operating
temperature, 85 C.
Aug 10 04:20:06 rayleigh kernel: kenvctrld: Shutting down the system now.
Aug 10 04:20:06 rayleigh shutdown[19130]: shutting down for system halt
Aug 10 04:20:08 rayleigh init: Switching to runlevel: 0
I have written some code to debug bbc_envctrl without any success and
bbc_envctrl seem to be bug free. When system is overheating, kernel
always claims against temp0 (never against temp1 even I swap CPU0 and
CPU1.).
Does someone see the same strange shutdown process on UltraSPARC-III? I
suspect that a bad temp is reported by sensors for some reasons.
Regards,
JKB
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
