Re: disk spinning continuously + cpu possibly overheating = sudden death

2011-07-10 Thread Manish Jain

   Hi Chuck,
   I ran memtest86+ by loading it from the loader prompt. It completed
   Pass 1 with no errors in about 45 minutes. While I'll re-run
   memtest86+ for at least 2 passes tonight, I more suspect bad blocks in
   my ffs slice. So I am planning to install sysutils/smartmontools and
   run smartctl on /dev/ad8s2 (the ffs slice). Is there any port for
   monitoring system thermals, particularly CPU ?
   Regards
   Manish Jain
   On 09-Jul-11 22:24, Chuck Swiger wrote:

On Jul 9, 2011, at 9:22 AM, Manish Jain wrote:

What do I make of it when the system runs flawlessly both on Win XP and FreeBSD
-8.0-amd64 ?

That's interesting but inconclusive.  Can you run prime95 testing overnight und
er WinXP without issues?  Or memtest86?

BTW, 8.2 does NOT - for reasons unfathomable to me - install /usr/sbin/apmd. I 
don't know how significant that is, but I have had to insert into rc.conf :

apm_enable="NO"
apmd_enable="NO"

ACPI replaced APM around 2000.  (I'm assuming you don't have a 11-year-old quad
-core Phenom system, since AMD didn't make such processors until 2007 or 2008.)

Regards,
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"


Re: disk spinning continuously + cpu possibly overheating = sudden death

2011-07-09 Thread Chuck Swiger
On Jul 9, 2011, at 9:22 AM, Manish Jain wrote:
> What do I make of it when the system runs flawlessly both on Win XP and 
> FreeBSD-8.0-amd64 ?

That's interesting but inconclusive.  Can you run prime95 testing overnight 
under WinXP without issues?  Or memtest86?

> BTW, 8.2 does NOT - for reasons unfathomable to me - install /usr/sbin/apmd. 
> I don't know how significant that is, but I have had to insert into rc.conf :
> 
> apm_enable="NO"
> apmd_enable="NO"

ACPI replaced APM around 2000.  (I'm assuming you don't have a 11-year-old 
quad-core Phenom system, since AMD didn't make such processors until 2007 or 
2008.)

Regards,
-- 
-Chuck

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"


Re: disk spinning continuously + cpu possibly overheating = sudden death

2011-07-09 Thread Manish Jain

   Hi Chuck,
   What do I make of it when the system runs flawlessly both on Win XP
   and FreeBSD-8.0-amd64 ? BTW, 8.2 does NOT - for reasons unfathomable
   to me - install /usr/sbin/apmd. I don't know how significant that is,
   but I have had to insert into rc.conf :
   apm_enable="NO"
   apmd_enable="NO"
   Thanks for any help.
   Regards
   Manish Jain
   [1]invalid.poin...@gmail.com
   On 09-Jul-11 19:59, Chuck Swiger wrote:

On Jul 9, 2011, at 3:53 AM, Manish Jain wrote:

I am running FreeBSD-8.2-amd64 on a Western Digital 320 GD disk and a
  quad-core AMD Phenom processor. After booting, the disk keeps
  continuously (meaning continuous I/O) and after a couple of hours (on
  the console with the X server not running), suddenly the whole system
  dies, powering off in a flash. I suspect the continuous I/O is causing
  is a CPU overheat which results in this behaviour.

While you probably do have flaky hardware causing problems, even under the wors
t case situation, a modern hard drive doesn't use more than about 10W even unde
r full load (ie, continuous writes).  Compare that with the 125W thermal design
 rating of a 4-core Phenom.  If your system is overheating, it's not because of
 what a hard drive is doing.

It should be easy to tell if thermals are an issue-- is the CPU heat sink getti
ng too hot to touch just before this shutdown?  If so, it's likely to be a prob
lem with a cooling fan or heatsink attachment.

Also, what is it that makes you think your disk is doing continuous I/O?
What does "iostat 5" or similar show?

Regards,

References

   1. mailto:invalid.poin...@gmail.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"


Re: disk spinning continuously + cpu possibly overheating = sudden death

2011-07-09 Thread Chuck Swiger
On Jul 9, 2011, at 3:53 AM, Manish Jain wrote:
> I am running FreeBSD-8.2-amd64 on a Western Digital 320 GD disk and a
>   quad-core AMD Phenom processor. After booting, the disk keeps
>   continuously (meaning continuous I/O) and after a couple of hours (on
>   the console with the X server not running), suddenly the whole system
>   dies, powering off in a flash. I suspect the continuous I/O is causing
>   is a CPU overheat which results in this behaviour.

While you probably do have flaky hardware causing problems, even under the 
worst case situation, a modern hard drive doesn't use more than about 10W even 
under full load (ie, continuous writes).  Compare that with the 125W thermal 
design rating of a 4-core Phenom.  If your system is overheating, it's not 
because of what a hard drive is doing.

It should be easy to tell if thermals are an issue-- is the CPU heat sink 
getting too hot to touch just before this shutdown?  If so, it's likely to be a 
problem with a cooling fan or heatsink attachment.

Also, what is it that makes you think your disk is doing continuous I/O?
What does "iostat 5" or similar show?

Regards,
-- 
-Chuck


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"


disk spinning continuously + cpu possibly overheating = sudden death

2011-07-09 Thread Manish Jain

   Hello all,
   I am running FreeBSD-8.2-amd64 on a Western Digital 320 GD disk and a
   quad-core AMD Phenom processor. After booting, the disk keeps
   continuously (meaning continuous I/O) and after a couple of hours (on
   the console with the X server not running), suddenly the whole system
   dies, powering off in a flash. I suspect the continuous I/O is causing
   is a CPU overheat which results in this behaviour. The strange thing
   is when I ran a script to ktrace and kdump, there is no sign of any
   unwarranted I/O. Can anyone suggest what possibly be wrong with my
   system ? I have disabled as much as possible in rc.conf and
   syslog.conf to prevent I/O. My script and the output are pasted below.
   BTW, the system runs flawlessly both on Win XP and FreeBSD-8.0-amd64.
   The reason I do not want to install FreeBSD-8.0-amd64 is that its Logo
   saver, one of my al-time favourites, does not work. In 8.2, it does
   work. One thing I should mention is that, apart from swap, I have only
   one partition / (35 GB). No other mountpoints.
   BTW, does FreeBSD have anything like Linux badblocks to check for any
   bad sectors in hard disks ?
   Thanks for any help.
   Regards
   Manish Jain
   [1]invalid.poin...@gmail.com
   #!/usr/local/bin/bash
   ktrace -dip 0
   sleep 10
   ktrace -C
   kdump -m64 | egrep '\|\' > ~/diskio.log
   exit 0
   Here are the results :
 1640 bash CALL  read(0xff,0x800e23080,0x77)
 1640 bash GIO   fd 255 read 82 bytes
 1640 bash RET   read 82/0x52
 1642 sleepCALL  read(0x3,0x7fffe170,0x80)
 1642 sleepGIO   fd 3 read 128 bytes
 1642 sleepRET   read 128/0x80
 1642 sleepCALL  read(0x3,0x800533000,0xb4)
 1642 sleepGIO   fd 3 read 180 bytes
 1642 sleepRET   read 180/0xb4
 1642 sleepGIO   fd 3 read 4096 bytes
  858 ppp  CALL  read(0,0x801420140,0x800)
  858 ppp  GIO   fd 0 read 18 bytes
  858 ppp  RET   read 18/0x12
  858 ppp  CALL  read(0,0x801420140,0x800)
  858 ppp  GIO   fd 0 read 53 bytes
  858 ppp  RET   read 53/0x35
  858 ppp  CALL  write(0x6,0x7fffd980,0x34)
  858 ppp  RET   write 52/0x34
  858 ppp  CALL  read(0x6,0x7fffe320,0x804)
  858 ppp  GIO   fd 6 read 44 bytes
  858 ppp  RET   read 44/0x2c
  858 ppp  CALL  write(0,0x80145b138,0x2d)
  858 ppp  RET   write 45/0x2d
  858 ppp  CALL  read(0,0x801420140,0x800)
  858 ppp  GIO   fd 0 read 53 bytes
  858 ppp  RET   read 53/0x35
  858 ppp  CALL  write(0x6,0x7fffd980,0x34)
  858 ppp  RET   write 52/0x34
  858 ppp  CALL  read(0x6,0x7fffe320,0x804)
  858 ppp  GIO   fd 6 read 44 bytes
  858 ppp  RET   read 44/0x2c
  858 ppp  CALL  write(0,0x80145b338,0x2e)
  858 ppp  RET   write 46/0x2e
 1640 bash CALL  read(0xff,0x800e23080,0x77)
 1640 bash GIO   fd 255 read 73 bytes
   kdump -m64 | egrep '\\|\\' > ~/diskio.log"
 1640 bash RET   read 73/0x49
 1643 ktrace   CALL  read(0x3,0x7fffe140,0x80)
 1643 ktrace   GIO   fd 3 read 128 bytes
 1643 ktrace   RET   read 128/0x80
 1643 ktrace   CALL  read(0x3,0x800534000,0xb4)
 1643 ktrace   GIO   fd 3 read 180 bytes
 1643 ktrace   RET   read 180/0xb4
 1643 ktrace   GIO   fd 3 read 4096 bytes

References

   1. mailto:invalid.poin...@gmail.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"