2 CPU's in timeout on my ES40

Syltrem · *nix forums Guru Wannabe Joined: 02 May 2005 Posts: 131

Has anyone experienced CPU's ceasing to be in service on an ES40 ?

I've never seen this before, there are no errors on the system or reported
by DIAG
2 CPUs out of 4 are not working right now.

$ sh cpu

System: HELIOS, AlphaServer ES40

CPU ownership sets:
Active 0,2
Configure 0-3

CPU state sets:
Potential 0-3
Autostart 0-3
Powered Down None
Not Present None
Failover None
$ sh err
Device Error Count
$1$DQB0: (HELIOS) 1
$1$DQB1: (HELIOS) 1

I tried:
$ set cpu/start 1
%SYSTEM-W-WRONGSTATE, CPU 1 is in the wrong state for the requested
operation
$

But I'm not familiar with these commands, such a problem never happened to
me (and we never had Galaxy)

If you have an idea why this would be, and have a solution, pls write !
Is this a situation where VMS needs to be rebooted just like Windows?

The 2 CPUs may not have restarted since the system was powered back on after
having failed because of the high temperature in the computer room. The
reading with f$getsyi("temperature_vector") showed 100 F at the time. The
system was rebooted when the temperature in the room started to go down.

Thanks

--
Syltrem
http://pages.infinit.net/syltrem (OpenVMS information and help, en français)

John Santos · *nix forums Guru Wannabe Joined: 25 May 2005 Posts: 189

Syltrem wrote:

Volker Halle · *nix forums beginner Joined: 29 Jul 2005 Posts: 33

Syltrem,

$ ANAL/SYS
SDA> CLUE CONFIG

may provide additional information about the state of those 2 CPUs.

Consider to capture the console output during the next boot to find
possible messages being logged on the console.

Volker.

Syltrem · *nix forums Guru Wannabe Joined: 02 May 2005 Posts: 131

"Volker Halle" <volker_halle@hotmail.com> a écrit dans le message de news:
1153294081.483595.17480@i3g2000cwc.googlegroups.com...

dave.baxter@bannerhealth. · *nix forums beginner Joined: 29 Jul 2005 Posts: 41

Syltrem wrote:

Volker Halle · *nix forums beginner Joined: 29 Jul 2005 Posts: 33

Dave,

once a CPU has entered the active set, it should not be able to get
into a TIMEOUT state all by itself, except through a STOP/CPU and
START/CPU, which would fail to start it.

If a CPU ceases operation while in the active set, you'll either get a
CPUSANITY or a CPUSPINWAIT bugcheck.From the dump, SDA> CLUE CONFIG
will show the state of the CPUs.

In your CPUSPINWAIT case, the fact that CPU #2 did not service the
bugcheck request, may be enough evidence to have a look at the state of
that CPU. Note that if AUTO_ACTION is not set to RESTART, you may get
strange CPUSPINWAIT crashes instead of HALT-crashes (like MCHECKPAL or
KRNLSTAKNV). Look at the state of CPU #2 with SDA> CLUE CONFIG
(especially the HALT code).

Valentin,

once your ES40 was hung, did you try to press the HALT button ? If the
system would not react when pressing HALT, this is most likely a
hardware hang, otherwise it could be software and you can force a crash
entering >>> CRASH.

Volker.

Google