Nobody logs on to the controller, so it was likely not a MORE/HOLDING
event. It was early in our day (the 16:xx:xx time is GMT), so we were
not at a peak. The controllers have a REL 3000 share with QUICKDSP on.
We rarely hit over 50% cpu utilization on the system, so their being
starved of cpu is not the answer. The controllers show numbers like
this: "CPU 00: Ctime=4 01:44:28  Vtime=0 00:00:01  Ttime=0 00:00:41" in
response to IND USER uid EXP, so they obviously do not consume cpu. We
are not memory constrained. We seldom see any paging. 

I guess that the cause will remain a mystery. Maybe I should blame it on
the OSA card :-) We are still using DTCVSW2 with F00 as the backup
device, so it remains an unknown.

Regards, 
Richard Schuh 

 

> -----Original Message-----
> From: The IBM z/VM Operating System 
> [mailto:[EMAIL PROTECTED] On Behalf Of Alan Altmark
> Sent: Thursday, February 28, 2008 11:21 AM
> To: [email protected]
> Subject: Re: VSWITCH/OSA Problem
> 
> On Tuesday, 02/26/2008 at 06:12 EST, "Schuh, Richard" 
> <[EMAIL PROTECTED]>
> wrote:
> 
> > 16:12:32 HCPSWU2843E A stall was detected for TCP/IP 
> Controller DTCVSW1.
> > HCPSWU2843E It was managing device 0F00 for VSWITCH SYSTEM VM3SW1.
> > 16:12:32 HCPSWU2830I VSWITCH SYSTEM VM3SW1 status is in 
> error recovery.
> > HCPSWU2830I DTCVSW2 is new VSWITCH controller.
> > 16:12:33 HCPSWU2831E CP Controller error 5 for DTCVSW1.
> > 16:12:33 HCPSWU2830I VSWITCH SYSTEM VM3SW1 status is ready.
> > HCPSWU2830I DTCVSW2 is VSWITCH controller.
> > 
> > There was a daylong stream of DTCOSD246I and DTCOSD247I 
> messages. The 
> > only other messages from the controller were:
> > 
> > OSA 0F00 DETACHED DTCVSW1 0F00 BY DTCVSW1 OSA 0F01 DETACHED DTCVSW1 
> > 0F01 BY DTCVSW1 OSA 0F02 DETACHED DTCVSW1 0F02 BY DTCVSW1
> 
> For whatever reason, CP was unhappy with the controller.  
> Perhaps your system is loaded down and the controller is 
> being starved of resources or something is driving it crazy 
> with NETSTATs or it is sitting in MORE/HOLDING or CP READ?  
> The book says to ignore "error 5" (so why display it?).
> 
> Alan Altmark
> z/VM Development
> IBM Endicott
> 

Reply via email to