On Monday, 05/05/2008 at 08:03 EDT, Shimon Lebowitz <[EMAIL PROTECTED]>
wrote:
> Hi,
> I have occasionally had an OSA go down, as happened today.
> Here are the messages I got from user TCPIP on the operator
> console:
>
> ***** 05/05/08 *****
> 06:37:34 DTCOSD309W RECEIVED ADAPTER-INITIATED STOP LAN
> ***** 05/05/08 *****
> 06:37:34 DTCOSD082E OSD SHUTTING DOWN:
> 06:37:34 DTCPRI385I DEVICE OSA-B5:
> 06:37:34 DTCPRI386I TYPE: OSD, STATUS: READY
> 06:37:34 DTCPRI387I ENVELOPE QUEUE SIZE: 0
> 06:37:34 DTCPRI388I ADDRESS: B500
> 06:37:34 DTCQDI001I QDIO DEVICE OSA-B5 DEVICE NUMBER B502:
> 06:37:34 DTCQDI007I DISABLE FOR QDIO DATA TRANSFERS
>
> I would like to prepare a routine to be triggered when these
> messages are seen by PROP. My problem is that the message
> lines which describe a problem ('stop lan', 'shutting down',
> 'disable') do not include the device address or link name, and the
> lines with the device do not mention the problem, but PROP
> works on the basis of a single line.
>
> How do other people deal with this?
1. Add AUTORESTART to the DEVICE definition
or
2. Use a virtual NIC on a VSWITCH. The VSWITCH has better error recovery
than the stack. It also gives you the ability to have a second OSA on
stand-by without having to code dual interfaces in the stack.
In any case, you don't want PROP to handle this. A "STOP LAN" is the
result of a switch failing or being recycled, or a cable pull. A manual
restart won't help.
Alan Altmark
z/VM Development
IBM Endicott
