I've got 2 z/VM lpars and run a Linux server in each lpar that should be
able to service common ip address.  Both lpars are connected via double OSA
cards to a real switch.

         VM1                                  VM2
/==================\                 /==================\
|                  |                 |                  |
|   /----------\   |                 |   /----------\   |
|   |  Linux1  |   |                 |   |  Linux2  |   |
|   \----+-----/   |                 |   \----+-----/   |
|        |         |                 |        |         |
|        |         |                 |        |         |
|   /----+-----\   |                 |   /----+-----\   |
|   | vswitch1 |   |                 |   | vswitch2 |   |
|   \--+----+--/   |                 |   \--+----+--/   |
|      |    |      |                 |      |    |      |
\=====OSA==OSA=====/                 \=====OSA==OSA=====/
       |    |                               |    |
       |    |                               |    |
       |    |                               |    |
    /--+----+--\                         /--+----+--\
    |  switch1 |-------------------------|  switch2 |
    \----------/                         \----------/


All IP addresses are in the same subnet and the vswitches are connected to
access ports on the switches.

Normally Linux1 has the common IP address, but when it becomes unavailable,
Linux2 should take over.
This process works, when Linux1 can normally terminate, or when it gets
logged off and CP cleans up its OSA IP registrations.
When it breaks, for instance with a kernel panic, then the Linux1 server
retains its registered IP addresses, so Linux2 should forcefully obtain the
common IP address.
To be sure that that happens, the common IP address is registered in Linux
with the "qethconf ipa add <ip_address>" command.

In the situation, were the take over is forcefully (eg, Linux1 has the IP
address still registered), something strange happens.
On VM1's vswitch and OSAs, the common IP address remains registered.
But on VM2, the vswitch shows the IP address registered, but with a "local"
flag and the console of the vswitch controller shows the familiar "Return
code E00A from SETIP for IPv4" error, indicating it could not register the
IP address to the switch.

We are running z/VM 5,2 and the VMLAN level is VM63989.
The OSAs are OSA-Direct Express Gigabit Ethernet types and the vswitches are
defined with only the RDEV parameters.

Does anyone know how to make the IP take over being forwarded to the real
switches and then to the other vswitch?

Ronald van der Laan

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to