I had this setup for our z/VM test machine ages back when we ran out of OSA
addresses on a 9672. But I no longer find the config files used then. What
I remember is
- you need to turn on PROXYARP in ASSORTEDPARM in the TCP/IP stack of VM
- you need to specify also the IP address of the guest below the HOME
statement
Later we had to experiment with some Linux guests and the z/VM host didn't
have free OSA addresses at that time, so I prepared a similar setup. This
is what I find back (from 2004)
This is what I did in VM's TCPIP stack to connect the VCTC of a guest
(somewhere in 2004). It appears we were also preparing the move from Token
Ring to Ethernet, so two setups)
DEVICE OSATRDEV LCS 180
LINK TRCCM IBMTR 0 OSATRDEV
DEVICE VCTCLIX0 CTC 800
LINK CTC_LIX0 CTC 0 VCTCLIX0
DEVICE VCTCLIX1 CTC 810
LINK CTC_LIX1 CTC 0 VCTCLIX1
DEVICE OSA1180 LCS 1180
LINK OSA1180L ETHEROR802.3 0 OSA1180
DEVICE VCTCL1 CTC 850
LINK CTC_L1 CTC 0 VCTCL1
DEVICE VCTCL2 CTC 860
LINK CTC_L2 CTC 0 VCTCL2
....
HOME
192.168.9.2 TRCCM
10.200.0.53 CTC_LIX0
10.200.0.53 CTC_LIX1
10.48.10.88 OSA1180L
10.46.55.6 CTC_L1
10.46.55.10 CTC_L2
GATEWAY
; (IP) Network First Link Max. Packet Subnet Subnet
; Address Hop Name Size (MTU) Mask Value
; ----------- ------------ ------- ----------- ----------- --------
10 = TRCCM 2052 0.255.0.0 0.200.0.0
10.200.0.230 = CTC_LIX0 1492 HOST
10.200.0.231 = CTC_LIX1 1492 HOST
10 = OSA1180L 1492 0.255.0.0 0.48.0.0
10 = CTC_L1 1492 0.255.255.252 0.46.55.0
10.46.55.4 = CTC_L1 1492 HOST
10.46.55.8 = CTC_L2 1492 HOST
I hope this helps a bit.
2010/8/26 KEETON Dave * SDC <dave.kee...@state.or.us>
> Greetings Listers,
>
> I'm wrestling with a routing issue between a real switch, a TCPIP
> service machine and another VM (on the same system) connected through a
> vCTCA.
>
> Assuming the formatting holds, here's a rough drawing of the connection:
>
> |-------------|
> | Real Switch |
> |-------------|<----->|--------|
> | OSA |
> |--------|<---->|-----------| .10
> | VSWITCH
> |<---->|--------|
> |-----------| |
> TCPIP | .11 vCTCA .12
>
> |--------|<---------------->|---------|
>
> | guest2 |
> |---------|
>
> We've got a Cisco switch connected to an OSA-Express2, which has a
> VLAN-unaware VSWITCH attached. TCPIP is attached to the VSWITCH. Also
> attached to the TCPIP machine is the .11 address of the left side of the
> vCTCA connection. Connected to the 2nd VM on the other end of the vCTCA is
> address .12. This was configured using the following IFCONFIG statement:
>
> IFCONFIG CTC0 159.121.93.11/25 VCTC 800 OLIS 1800 PTP 159.121.93.12
> PORTNUMBER 0
>
> Before configuring and activating the vCTCA connection, this is what
> NETSTAT GATE displays:
>
> Known IPv4 gateways:
>
> Subnet Address Subnet Mask FirstHop Flgs PktSz Metric Link
> -------------- ----------- -------- ---- ----- ------ ------
> Default <none> 192.168.93.1 UGS 1500 <none> DV1000
> 192.168.93.0 255.255.255.128 <direct> UT 1500 <none> DV1000
>
> IFCONFIG CTC0 192.168.93.11/25 VCTC 800 guest2 1800 PTP 192.168.93.12
> PORTNUMBER 0
>
> The NETSTAT output from this is:
>
> Known IPv4 gateways:
>
> Subnet Address Subnet Mask FirstHop Flgs PktSz Metric Link
> -------------- ----------- -------- ---- ----- ------ ------
> Default <none> 192.168.93.1 UGS 1500 <none> DV1000
> *192.168.93.0 255.255.255.128 <direct> UT 9216 <none> CTC0*
> 192.168.93.0 255.255.255.128 <direct> UT 9216 <none> DV1000
> 192.168.93.12 HOST <direct> UHS 9216 <none> CTC0
>
> What happens is pings work from .12 to .11 and .10, but not to the GATEWAY
> of .1. I'm wondering if the problem is packets leave the system to the
> switch, but cannot find their way back because there are two identical route
> statements on two different interfaces. Is there a way to remove the
> dynamically added route listed here in line 2? Everything needs to come back
> via DV1000, I would think.
>
> Thanks in advance,
> Dave Keeton
>
--
Kris Buelens,
IBM Belgium, VM customer support
