Hi all,
I have been working with the zPDT/ZD&T platform for about a decade now. I have always used the Parallel Sysplex environment and when we first configured this, I set up z/VM to have full network access and set up VSWITCH for the z/OS guest images to use to access the network.
The devmap configuration is as follows:
name awsosa 0009 --path=F0 --pathtype=OSD
device 0400 osa osa --unitadd=0
device 0401 osa osa --unitadd=1
device 0402 osa osa --unitadd=2
name awsosa 0019 --path=F1 --pathtype=OSD
device 0404 osa osa --unitadd=0
device 0405 osa osa --unitadd=1
device 0406 osa osa --unitadd=2
I use the first manager for z/VM and the second for the VSWITCH and have two NICs attached/defined to the system. This was done originally for load balancing, but it worked, so I kept it.
The original configuration was a DELL Server running SUSE and this environment worked perfectly for 8 years. A few years ago, we migrated to an on prem cloud and Ubuntu as the linux host operating system. Again, the devmap and networking remained the same and worked perfectly. Users were able to access z/VM and the z/OS images via hostnames and no port forwarding was needed. The users never knew if they were on a zPDT/ZD&T or a real z/VM guest image.
A few months ago we started a migration to the Google Cloud, specifically the CGE environment. I have yet to be able to get this configuration to work within GCE. Has anyone here attempted to do the same thing?
I can provide the GCE networking configurations, but didnt want to overload with useless information. I have spent time with Google support and many, many hours testing various configurations. My belief is that it is the routing that Google is doing that is causing the issues, the subnet mask for every interface is 255.255.255.255 and there is no real "static" IP that can be reserved, you have to define an alias for the guest systems.
Thanks in advance,
Steve
------------------------------
Stephen Norris
------------------------------