guidance request for NIC/drive mapping and target configs

Looking for some guidance to help automate/modularize repeatable deployments ...

1) NIC setup scenario:

- The target server systems have multiple NICs
- The respective MAC addresses of each NIC can be found using BMC interface
- User wishes to seed desired NIC usage by providing target MAC Address(es)
   (e.g. MACAddress1 - dhcp, MACAddress3 - IP/mask,gateway,...)

After (re)reading the docs many times, seems like a possible approach
might be (albeit have yet to succeed doing this)
- generate an autoYaST pre-script with
   - desired MACAddress
   - desired "bootproto"
   - output = mapping to associated "device" name found during boot
- leverage rules, classes, or even asks (to confirm)
   - pull in "interface" fragment, substituting "device" name and
config specifics

Bonus points / challenges
- also seems doable to have "interface" fragments for more complicated setups:
   - bond these two devices
   - bond/bridge these N devices, etc.
- managing firewall settings
- perhaps using post-scripts (for anything beyond the one needed NIC
for install)

2) Disk setup scenario

- The target server systems have multiple drives
- The respective drive (model, serial number, size, etc) of each drive can
be found using BMC interface
- User wishes to seed desired partitioning by providing a physical attribute
   (e.g. serial number - root, serial number - data)

As above, after (re)reading the docs many times, seems like a possible approach
might be (albeit have yet to succeed doing this)
- generate an autoYaST pre-script with
   - desired serial number
   - desired partitioning type/label
   - output = mapping serial number to associated "device" name found
during boot
- leverage rules, classes, or even asks (to confirm)
   - pull in "drive" fragment targeting mapped "device" name and
config specifics

Bonus points / challenges
- more difficult if storage controller has created LUNs of multiple drives
- perhaps using post-scripts (for anything beyond the one needed drive
for install)

Thanks in advance for any guidance and suggestions on how to implement
this, for a large set of systems that are to be (re)installed over
time and may not be entirely homogeneous. The goal though is to only
pre-seed those specifics (that can be found outside the installer) and
generate/rely upon such mappings into for the fragments to provide a
good level of consistency (vs. just using the interactive installer
interface). Of course, either post-install or typical
configuration management options could be layered on after the fact to
complete the system configuration, beyond needing one (desired) NIC to
access packages/repos and one (desired) drive to install upon.
-- 
To unsubscribe, e-mail: [email protected]
To contact the owner, e-mail: [email protected]