I have an interesting conundrum with jumpstart, to which I need a resolution AND hope I may influence development of the new installation tools to make this work more... usefully.
Take a simplified environment of a jumpstart server and a new server to be installed. There are 4 networks, both servers are on all 4 - production, backup, provisioning, and admin. In order to fully automate the configuration through jumpstart, there are network clauses for all 4 interfaces in the sysidcfg, with the production network flagged as 'primary' and having a default route set. The jumpstart is configured to happen on the provisioning network, and when kicked off does indeed tftp and start via the provisioning network. When the sysidcfg file is read, however, the interface marked 'primary' is brought up and all others brought down. Since the provisioning network is not routed, the install then hangs trying to mount the installation image. Is there any particular reason the sysidcfg file is used on the spot, rather than just being used to configure the OS image being installed? I've specified which network I want the installation to run on via 'boot net1 -install' but the installation seems to think it is smarter than I am. Unfortunately, it fails because of that. This has finally explained for me why the N1 provisioning tools I tried to make work ages ago don't work when installed as recommended - with their own provisioning network. --- Joel Lord Manager of Research and Development, Information Systems Garnet River LLC www.garnetriver.com -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/install-discuss/attachments/20090205/055c0d98/attachment.html>
