On Jul 14, 2011, at 6:41 PM, Rob Lagunas wrote: > Jesse, > > Thanks for the response . . . my comments are below. > >> One of the more noticeable changes with 169 onward is the migration of the >> root TFTP directory from /tftpboot to /etc/netboot. That's the place to look >> for validation of your bootfiles, configurations, etc. >> >> Note if you're also running Jumpstart on the same server, there is some >> manual work that's required to relocate the files in /tftpboot over to >> /etc/netboot, then replace /tftpboot w/ a symlink to /etc/netboot. If there >> were non-AI files in /tftpboot, the installadm-convert.py utility should >> have informed you of this step. > > The system I'm using is dedicated to AI services only; just the way the lab > AI servers are configured. > >> A second change related to what you're seeing is that we no longer have >> bootfiles named as "I86PC*", but rather we have default services and >> client-specific bootfiles named by their Client ID (which is "01" + client >> mac address). The conversion utility should have created an >> /etc/netboot/<client id> bootfile for each client-specific bootfile in >> /tftpboot. And, any new client bindings that you add from 169 onward via >> create-client will create a Client ID bootfile in /etc/netboot. > > I see this now and re-ran 'installadm create-client' for my client and would > have hoped the client-ID specific menu.lst file would have the proper > entries, but it doesn't. > > Does this mean I need to delete the service and then re-create it?
You should not have to, no. With that create-client, you should have gotten a client-ID bootfile created in /etc/netboot which points at the snv_166's service bootfile (/etc/netboot/<service name>/boot/grub/pxegrub), and also a menu.lst.<client ID> file in /etc/netboot to go along with it which should reference that service (e.g. kernel$ /<service name>/platform... ). Is that the menu.lst you're referring to? If so, can you tell me which part looks broken? It could be a known bug (there are a couple in 169). If you want to cut to the chase on this, I'd be happy to log into the system and poke around a bit, you could send me an off-alias email w/ login info if that suits. Best Jesse > >> Let me know if that gets you going on the right track. If not, we can >> continue here, or we can get on the phone quickly if that suits. > > I think I'm closer, thanks! > > Rob > >>> Just updated a test AI server from 165 to 169 and ran >>> /usr/sbin/installadm-convert to get the AI server running again. When I try >>> to install an X4200 with a pre-existing 166 AI service, I see the following >>> on the client's console: >>> >>>> kernel$ /I86PC.Solaris-2/platform/i86pc/kernel/$ISADIR/unix -B >>>> console=ttya,install=true,install_media=http://10.134.7.48:5555/pool01/ai/services/sol-11-dev-166-ai-x86,install_service=sol-11-dev-166-ai-x86,install_svc_address=10.134.7.48:5555 >>>> loading '/I86PC.Solaris-2/platform/i86pc/kernel/$ISADIR/unix -B >>>> console=ttya,install=true,install_media=http://10.134.7.48:5555/pool01/ai/services/sol-11-dev-166-ai-x86,install_service=sol-11-dev-166-ai-x86,install_svc_address=10.134.7.48:5555' >>>> ... >>>> cpu: 'AuthenticAMD' family 15 model 1 step 2 >>>> [BIOS accepted mixed-mode target setting!] >>>> module$ /I86PC.Solaris-2/platform/i86pc/$ISADIR/boot_archive >>>> loading '/I86PC.Solaris-2/platform/i86pc/$ISADIR/boot_archive' ... >>>> >>>> Error 20: Multiboot kernel must be loaded before modules >>>> >>>> Press any key to continue... >>> >>> There are no /tftpboot/I86PC* directories/mountpoints and no entries in the >>> vfstab. I'm aware of bug 7040399 and 6999502 (no nge adapters on the >>> client), but wonder if my AI server is in this state because of >>> installadm-convert. >>> >>> Can anyone help me understand what's going on? >>> >>> Rob >>> _______________________________________________ >>> caiman-discuss mailing list >>> [email protected] >>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >> > _______________________________________________ caiman-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
