I'm not sure how the opensolaris tools for bfu work, but given that Clearview-UV made many changes to bfu, I'm cc-ing the clearview mailing list to see if they have any suggestions.
Clearview-experts: Jason is trying to convert the nge driver to brussels, but seems to be running into trouble with post-attach (even without his changes). The original thread is at http://mail.opensolaris.org/pipermail/driver-discuss/2008-February/005096.html Do you have any insights into what he should look at? --Sowmini On (02/16/08 20:27), Jason King wrote: > > On Feb 16, 2008 5:10 PM, Jason King <jason at ansipunx.net> wrote: > > > > On Feb 16, 2008 4:46 PM, <sowmini.varadhan at sun.com> wrote: > > > On (02/16/08 15:11), Jason King wrote: > > > > > > > > Actually, it's the 'postattach' (though I'm not exactly sure what that > > > > actually does) that is saying is not working. > > > > > > nothing in your webrev should affect postattach- can you generate > > > a baseline (i.e., an image without any of your changes) that works? > > > Since nv83 had a lot of major projects like Clearview-UV, I wonder if > > > there is some other error like using the wrong bfu script, or something > > > about the process of generating the image itself that is messed up? > > > > No problem.. it'll take a couple of hours. I will pull down the > > latest ON copy (changeset 6029:467926d47d94), and build that and see > > what happens. > > > > Well it looks like even just the plain bfu is giving the same error. > For completeness (I wouldn't call me self an bfu expert), what I did > was this: > > hg pull > setup setenv.sh > ran nightly setenv.sh > made copy of existing b81 / into new fs, mounted as /a > bfu /path/to/archives /a > bootadm update-archive -R /a > umount /a > edit grub > reboot > > I glanced through the files that it complained as having a conflict > due to local modifications), but nothing stood out (it was stuff like > me setting 2a as the the default crypt algo, etc). Might there be > specific files that could be related to a postattach failure I should > examine in closer detail (just to cover all bases). > > I didn't see anything in the flag days on opensolaris.org that seemed > relevant either. > > If there is anything you can suggest to try to debug it, I'm more than > willing to try.
