krowitz@RICHTER.MIT.EDU (David Krowitz) (09/07/89)
Here's a new one ... I just tried to boot one of my DN3000's (normally running SR9.7.1) off of a DN460 running SR10.1 so I could test out some DN3000 specific software under SR10. I copied the /sau8 directory from the AA machine over to the DN460 and then booted the DN3000 diskless off of the DN460. The machine booted just fine, but when I try to login it pops up my two initial windows and they are still-born. I never get a shell prompt, typing control-Q gets me a "level 2 process not found" message, and typing control-N closes the window. Doing a "cp /com/sh" or a "cp /bin/csh" DM command also pops up a dead window. So ... just for tricks ... I tried booting my DN560 (my AA machine) off of the DN460. I copied the /sau5 directory over and then tried to boot the machine. Much to my surprise, it would not boot. It starts loading the OS and then gets a "boot error: uid request failed" message just after the low, high, and start addresses are printed. Are these known bugs? Are there known fixes? Am I just doing something stupid? -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)
dbfunk@ICAEN.UIOWA.EDU (David B Funk) (09/08/89)
Re the diskless node booting failure with "boot error: uid request failed" error:
Yes there is a problem with netman at sr10.1 and diskless node
booting. The really strange part of it is it only occurs when you
do a forced partnering. If you partner via the "diskless_list" file
then things work OK. IE the problem only shows up when you use the
MD command "DI N node-id" and then try to boot. If you edit the
"diskless_list" file on the sr10 disked node and then let the
diskless node auto-partner then its OK. I've found 3 different
work-arounds for this problem.
1) Just use the "diskless_list" and auto-partner.
If you can't do that, or don't want to because it's a temporary
partnering, then there are 2 options.
2) do the "DI N node-id", then do a "LD", and then try to boot.
For some strange reason, doing the MD command "LD" to get a
directory listing from the disked node will make the boot
succeed (most of the time). IE:
> DI N 12345
> LD
> EX DOMAIN_OS
3) If things are really tough, (IE the "LD" trick won't do it),
then run netman in debug mode on the disked node. For some
strange reason if netman is running in debug mode then it's
better behaived. To run netman in debug mode, invoke it in a
window and give it the "-db" option.
Be aware, if you have this booting problem then it's going to leave some
strange trash processes on your disked node after the boot request fails.
After you get the "boot error: uid request failed" error, do a "/com/pst -l1"
command on the disked node. You'll see some strange "level 1" process
appear. An additional one of these will appear for each failed boot attempt.
I don't know what the ramifications of these things are but I don't like
them. I'd reboot the disked node to get rid of them if I were you.
Dave Funk