Post by YTC#1Post by Cydrome LeaderPost by YTC#1Post by Cydrome LeaderPost by Richard B. GilbertPost by Cydrome LeaderPost by Richard B. GilbertPost by b***@gmail.comPost by Richard B. GilbertPost by b***@gmail.comPost by Richard B. GilbertPost by b***@gmail.comAny ideas what to check....
/etc/bootparams seems to be doing its job because client boots up ok
and seems to find the correct sysidcfg.
However, for some reason its saying rules is wrong.
What to check first?
Your rules file?
BTW. Should mention it is there....
Is are the contents syntactically and semantically correct?
I dimly recall that there are two files involved: sysidconfig is one and
the other is your rules file.
The "Advanced Installation Manual" goes into all this in some detail
Thing is I saved a few of the files and the complete /jumpstart
directory off another working js server.
It seems to find the correct profile and then fail to find the
jumpstart directory? Weird - because I think I'm right in saying this
is all identified in the bootparams file, isnt it?
Don't know! The one time in my career that it would have been useful, I
hadn't a clue. Somebody showed me how to install from the local CD
drive. I repeated the steps for eight machines, one at a time. Since
then I've read TFM but had no need to do mass installations.
The best advice I can give you is RTFM and follow the directions.
jumpstart is picky and overly complex monster. On x86, it's just
worthless.
always add the -v flag to your boot install whatever command so you can
see what's happening and where it blows up.
You're at least able to get an address, and boot the kernel off your
jumpstart machine if you're getting rules errors- this is good.
If Jumpstart fails to work as documented, you need to complain to Sun.
I don't know how seriously they take reports/complaints from people who
are not paying customers. I'm fairly sure that they will fix it
they don't care if paying customers complain. they don't even care if
sales reps trying to peddle new hardware get complaints during
trial runs. Then of course, sun fired off their sales people, so it's
not important anyways anymore.
Post by Richard B. Gilbertsomeday. I couldn't even attempt to guess what priority the problem
would be given unless the complaint came from a customer paying support.
been there done, that, I have a long list of support contracts here.
need a list of cases about jump start on x86 being a total piece of shit?
Which bits in particular are causing trouble ?
By using a non-sun DHCP server, we finally got a kernel to boot. rules
never worked, ever. Nobody at sun could get it to work either, which
Which is an outright slur, no one you asked at Sun gave you the right
anser, as I keep pointing out JET (which is also the engine behinbd N1 and
XvM JumpStarts can do all you requested so far, including using non Sun
DHCP. And it was written and supported by Sun staff.
I should not have to use a side project tool to jumpstart a machine. If I
do, the process is just garbage.
Jet isn't a flawless tool either by the way.
Post by YTC#1as they keep changing all the secret paramaters and
Post by Cydrome Leadersettings every 9 seconds with their junior attempts as PXE booting.
PXE booting is a bit of a nightmare, very few seem to understand it well.
Mike Ramchand has done a lot of work in that area to simplyfy the usage
for JumpStart.
PXE booting in ONLY a nightmare when sun is involved.
microsoft has had a RIS service forever, it works perfectly. It doesn't
even require registry hacks to get working.
you hit f12 (or whatever) to netbook your client, possibly login to a
small DOS looking program that loads, and come back to a fully installed
workstation.
go bother some IT guy to see how it works, and has worked for like 10
years now.
It's really easy stuff. You don't have to add console flags to grub menus
and stupid shit like that.
linux people are mostly retarded, but they sure like reinstalling their
machines every 15 minutes and have PXE booting all figured out too.
Last time I looked, the FreeBSD people just didn't get it either. but
their entire project is becoming increasingly obsolete anyways which is
too bad.
Post by YTC#1Post by Cydrome LeaderTo this day, there's still no accurate or even close to upto date doc on
jumpstarting sun stuff with PXE where the commands as they show actually
work. There's some "blueprint" from 2006 or so. it's completely useless.
Ok, I'll go for that one, but bear inb mind that PXE booting isnot Sun
technology.
Sun has no problem making and touting intel hardware, so mayb ethey need
to attend some seminar at intel HQ about PXE.
Post by YTC#1Post by Cydrome LeaderSun's attempt at remote console for their own x86 hardware is a complete
joke as well.
Depends which one you use :-)
But as I always prefer to stay away from GUIs I seem to have less
problems, and the emote console has nothing to do with JumpStart.
Sun remote consoles on their PC harware is completely bogus. If somebody
leaves a monitor cable connected to the server, you can't grab the
non-serial consoles remotely, vs. HP's ILO where both people can type and
see the same thing at the same time as their remote console REALLY is the
same as the hardare one.
Post by YTC#1Post by Cydrome Leaderwan booting sparc machines with newer firmware works better, but only if
you use JET to make the 6000 character long commands.
Aha, so you have played with it then :-)
By and large, WAN boot is a PITA, and not that many *really* need it, a
combo of DHCP/NFS often works better.
Anything works better than obsolete nonsense like RAPR servers and
symlinks in some tftpboot directory with lists of mac addresses.
Any time I have to deal with mac addresses, I feel like it's early 1990s
and I'm installing the Novell client on some PC running DOS, or I'm
configuring a JetLan print server for some "letter quality" dot matrix
printer.
You also mentioned you prefer CLI stuff over a GUI, so that rules out
using the sun DHCP server. That thing is a strange hunk of junk. MS and
ISC did a far better job. Sun should just drop their attempt and move on.
Oh wait, maybe a team of bored sun people in the UK office wrote a front
end to the command line for the sun DHCP server that inputs and outputs
sensible values because the people who wrote the DHCP server never
bothered to do it themselves.
Post by YTC#1Post by Cydrome LeaderIf anybody can explain why a MAC address and hostname have to be used at
least 50 times per client, I'd love to hear the logic behind that.
Again, not sure what your problem is here, you supply it *once* to a
JumpStart tool like JET. The tool then handles how and when they get
handed to the client.
I think 50 is an over exaggeration.
ok how about 35 times?
To any nay-sayers, just watch how you net install windows.