Getting Started, Documentation, Tips & Tricks

Hi,

I'm trying to do a network install on a clean (ex PC) disk in a Fuel (from FreeBSD) using 6.5.30 from supportfolio, and the rest from base, copied (using IRIX) to disk, and the copied over to the FreeBSD box.

I do have the CDs, and I do have a working CD drive in the Fuel, and I can do a CD install without any problem at all (which is how I copied the contents of the CDs) however I'm trying to learn how to do a network install, I like doing "bare metal" network installs

On the FreeBSD box I've started with two directory trees setup in /usr /tftpboot with just fx.64, sa, unix.IP35 and uni_0004.ip3 and /sgi_os_software with all of the copies of teh CDs (full set of from 6.5.14, copies of the overlays from 6.5.22 and copies of the downloaded 6.5.30)

I've setup dhcp, tftp (with -s /usr/tftpboot) and rsh (bootp and dhcp don't actually seem to be needed if you set netaddr, netmask, and srvaddr in the PROM

I can boot fx.64 and partition the disk



I can then boot sash



I can also boot uni_0004.ip3



but, it then complains about there being no file system, which is valid as mkfs hasn't been run.

If I try to run install and select a remote dir it produces the following error (debug is on) when trying to start the same version of SA that I can boot as above



Which again I'm guessing is to do with the fact that there isn't a filesytem on the disk.

I know I'm either probably doing something really stupid, or missing out a basic step, but I'm now at the can't see the wood for the trees stage on this and would really appreciate any help

Cheers

sholdn't dest be dksc(0,1,0) ?

Hi,

I agree, but its not me selecting it :-)

I get the same result with 6.5.30 and 6.5.22

Full trace below

thermionic wrote:

Code: Select all

...
need to seek backward to offset 64 (currently at 266768).
Closing and reopening file
bootpclose: ctlr 0
tftpabort: TFTP send ERROR packet


I don't know exactly what's happening here, but I don't think that's good.

thats just the end of the TFTP connection

and I've just fixed it

I was doing all of this over console, on a whim I plugged in a monitor, keyboard and mouse

and it worked perfectly

I'll put something on the wiki about this

in precis, "don't try and install on a clean disk on a fuel using the serial console, use a local keyboard and monitor"

thermionic wrote: "don't try and install on a clean disk on a fuel using the serial console, use a local keyboard and monitor"

Was this because the disk did not have an SGI style disk-label? If you had booted into sash over TFTP you should have been able to re-label the disk.

porter wrote:

thermionic wrote: "don't try and install on a clean disk on a fuel using the serial console, use a local keyboard and monitor"

Was this because the disk did not have an SGI style disk-label? If you had booted into sash over TFTP you should have been able to re-label the disk.

well that's an explanation. i assumed the disk was prepared *properly* before.
how often did we have that? that'd be a hot candidate for the wiki. a category like 'common errors' or something.

I thought that you labled the disk in fx not sash, did I miss something ?

It was interesting going through it. I did discover that there are quite a lot of environment variables you can set in the PROM. And certainly with a Fuel, you don't need a BOOTP server or DHCP server to do a network boot.

The TFTP setup was trivial. I've booted fx, sash and the miniroot from several TFTP servers, including (gasp) the Windows2003 one :-)

Getting RSH to work the way that inst wants it to though, is proving a little (a LOT) harder, I've got a nasty feeling that this will take a while.... FC6 just wasn't having any of it, FreeBSD6.1 also is proving difficult. I can RSH onto the server, but getting a remote command to work is proving harder. However from what I've seen of IRIX so far, its really not intended as a secured OS, its designed to be inside a secure environmet </flame>

What I would really like, would be a way of logging the on screen output from when I booted the miniroot on the console (not the serial console) to see if it is also installing it to "dksc(0,1,1)" (the raw disk) rather than to dksc(0,1,0) the XFS partition. Biut as it doesn't have a network interface at that point in time then it can't be redirected to syslog, and as there isn't a formatted disk then it can't go to disk, and as at the end of it it boots the miniroot, then I'm guessing that putting it into memory isn't going to work either. I'll just have to wait untill the Origin200 becomes free and try it on that.

list init 128c000 12ff654 73654
mrboot
source : bootp()192.168.53.171:/30/sa(mr)
dest : dksc(0,1,1)

What's the problem here? Perhaps I'm reading it wrong, but it looks like it's copying the miniroot over to partition 1-which is what it should be doing.
Following this, if everything goes the way it should (which it isn't), your IRIS should then commence to boot from the installed miniroot and start inst to install IRIX to partition 0. The miniroot is a complete filesystem unto itself that just needs to be copied raw into what used to be the swap partition, so having a filesystem doesn't matter (it would just be clobbered by the miniroot).

Not skilled at IRIX netinstalls, but doesn't what happens next involve fetching the kernel from /dist/miniroot/unix.IPxx and integrating it in with the miniroot? Perhaps it's falling down here at the remote end. Of course, this would depend on whether the PROM has successfully copied the common miniroot yet.

SAQ wrote: Perhaps I'm reading it wrong, but it looks like it's copying the miniroot over to partition 1-which is what it should be doing.

But if the disk did not have a label/volume-header it wouldn't know where the swap partition was supposed to be....

Hi,

The only change between failing on serial console, and working on the console, was changing where I was running it from.

When I used the console I went into the command monitor, did a "setenv debug 1" then went to "install software" and selected remote etc.

Cheers

thermionic wrote: The TFTP setup was trivial.

It's called Trivial FTP for a reason

Getting RSH to work the way that inst wants it to though, is proving a little (a LOT) harder, I've got a nasty feeling that this will take a while.... FC6 just wasn't having any of it, FreeBSD6.1 also is proving difficult. I can RSH onto the server, but getting a remote command to work is proving harder. However from what I've seen of IRIX so far, its really not intended as a secured OS, its designed to be inside a secure environmet </flame>

Two things:
1) The installation process utilizes r-* commands (rexec, rsh, ...). The reason you're not getting an love from an FC6 box is that e.g. "rsh ls *.idb" gives a different output then what the IRIX miniroot expects (something with XPG4 compliant wildcard expansion, 'color'-ls etc. etc.). Network installations from another SGI are a breeze. I used to run a Linux based install server but gave up. I'm moving it to a Challenge S (which still NFS-mounts the media from the Linux box).

2) Security: yes, this install method is from the days that the internet was still a friendly place Look at it this way: you can install any SGI (including the oldest 4D that otherwise need to be booted from quarter inch tape) over the network from any other for any IRIX version (3.x ... 6.5.xx) without modifications to the install server. I'd say this install server usually lives in the sysadmins room on a private network segment.

You can improve security of your IRIX box by installing IPFilter ( http://www.sgi.com/products/evaluation/ ... er_3.4.27/ )

FreeBSD rshd is now working with inst :-) It needed several changes to /etc/pam.d/rsh to get it "open" enough.

Then a quick change of sh to ksh and inst can see the distributions, YAY !

removed the conflicts, typed go, and its doing its install!

I'm not going to bother with ipfilter, but its good to know its available.

I am still curious as to why it was failing with the TLB exception when running from the serial console.

porter wrote:

SAQ wrote: Perhaps I'm reading it wrong, but it looks like it's copying the miniroot over to partition 1-which is what it should be doing.

But if the disk did not have a label/volume-header it wouldn't know where the swap partition was supposed to be....

He said he labeled it (or at least booted fx.64, and why else use standalone fx?)

Perhaps I was confused by the use of the term "raw disk" - I interpreted raw disk to be the whole partition (10 in IRIX) rather than swap (1 in IRIX).

thermionic wrote: FreeBSD rshd is now working with inst It needed several changes to /etc/pam.d/rsh to get it "open" enough.

Then a quick change of sh to ksh and inst can see the distributions, YAY !

Hi Thermionic,
I guess that you've once found a solution to my actual problem ... Could you please tell me more about the changes you've applied to rsh and/or rshd to have a network installation from FreeBSD work ?
Thanks in advance,
PB

thermionic wrote: FreeBSD rshd is now working with inst It needed several changes to /etc/pam.d/rsh to get it "open" enough.

Now that's starting to sound like a wiki article!

josehill wrote: Now that's starting to sound like a wiki article!

As soon as I've managed to net install my box, I'll publish my experiences on this subject in the wiki. But for now...
PB

Hi PB

I'll fire up the FreeBSD box tomorrow (little to drunk to do it now), and go through some of my assorted notes to see what I did back then to get it working.

Cheers

thermionic wrote: (little to drunk to do it now)

Hello Thermionic,
Thank you very much, but I think I've solved my problem (see this post ).
In the next few days, I'll post an article in the wiki, to explain all the steps one have to perform to net-install a box. My contribution will be about installation from FreeBSD. Other people may add contributions about installation from other systems.
Your posts in this thread have helped me a lot. Thanks again.
Regards,
PB

Hello,
Wiki article about net-installing an Octane from a FreeBSD box added here !
HTH,
PB