Wokstation hangs after "Boot Client" - TFTP timeout

[Alexander Skwar]

Hello.

I installed Ignite-UX-11-ALL_C.6.8.152 on a HP-UX 11.11 server and would like to use this system as a Ignite server; ie. I'd like to be able to install other workstations from this system. This server is in a different subnet than the workstations are. Because of this, I installed Ignite-UX on a system which is in the same subnet as the to-be-installed workstation; this system should act as a boot helper.

After having installed Ignite-UX on both server and boot helper, I started the Ignite-UX GUI on the server and selected Actions -> Boot client. I entered the host name of a 2nd workstation (ie. *NOT* the hostname from the boot helper). After a short time, the workstation booted.

When I now have a look at the to-be installed workstation, I see some error messages:

Note: Retrying loadfile command...
* tftp error log follows:
tftp> tftp> tftp> tftp> tftp> Transfer timed out.
tftp>
ERROR: Failed to read "INDEX" file from install server. Check that the install server's IP address is correct and the server has the "Ignite-UX" product loaded and is available via the tftp(1) service.
delete net default: gateway 10.233.16.250
Failed to startup networking, user interaction required.

But tftp *IS* working. In /etc/inetd.conf of the IgniteUX server, I've got:

[10:42:05 vz6tml@dewudb05:~] $ grep tftp /etc/inetd.conf
# Before uncommenting the "tftp" entry below, please make sure
# that you have a "tftp" user in /etc/passwd. If you don't
# have one, please consult the tftpd(1M) manual entry for
tftp dgram udp wait root /usr/lbin/tftpd tftpd -l /opt/ignite /var/opt/ignite

From yet another workstation, which is in the same subnet as the to-be installed WS, I can tftp to the server just fine and fetch the INDEX file:

[11:12:01 vz6tml@tc56:/tmp] $ tftp dewudb05
tftp> get /var/opt/ignite/INDEX
Received 1933 bytes in 0.0 seconds
tftp>

In /etc/opt/ignite/instl_boottab on the boot helper, I've got:

10.233.16.60:::
10.233.16.61:::
10.233.16.62:::
10.233.16.63:::
10.233.16.64:::
10.233.16.65:::

These IPs (and some more) are also in the instl_boottab on the Ignite UX server.

The Ignite server has 4 NICs but only 1 is activated.

When I go to the to-be installed workstation and enter

boot lan.10.233.16.156 install

I get to the install menu. 10.233.16.156 is the IP of the boot helper.

When booting a client from the Ignite UX GUI, how do I make the client use a boothelper and specifically THIS boothelper?

Thanks,
Alexander


Some more data:

On the boothelper:
netstat -rn -> http://vz6tml.pastebin.ca/124895
inetd.conf -> http://vz6tml.pastebin.ca/124898
lanscan -> http://vz6tml.pastebin.ca/124899
instl_adm -d -> http://vz6tml.pastebin.ca/124900
/etc/opt/ignite/instl_boottab -> http://vz6tml.pastebin.ca/124902

On the server:
netstat -rn -> http://vz6tml.pastebin.ca/124903
inetd.conf -> http://vz6tml.pastebin.ca/124904
lanscan -> http://vz6tml.pastebin.ca/124905
instl_adm -d -> http://vz6tml.pastebin.ca/124906
/etc/opt/ignite/instl_boottab -> http://vz6tml.pastebin.ca/124908

[Steven E. Protter]

Shalom,

A few things to look at:

1) You can't boot Ignite client off anything other than its built in NIC. This may have changed, but I didn't get the memo.

2)Check the tail -f /var/adm/syslog/syslog.log on the Ignite server and see if there are errors. Report or correct them.

3) check /var/adm/inetd.sec on the server and make sure its not blocking something important.

SEP

[Alexander Skwar]

Shalom to you as well (although there isn't that much shalom... :( )!

I don't get what you mean with "1)".

2) There's *NOTHING* at all in the syslogs. Neither on the boothelper nor on the ignite ux server. I enabled logging with inetd -l, and the only thing I see on the boothelper is:

Aug 8 13:03:13 tc56 inetd[792]: Connection logging enabled

Because of that, I don't think that the to-be installed client actually reaches the boothelper. That's why I asked, how to tell the client to use a boothelper.

3) The file doesn't exist.

Thanks,
Alexander

[IT_2007]

Make sure that new MAC address for the client is properly linked to /var/opt/ignite/clients/ on ignite server. Also check /etc/exports file on Ignite server which should have directory exported for NFS mounts to the client.

Thanks,
Srini

[Alexander Skwar]

Hi.

The clients MAC isn't listed in /var/opt/ignite/clients/. Should it be? So soon? Even before TFTP transfer works?

Thanks,
Alexander

[IT_2007]

No. I am sorry. I misunderstood. You are trying to take first time setting up client and taking backup so what you need to do is setup .rhosts, add root in the file for Ignite server and see if you can rlogin from Ignite server to client then you can try.

[Alexander Skwar]

rlogin login (actually, ssh is used) is done earlier - thru this channel, the reboot command is sent. After the reboot is done, TFTP is tried.

[Sameer_Nirmal]

You need to have an entry for "bootps" in "inetd.conf" with "instl_bootd" on the Ignite-UX server.
The entry would look like
bootps dgram udp wait root /opt/ignite/lbin/instl_bootd instl_bootd

[Alexander Skwar]

Sameer,

I added the line, you told me to add. Didn't change anything.

Actually, I did NOT expect this to work. To me, it seems that the client is either not using a boothelper at all or maybe not the correct boothelper. Thus my question:

How do I tell the client to use a boothelper and how do I tell the client to use this specific boothelper?

How do I tell, if the client is using a boothelper AT ALL?

Thanks,
Alexander

[IT_2007]

You can interrupt at BCH and scan for devices and boot with boothelper and fill out the information it asks for hostname, ip address, gateway, subnetmask and Ignite Server IP address. If Ignite server is on the same subnet then you can pull the image.

[Alexander Skwar]

Srini_2007,

when I manually enter the information, the client boots. But that's not an acceptable workaround. Especially not, as we used to be able to automatically install clients without going to the workstations.

When everything was working, we used a different server hardware and an older IUX release.

IUX server is on a different subnet (it now is and it always used to be).

Once more: How do I see, if the client is using a boothelper, and if it is, what boothelper the client is using? Is this possible?

Thanks,
Alexander

[Alexander Skwar]

Solution: When I initially started, I (somehow) had a wrong IP adress of the install_tftp_server set. This was shown by running "instl_adm -d". I noticed this, and corrected this by doing "instl_adm -t $correct_ip".

BUT: When I initially booted the client from the IUX GUI, I had the wrong IP set. To rectify this, I had to boot the client into the real OS (11.00) and went to the IUX GUI and did Actions -> Boot Client once more. After I did this, the client booted and I could install it from the IUX GUI.

So, what I had to do, was to boot once again into OS.