Solaris 2.4 Boot server won't cooperate.

Solaris 2.4 Boot server won't cooperate.

Post by Mike Co » Sat, 18 Mar 1995 08:05:40



I've been having *BIG* problems setting up a Network install server for
Sol 2.4.  I've configured a SparcClassic as a boot/install server, and
when attempting to JumpStart another system (Sparc 20) on the same subnet
I get the following symptoms:

        1) The client sends out it's RARP broadcast.
        2) the server's  rarpd sees it and responds.

However, the client never seems to receive the ARP/RARP packet, it just
times out. Details...

    From the server, "in.rarpd -d -a" shows:

        in.rarpd:  RARP_REQUEST for 8:0:20:72:f5:bf
        in.rarpd:  good lookup, maps to 129.100.1.60
        in.rarpd:  immediate reply sent
        in.rarpd:  RARP_REQUEST for 8:0:20:72:f5:bf
        in.rarpd:  good lookup, maps to 129.100.1.60
        in.rarpd:  immediate reply sent
        ...

    "snoop | grep ARP" shows:

       BROADCAST -> (broadcast)  RARP C Who is 8:0:20:72:f5:bf ?
       blizzard.ccs.uwo.ca -> xotic.ccs.uwo.ca RARP R 8:0:20:72:f5:bf is 129.100.1.60, xotic.ccs.uwo.ca

    Finally, on the client:

        Timeout waiting for ARP/RARP packet
        Timeout waiting for ARP/RARP packet
        ...

Has anyone else encountered/solved this one?   Any comments are much appreciated.   Thank you.

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Chris Ch » Sun, 19 Mar 1995 06:49:26


: I've been having *BIG* problems setting up a Network install server for
: Sol 2.4.  I've configured a SparcClassic as a boot/install server, and
: when attempting to JumpStart another system (Sparc 20) on the same subnet
: I get the following symptoms:

- Did you run add_install_client from the server?
- Try boot net at the OK prompt
- Any messages in tftp.log in /var/spool/? ( I think)

--- Chris Tai Chan
    System/Network Admin.
    Wilshire Assocites

:     From the server, "in.rarpd -d -a" shows:

:       in.rarpd:  RARP_REQUEST for 8:0:20:72:f5:bf
:       in.rarpd:  good lookup, maps to 129.100.1.60
:       in.rarpd:  immediate reply sent
:       in.rarpd:  RARP_REQUEST for 8:0:20:72:f5:bf
:       in.rarpd:  good lookup, maps to 129.100.1.60
:       in.rarpd:  immediate reply sent
:       ...

:     Finally, on the client:

:       Timeout waiting for ARP/RARP packet
:       Timeout waiting for ARP/RARP packet
:       ...

: Has anyone else encountered/solved this one?   Any comments are much appreciated.   Thank you.

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Michael Bend » Wed, 22 Mar 1995 12:50:28


: I've been having *BIG* problems setting up a Network install server for
: Sol 2.4.  I've configured a SparcClassic as a boot/install server, and
: when attempting to JumpStart another system (Sparc 20) on the same subnet
: I get the following symptoms:

I'm not an expert on this, but I have setup a few boot/install servers, and
when I've seen the "Timeout waiting for ARP/RARP packet" it usually meant
that I was trying to bring up a client that wasn't in the ethers map (if
using NIS), but in your case, it looks like the RARP is being done
sucessfully, since your host knows your client's ethernet and IP addresses.
Did you use the "addclient" command to add the client, or are you doing it
by hand?

mike

--
----------------------------------------------------------------------------
                                OJ - Who Cares?
----------------------------------------------------------------------------

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Mike Co » Sat, 25 Mar 1995 06:10:37



|>
|> : I've been having *BIG* problems setting up a Network install server for
|> : Sol 2.4.  I've configured a SparcClassic as a boot/install server, and
|> : when attempting to JumpStart another system (Sparc 20) on the same subnet
|> : I get the following symptoms:
|>
|> I'm not an expert on this, but I have setup a few boot/install servers, and
|> when I've seen the "Timeout waiting for ARP/RARP packet" it usually meant
|> that I was trying to bring up a client that wasn't in the ethers map (if
|> using NIS), but in your case, it looks like the RARP is being done
|> sucessfully, since your host knows your client's ethernet and IP addresses.
|> Did you use the "addclient" command to add the client, or are you doing it
|> by hand?

Mike,

I tried adding the new hostname (and ethers entry) both by hand and using
admintool, and used the add_install_client script from the install directory
to set the rest of it up.

Other details...

There are other systems on the subnet running rarpd, but none of them have
the ethers info, so they shouldn't respond to the broadcast (and don't, as
far as I can see with 'snoop').

In /etc/nsswitch.conf (on the install server) I've shuffled things a bit,
so it looks at 'files' first for the hosts, ethers, and bootparams tables
(normally I use dns and nis).

It was suggested that the problem might be caused by using full hostnames,
but I tried changing the hosts table to list the short hostname as the primary,
still no effect.

Another suggestion was that the rarpd on the server was responding to the IP
address of the client, but since the client didn't yet know it's IP address it
wasn't receiving.  This one sounds pretty unlikely to me though.

Thanks for the help.

        - Mike C.

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Michael Bend » Sun, 26 Mar 1995 14:27:06




: |>
: |> : I've been having *BIG* problems setting up a Network install server for
: |> : Sol 2.4.  I've configured a SparcClassic as a boot/install server, and
: |> : when attempting to JumpStart another system (Sparc 20) on the same subnet
: |> : I get the following symptoms:
: |>
: |> I'm not an expert on this, but I have setup a few boot/install servers, and
: |> when I've seen the "Timeout waiting for ARP/RARP packet" it usually meant
: |> that I was trying to bring up a client that wasn't in the ethers map (if
: |> using NIS), but in your case, it looks like the RARP is being done
: |> sucessfully, since your host knows your client's ethernet and IP addresses.
: |> Did you use the "addclient" command to add the client, or are you doing it
: |> by hand?

: I tried adding the new hostname (and ethers entry) both by hand and using
: admintool, and used the add_install_client script from the install directory
: to set the rest of it up.

: Other details...

: There are other systems on the subnet running rarpd, but none of them have
: the ethers info, so they shouldn't respond to the broadcast (and don't, as
: far as I can see with 'snoop').

Do you even see your server responding to the broadcast from the client?

: In /etc/nsswitch.conf (on the install server) I've shuffled things a bit,
: so it looks at 'files' first for the hosts, ethers, and bootparams tables
: (normally I use dns and nis).

I have mine setup this way as well. Is it something non-obvious like a typo
in the ethernet address, hostname or IP address in one of the files? Did you
reboot after adding entries to these files and changing your nsswitch.conf
file? Or, if you didn't reboot, you should probably at least kill and restart
rpc.bootparamd and in.rarpd.

: It was suggested that the problem might be caused by using full hostnames,
: but I tried changing the hosts table to list the short hostname as the primary,
: still no effect.

You mean the problem might have been something like:

        $ cat /etc/hosts
        123.456.789.123 myhost.mydomain.org

and changing it to:

        $ cat /etc/hosts
        123.456.789.123 myhost

didn't fix the problem?

: Another suggestion was that the rarpd on the server was responding to the IP
: address of the client, but since the client didn't yet know it's IP address it
: wasn't receiving.  This one sounds pretty unlikely to me though.

This I don't understand. When the client does a broadcast with it's ethernet
address, your server should send the client it's IP address back to the
ethernet address that the client did the broadcast from. Maybe the ethernet
address of your client is incorrect in your server files?

mike

--
----------------------------------------------------------------------------
                                OJ - Who Cares?
----------------------------------------------------------------------------

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Mario Klebs » Wed, 29 Mar 1995 03:38:18



>I have mine setup this way as well. Is it something non-obvious like a typo
>in the ethernet address, hostname or IP address in one of the files?

I used to enter the ethernet addresses using two nibble
hexadecimal values (08:00:20:00:11:03). After upgrading to Solaris
2.3 (from Sunos 4.3), I had problems with these entries, and it worked
fine, when I removed the leading zeros (now I use 8:0:20:0:11:3). I
don't know, wether it ever worked with leading zeros, since I never
did a diskless boot to install SunOS4, I installed the miniroot when
running the previous system single user. So it's possible, that it
never worked.

This is just a guess, but perhaps it can help.

P.S. in.rarpd does have a -d option. Perhaps it can be usefull.

73, Mario
--

Institut fuer Robotik und Prozessinformatik der TU Braunschweig
Hamburger Strasse 267, 38114 Braunschweig, Germany

 
 
 

Solaris 2.4 Boot server won't cooperate.

Post by Casper H.S. D » Thu, 30 Mar 1995 04:00:00



}I used to enter the ethernet addresses using two nibble
}hexadecimal values (08:00:20:00:11:03). After upgrading to Solaris
}2.3 (from Sunos 4.3), I had problems with these entries, and it worked
}fine, when I removed the leading zeros (now I use 8:0:20:0:11:3). I
}don't know, wether it ever worked with leading zeros, since I never
}did a diskless boot to install SunOS4, I installed the miniroot when
}running the previous system single user. So it's possible, that it
}never worked.

Leading zeros in the ethers map never worked.

Casper

 
 
 

1. Solaris x86 2.4 won't boot after upgrade

I upgraded my 486 system which was running Solaris x86 on the primary

Tomcat I motherboard (Intel Triton II chipset) running a Pentium 133.
Now when I boot off the same IDE drive on the primary IDE controller, I
get pages of BAD TRAP, Page Fault, kernel fault, panic:Page Fault and
finally it locks up.

I figured I would have to reload so I booted off of DU10 and then the
CDROM to reinstall on the IDE drive.  It bombed when it tried a fdisk on
the IDE to create the partitions.  Out of desperation, I loaded 2.4 on a

in Solaris from the SCSI and found I could completely access all of the
IDE Solaris partitions (ie mount /dev/dsk/c0d0s6 /mnt mounted the old
IDE /usr directory).  My questions:

  1. Can I do anything to boot off and use the existing IDE x86 2.4?
  2. If not, then how can I reinstall x86 2.4 on the IDE drive?

Any info is greatly appreciated.
Yatish.

2. Anyone ever get java to work in Mozilla?

3. Linux 2.4-based systems won't boot on SiS-based motherboard

4. All my AuthNames show up as "unknown"

5. Linux 2.4 won't boot without a VGA card

6. Is there a LINUX for me?

7. Solarix_x86 2.4 boot manager and Win'95

8. three nic woes

9. kernel 2.4 won't boot

10. SMP kernel won't boot on Sparc 1000E, get watchdog (2.2 and 2.4)

11. LILO won't boot 2.4

12. Solaris x86 2.4 on ASUS PCI/E-P54NP4 (won't install)

13. Solaris 2.4, HP4 w/postscript, won't print