Re: PXE+DHCP/BOOTP+Kickstart problem

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



We had similar problems on 7.2. Portfast should solve the problem though if someone plugs a hub into the port later and that hub loops back you could get all sorts of ugly things. That's fairly unlikely though.

			- Matt

Chris Meisinger wrote:
I ran into a similar problem here.  If you're using a switched network,
and you are connected to Cisco Switches, then you might be running into
the same thing I did.  Cisco Catalysts have a feature called 'Spanning
Tree', where the switch broadcasts your mac address and figures out
shortest route, etc etc.  This will take about 5-10 seconds, depending
on the network.

Some network cards will bring the interface up and then down multiple
times during the initial phases of a kickstart (sis 900's are bad about
this), and so your port on the switch spends most of it's time in the
Spanning mode, where it can't forward any packets, and so your kickstart
times out before it can get a valid DHCP request through.

You can tell if this is your problem by checking out where the port your
plugged into terminates, and seeing if the status LED for that port is
amber, then change to green, then back to amber, and so on.


If this IS the problem, speak to your network admin about turning on
portfast for the ports you're trying to use to kickstart off of.  This
solved the problem, for me, but be away portfast has it's drawbacks too.


Chris Meisinger


-----Original Message-----
From: kickstart-list-admin@xxxxxxxxxx
[mailto:kickstart-list-admin@xxxxxxxxxx] On Behalf Of Michael Komitee
Sent: Monday, August 05, 2002 5:25 PM
To: kickstart-list@xxxxxxxxxx
Cc: Michael Komitee
Subject: PXE+DHCP/BOOTP+Kickstart problem

I'm using pxelinux, dhcpd (v2.0pl5), tftp, and kickstart to install
redhat 7.2 on IBM x330's.

Anaconda cannot get a dhcp address, mind you it gets one fine during the
pxeboot process, and gets one fine after the kickstart install fails and
i configure dhcp within anaconda and do a manual install.

I ran a sniffer and observed a problem with the dhcp transaction.
Normally, the transaction is as follows

Discovery	(client)
Offer		(server)
Request	(client)
ACK		(server)

all 4 packets have 1 transaction id.

the attempted transaction in question works as follows:

Discovery	(client)
Discovery	(client)
Offer		(server)
Offer		(server)

--and it fails... all 4 of these packets have a single transaction id as
well.


I'm not sure if its a problem with the client or the server. If it's the
client, Id guess theres a problem with my default config file for
pxelinux, and if it's the server i'd guess its a problem with my
dhcpd.conf. Any suggestions?

Here's my setup:

/tftpboot/initrd.img
/tftpboot/vmlinuz
/tftpboot/pxelinux.cfg/default

#/tftpboot/pxelinux.cfg/default
default linux
prompt 1
timeout 30
label linux
  kernel /vmlinuz
  append load_ramdisk=1 ramdisk_size=32768 initrd=/initrd.img
ksdevice=eth0 ks=nfs:/172.16.6.2/kickstart/ks/ksf.cfg

(it's one line in the file)

#/etc/dhcpd.conf
allow bootp;

default-lease-time 259200;        #3 day lease
max-lease-time 691200;            #8 day maximum lease
option routers  172.16.6.1;       #default gateway
option subnet-mask      255.255.255.0;
option domain-name      "ourdomain.net";
option domain-name-servers      172.16.1.6,172.16.1.7;
option broadcast-address        172.16.6.255;

filename        "/kickstart/ks/ksf.cfg";
next-server     172.16.6.2;

group {
        #PXE Boot Information
        filename "pxelinux.0";
        use-host-decl-names on;
        host pxeboot1 {
                hardware ethernet 00:02:55:C7:FD:CD;
                fixed-address 172.16.6.254;
        }
        host pxeboot2 {
                hardware ethernet 00:02:55:C6:1B:C9;
                fixed-address 172.16.6.253;
        }
}

#range
subnet 172.16.6.0 netmask 255.255.255.0 {
        range dynamic-bootp 172.16.6.129 172.16.6.252;
}

my exports are fine. /kickstart/ks is exported to the correct network.

(the reason I have the seemingly redundant directories of /kickstart/ks
is so that I can have a seperate share with different permissions
(writable) called /kickstart/logs where I store mac-address labeled logs
of the kickstart install.
when i used the same kickstart file (which i wont include because thats
not the problem) and boot via a disk and use this dhcp server (without
the allow bootp, or the pxe boot group) the kickstart install works
fine.

Thanks for the help in advance.



_______________________________________________
Kickstart-list mailing list
Kickstart-list@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/kickstart-list



_______________________________________________
Kickstart-list mailing list
Kickstart-list@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/kickstart-list


--

---------------------------------------------------------------------
Matt Fahrner                                    2 South Park St.
Manager of Networking                           Willis House
Burlington Coat Factory Warehouse               Lebanon, N.H.  03766
TEL: (603) 448-4100 xt 5150                     USA
FAX: (603) 443-6190                             Matt.Fahrner@xxxxxxxx
---------------------------------------------------------------------





[Index of Archives]     [Red Hat General]     [CentOS Users]     [Fedora Users]     [Fedora Maintainers]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]

  Powered by Linux