Re: PXE/DHCP/tftpd boot failure in FBSD 4.4-STABLE since yesterday!! Please Help!
On Wed, Oct 31, 2001 at 11:50:35PM +0100, Hartmann, O. wrote: > On Wed, 31 Oct 2001, Chris Peiffer wrote: > > > For testing purposes, I switched the sources back to the 20th October, > I know that this codestage definitely worked! I recompiled a 'make world', > all the terminal related stuff, new kernels and installed them, then I > rebooted. After reboot I recompiled the dhcp code from the ports, with the > same effect: nothing has changed, the stations still fail to boot. I think this is > a problem with our LAN and the maintanace of the computer center. > > Now is the essentiell question: how to figure out what's going wrong? > I need to examine what the DHCP recieves and delivers and I need to > know what type of packets are exchanged between the DHCP server, > its client and what happens when the client has recieved its dhcp > config info and trys to get pxeboot image. How can I watch whether the > diskless client gets its pxeboot image? Well, I would like to know > some details about the protocol. The reason is, that the guys of our computer > center tend to push away responsibilty and I would like to avoid having > all the trouble I did not produce carrying on my shoulders ... tcpdump is your friend I guess, possibly together with ethereal. It should at least be possible to see dhcp requests, offers, etc. If the kernel is transfered using TFTP, it should be difficult to see that in tcpdump. With ethereal you will be able to analyze traffic captured with tcpdump more closely. You should run tcpdump on the DHCP server if you've a switched network (I guess so). Also, try to enable as much logging (or debugging) as possible in dhcpd, tftp (-l switch), etc. HTH, Alson -- ,---. > Name: Alson van der Meulen < > Personal:[EMAIL PROTECTED]< > School: [EMAIL PROTECTED]< `---' what's this hash prompt on my terminal mean? - To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: PXE/DHCP/tftpd boot failure in FBSD 4.4-STABLE since yesterday!!Please Help!
[snip] did you try to ask the ISC dhcp client/server mailing lists ? http://www.isc.org/services/public/lists/dhcp-lists.html Cyrille. -- Cyrille Lefevre mailto:[EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: PXE/DHCP/tftpd boot failure in FBSD 4.4-STABLE since yesterday!!Please Help!
On Wed, 31 Oct 2001, Chris Peiffer wrote: For testing purposes, I switched the sources back to the 20th October, I know that this codestage definitely worked! I recompiled a 'make world', all the terminal related stuff, new kernels and installed them, then I rebooted. After reboot I recompiled the dhcp code from the ports, with the same effect: nothing has changed, the stations still fail to boot. I think this is a problem with our LAN and the maintanace of the computer center. Now is the essentiell question: how to figure out what's going wrong? I need to examine what the DHCP recieves and delivers and I need to know what type of packets are exchanged between the DHCP server, its client and what happens when the client has recieved its dhcp config info and trys to get pxeboot image. How can I watch whether the diskless client gets its pxeboot image? Well, I would like to know some details about the protocol. The reason is, that the guys of our computer center tend to push away responsibilty and I would like to avoid having all the trouble I did not produce carrying on my shoulders ... Thanks ... :>On Wed, Oct 31, 2001 at 07:14:39PM +0100, Hartmann, O. wrote: :>> Hello. :>> :>> Since yesterday morning we have massive problems with PXE booting :>> diskless stations! :>> :>> The server has isc-dhcpd2 and since two hours ago isc-dhcpd3. The server :>> runs FreeBSD 4.4-STABLE, the last cvsupdate has been done two hours ago, :>> the previous cvsupdate has been done two days ago, but AFTER the first occurence :>> of the following described problem. :> :>So you changed the dhcp server after the problem started occuring? :>What change preceded the failure? What does your dhcpd.conf look like? :>What about the kernel config of the kernel sent to the netbooting :>clients? :> :>I am currently using ISC dhcpd V3.0rc12 with FreeBSD :>successfully. I've attached my dhcpd.conf if it helps you any. :> -- MfG O. Hartmann [EMAIL PROTECTED] IT-Administration des Institutes fuer Physik der Atmosphaere (IPA) Johannes Gutenberg Universitaet Mainz Becherweg 21 55099 Mainz Tel: +496131/3924662 (Maschinenraum) Tel: +496131/3924144 FAX: +496131/3923532 To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
Re: PXE/DHCP/tftpd boot failure in FBSD 4.4-STABLE since yesterday!!Please Help!
On Wed, 31 Oct 2001, Chris Peiffer wrote:
Hello.
I'm currently usinf isc-dhcpd3 V 3.0.1.rc2 from the ports collection.
I use this server since today afternoon, before I ran isc-dhcpd2 2.0.5pl2
or similar from the ports collection. Both show the same phenomenon.
Since more than half a year we use the same config file dhcpd.conf
and never changed the dhcp/tftp config so this __can-not-be__ the
cause! My dhcpd.conf looks like similar to yours, but I have for each
known ether address its own entry:
# stud04
host 134.93.180.239 {
option root-path "/usr/diskless/amd";
#filename "/boot/pxeboot";
hardware ethernet 00:02:b3:17:89:0e;
fixed-address 134.93.180.239;
}
I have a more 'per client config' and not a more sophisticated 'per class'
configuration as yours, but the eefect is the same.
Our computing center has several renewals of their power supplying infrastructure
and theirefore the backbone and all switches were shut down that time. Since
then our problems began on some terminals, but this phenomenon was of sporadic
nature. Today it is a 'stable' situation: none of them work.
I need to exmamine what's going on on the network to see whether my local installation,
e.g. FreeBSD itselfs, corrupts diskless booting or the switch device or gateway of
the computing center (which maintains our backbone and infrastructure) has some
problems.
Thanks,
Oliver
:>On Wed, Oct 31, 2001 at 07:14:39PM +0100, Hartmann, O. wrote:
:>> Hello.
:>>
:>> Since yesterday morning we have massive problems with PXE booting
:>> diskless stations!
:>>
:>> The server has isc-dhcpd2 and since two hours ago isc-dhcpd3. The server
:>> runs FreeBSD 4.4-STABLE, the last cvsupdate has been done two hours ago,
:>> the previous cvsupdate has been done two days ago, but AFTER the first occurence
:>> of the following described problem.
:>
:>So you changed the dhcp server after the problem started occuring?
:>What change preceded the failure? What does your dhcpd.conf look like?
:>What about the kernel config of the kernel sent to the netbooting
:>clients?
:>
:>I am currently using ISC dhcpd V3.0rc12 with FreeBSD
:>successfully. I've attached my dhcpd.conf if it helps you any.
:>
--
MfG
O. Hartmann
[EMAIL PROTECTED]
IT-Administration des Institutes fuer Physik der Atmosphaere (IPA)
Johannes Gutenberg Universitaet Mainz
Becherweg 21
55099 Mainz
Tel: +496131/3924662 (Maschinenraum)
Tel: +496131/3924144
FAX: +496131/3923532
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
PXE/DHCP/tftpd boot failure in FBSD 4.4-STABLE since yesterday!!Please Help!
Hello. Since yesterday morning we have massive problems with PXE booting diskless stations! The server has isc-dhcpd2 and since two hours ago isc-dhcpd3. The server runs FreeBSD 4.4-STABLE, the last cvsupdate has been done two hours ago, the previous cvsupdate has been done two days ago, but AFTER the first occurence of the following described problem. All clients are AMD Doron 700 MHz machines, all of our servers using Intel EtherExpress Pro NICs (fxp), and all clients are at the same code level as the server, means: they have the same source base and therefore the same FBSD 4.4-STABLE. On serverside we use IPFW as our filtering system (and I saw several changes in ipfw.c the last days). The server has tftp enabled in /etc/inetd.conf and we did not change the configuration since the last two weeks and in this time the terminals booted as usual. I would like to avoid questions about configuration errors and therefore I would like to tell you, that our configuration hasn't been changed the last month, only the installation of cvsupdated code from 4.4-STABLE has been done the regular way. Phenomenon: This morning our students worked as usual with the terminals. They booted the normal way, got theire IP from DHCP server, their gate, their DNS and so on. Then they received their pxeboot image and bootstraped the kernel. So far. Then, at lunchtime, most of the students left their places to get their meal and some of them switched off the terminals, others left their terminals switched on - and they ran well. When the others came back and switched their terminals on the got this: Response from DHCP server, the Intel NIC reports that DHCP delivered the appropriate informations like IP, gateway, broadcast address, mask and IP of the DHCP server. But then we got and still get this message: Press F12 for network service boot PXE-M0F:Exiting Intel PXE-ROM Then I get a BIOS message of a boot failure. Two terminals have the most recent PXE boot image from intel, version 4.0.19, several others have the older one, 4.0.17. Two or three hours ago I compiled a complete new FBSD after a cvsupdate, installed the necessary binaries at the NFS location where the clients should get it. But it this seems to have no effect: tftp sends no pxeboot and as I understood this procedure, for pxeboot loading, starting and then bootsraping the kernel NFS is not needed yet, it comes into play after the kernel booted and tries to get its system running - but this stage never gets reached now. Since yesterday we have had problems with our network since our computer center changed several subjects within the USV basics, but they told me nothing has been changed to the switches or other LAN facilities. When looking into the ARPA chache of the DHCP server, I usually see the MAC of each connected system. But I do not know whether the MAC occurs herein when a DHCP connection has been done or after a kernel has been booted and the appropriated IP has been broadcasted by the running UNIX/FBSD. I'm not very familiar in how to track down this problem due to the fact that I have not access to the switches and gateways. How can I test what response the DHCP server offers and waht the the terminal tries to do? I lloks like that after the terminals gottheir IP and other stuff, the connection gets cut off and nothing happend, but I can not track down the problem that way to say its the fault of FreeBSD or it is some strange behaviour in the network caused by a faulty LAN switch or gateway. I try to attach a terminal directly to the server, but doing this, the server must be detached from the LAN and I'm afraid of the problems in lacking of a suitable DNS. Can anyone respond to that problem? Does FBSD have problems with pxeboot environment now? I do not belive this right now, due to the LAN problems from yesterday on, but I need to be sure ... Thanks a lot, Oliver -- MfG O. Hartmann [EMAIL PROTECTED] IT-Administration des Institutes fuer Physik der Atmosphaere (IPA) Johannes Gutenberg Universitaet Mainz Becherweg 21 55099 Mainz Tel: +496131/3924662 (Maschinenraum) Tel: +496131/3924144 FAX: +496131/3923532 To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-stable" in the body of the message
