UEFI pxe boot problem from a network

[lebrun78]

@Sebastian-Roth said in UEFI pxe boot problem from a network:

ip a s

root@fogus ~]# ip a s
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: ens2f0np0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether b0:26:28:78:ce:d0 brd ff:ff:ff:ff:ff:ff
    inet 148.60.4.1/21 brd 148.60.7.255 scope global noprefixroute ens2f0np0
       valid_lft forever preferred_lft forever
    inet6 fe80::b226:28ff:fe78:ced0/64 scope link 
       valid_lft forever preferred_lft forever
3: eno1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
    link/ether 4c:d9:8f:8e:41:0f brd ff:ff:ff:ff:ff:ff
4: ens2f1np1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
    link/ether b0:26:28:78:ce:d1 brd ff:ff:ff:ff:ff:ff
5: eno2: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
    link/ether 4c:d9:8f:8e:41:10 brd ff:ff:ff:ff:ff:ff
6: virbr0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
    link/ether 52:54:00:1e:69:b9 brd ff:ff:ff:ff:ff:ff
    inet 192.168.122.1/24 brd 192.168.122.255 scope global virbr0
       valid_lft forever preferred_lft forever
7: virbr0-nic: <BROADCAST,MULTICAST> mtu 1500 qdisc fq_codel master virbr0 state DOWN group default qlen 1000
    link/ether 52:54:00:1e:69:b9 brd ff:ff:ff:ff:ff:ff

ip r s
default via 148.60.7.254 dev ens2f0np0 proto static metric 102 
148.60.0.0/21 dev ens2f0np0 proto kernel scope link src 148.60.4.1 metric 102 
192.168.122.0/24 dev virbr0 proto kernel scope link src 192.168.122.1 linkdown 

[lebrun78]

My fog server is not dhcp server

[Sebastian Roth]

@lebrun78 said in UEFI pxe boot problem from a network:

My fog server is not dhcp server

Then you might post the output of the same commands for the DHCP server.

[lebrun78]

@Sebastian-Roth

[root@sybille2 ~]# ip a s
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: em1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7d brd ff:ff:ff:ff:ff:ff
    inet 148.60.15.109/24 brd 148.60.15.255 scope global noprefixroute em1
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507d/64 scope link 
       valid_lft forever preferred_lft forever
3: em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
4: em2.14@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.14.252/24 brd 148.60.14.255 scope global noprefixroute em2.14
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
5: em2.2@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.4.3/21 brd 148.60.7.255 scope global noprefixroute em2.2
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
6: em2.11@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.11.249/24 brd 148.60.11.255 scope global noprefixroute em2.11
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
7: em2.3@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 129.20.15.252/24 brd 129.20.15.255 scope global noprefixroute em2.3
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
8: em2.10@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.10.252/24 brd 148.60.10.255 scope global noprefixroute em2.10
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
9: em2.13@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.13.249/24 brd 148.60.13.255 scope global noprefixroute em2.13
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever
10: em2.12@em2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:9a:4c:82:50:7e brd ff:ff:ff:ff:ff:ff
    inet 148.60.12.252/24 brd 148.60.12.255 scope global noprefixroute em2.12
       valid_lft forever preferred_lft forever
    inet6 fe80::529a:4cff:fe82:507e/64 scope link 
       valid_lft forever preferred_lft forever

[root@sybille2 ~]# ip r s
default via 148.60.15.254 dev em1 proto static metric 100 
129.20.15.0/24 dev em2.3 proto kernel scope link src 129.20.15.252 metric 406 
148.60.0.0/21 dev em2.2 proto kernel scope link src 148.60.4.3 metric 400 
148.60.10.0/24 dev em2.10 proto kernel scope link src 148.60.10.252 metric 403 
148.60.11.0/24 dev em2.11 proto kernel scope link src 148.60.11.249 metric 402 
148.60.12.0/24 dev em2.12 proto kernel scope link src 148.60.12.252 metric 404 
148.60.13.0/24 dev em2.13 proto kernel scope link src 148.60.13.249 metric 405 
148.60.14.0/24 dev em2.14 proto kernel scope link src 148.60.14.252 metric 401 
148.60.15.0/24 dev em1 proto kernel scope link src 148.60.15.109 metric 100

[lebrun78]

@george1421
Here is the dhcp extracted log for the not working sequence:

Mar  9 16:43:36 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar  9 16:43:36 sybille2 dhcpd: DHCPDISCOVER from 10:65:30:83:5c:4b via em2.10
Mar  9 16:43:36 sybille2 dhcpd: DHCPOFFER on 148.60.10.140 to 10:65:30:83:5c:4b via em2.10
Mar  9 16:43:39 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar  9 16:43:39 sybille2 dhcpd: DHCPREQUEST for 148.60.10.140 (148.60.10.252) from 10:65:30:83:5c:4b via em2.10
Mar  9 16:43:39 sybille2 dhcpd: DHCPACK on 148.60.10.140 to 10:65:30:83:5c:4b via em2.10

[george1421]

@lebrun78 Sorry I was not clear. From the target computer side, how do you know it didn’t work. Did you get a PXE-XXXX error?

[lebrun78]

@george1421
Sorry No error on the dell lattude 5900 laptop on booting except “No bootable media found”

[lebrun78]

@george1421
Do you have an example for a more wordy client ? I use libvirt for virtualisation, but I think it does not support uefi pxe boot.

[george1421]

@lebrun78 ok I think what we will need next is to have you install wireshark on a witness (extra) computer and use the capture filter of udp.port 67 or udp.port 68 Lets see what your dhcp server is telling that target computer to boot in uefi mode.

[lebrun78]

How could I have a witness ?

[george1421]

@lebrun78 said in UEFI pxe boot problem from a network:

How could I have a witness ?

Yes I was concerned that the name would not translate good.

What I meant is to have an extra computer plugged into the same subnet as the one that will not pxe boot in uefi mode. On this extra computer load wireshark (free program) and set the capture filter to what I said, so that it only sees the dhcp packets from your dhcp server. This will give us an idea where the problem is. If you save the pcap file (output from wireshark) to an internet file share site, then post the link here I will look at it quickly to see what is the pxe booting client being told.

[Sebastian Roth]

@lebrun78 said in UEFI pxe boot problem from a network:

Mar 9 16:43:36 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar 9 16:43:36 sybille2 dhcpd: DHCPDISCOVER from 10:65:30:83:5c:4b via em2.10
Mar 9 16:43:36 sybille2 dhcpd: DHCPOFFER on 148.60.10.140 to 10:65:30:83:5c:4b via em2.10
Mar 9 16:43:39 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar 9 16:43:39 sybille2 dhcpd: DHCPREQUEST for 148.60.10.140 (148.60.10.252) from 10:65:30:83:5c:4b via em2.10
Mar 9 16:43:39 sybille2 dhcpd: DHCPACK on 148.60.10.140 to 10:65:30:83:5c:4b via em2.10

We are sure the MAC is the one from the UEFI machine not properly PXE booting? Just want to make sure. If it is we do know the DHCP handshake is actually happening.

Looking through the config again I had the idea that maybe naming the classes all the same might be causing the issue. Though this doesn’t explain why only UEFI PXE boot is failing. Still you might want to change the names for one of the groups.

Other than that we’d need to get a packet dump from the DHCP information as suggested by George. You should be able to capture in the command line with a filter: tcpdump -w /tmp/foo.pcap ether host 10:65:30:83:5c:4b

Upload the pcap to a fileshare and post a link here.

[lebrun78]

@Sebastian-Roth said in UEFI pxe boot problem from a network:

Looking through the config again I had the idea that maybe naming the classes all the same might be causing the issue. Though this doesn’t explain why only UEFI PXE boot is failing. Still you might want to change the names for one of the groups.

Which parameter should I change in the class declaration ?

[Sebastian Roth]

@lebrun78 Just the names. Make it class "10Legacy" and so on for that subnet.

[lebrun78]

@Sebastian-Roth
Hello
I tried to rename class but no success.

I listen with wireshark and I discovered a problem with dhcp:
[https://filesender.renater.fr/?s=download&token=4e7a028b-59d6-afb9-d4bf-b08eb97b9e92](url du lien)

mask and rouer do not correspond with ip obtained which is in the good subnet???

subnet 148.60.10.0 netmask 255.255.255.0 {
##########################################                                                                                                                                                                                                                                      
option domain-name-servers 148.60.15.109,148.60.15.106 ;
option domain-name "istic.univ-rennes1.fr" ;
option routers 148.60.10.254 ;
option subnet-mask 255.255.255.0 ;
default-lease-time 600 ;
max-lease-time 1200 ;

group {
# On commente les deux lignes suivantes pour eviter le menu de Fog                                                                                                                                                                                                              
        next-server 148.60.4.1;

class "Legacy" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00000";
        filename "undionly.kkpxe";
    }
    class "UEFI-32-2" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00002";
        filename "i386-efi/ipxe.efi";
    }
    class "UEFI-32-1" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00006";
        filename "i386-efi/ipxe.efi";
    }
    class "UEFI-64-1" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00007";
        filename "ipxe.efi";
    }
    class "UEFI-64-2" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00008";
        filename "ipxe.efi";
    }
    class "UEFI-64-3" {
        match if substring(option vendor-class-identifier, 0, 20) = "PXEClient:Arch:00009";
        filename "ipxe.efi";
    }
host hobart { hardware ethernet 00:0C:29:CF:D4:2F; fixed-address hobart; option Host-name "hobart";}
host perth { hardware ethernet 00:0C:29:DA:4C:8B; fixed-address perth; option Host-name "perth";}
                                                                                                           
# marque debut pour dhcp-vm vlan 10, pas touche SVP.                                                                                                                                                                                                                            
host winrm10 {hardware ethernet 00:11:E2:61:00:03; fixed-address winrm10; option Host-name "winrm10";}     # Ajout par dhcp-vm le 22/08/2016 (11:32:32)                                                                                                                         
# marque fin pour dhcp-vm vlan 10, pas touche SVP.                                                                                                                                                                                                                              

}

        pool {
                deny members of "telephones-ip";
                range 148.60.10.180 148.60.10.220;
                next-server 148.60.4.1;
                #filename "pxelinux.0";                                                                                                                                                                                                                                         
        }

        ####################################################                                                                                                                                                                                                                    
        # pool d'adresse dynamique reserve aux telephones IP                                                                                                                                                                                                                    
        # testsip                                                                                                                                                                                                                                                               
        pool {
                allow members of "telephones-ip";
                range 148.60.10.224 148.60.10.239; #avec un masque de 255.255.255.240 pour erebus                                                                                                                                                                               
        }
}

[Sebastian Roth]

@lebrun78 said in UEFI pxe boot problem from a network:

mask and rouer do not correspond with ip obtained which is in the good subnet

You are right. If the posted config is the one in place we shouldn’t see router being set to 148.60.7.254 and mask to 255.255.248.0…

So something must be wrong with the config I suppose but I can’t find it. I’ve looked over it several times but just can’t see the obvious.

Is the config you posted really the whole one you have?

[lebrun78]

@Sebastian-Roth
Yes I only have one dhcp server.
I search to resolve the problem all last day, without success.
I change a different lease time for ech subnet.
Subnet 10 (Vlan10) clients receive Subnet 0 (vlan2) parameters. Why ?

Would you know how to make dhcp los more verbose ?

[lebrun78]

More strange, ip parameters are correct when the os (ubuntu) is running !

[Sebastian Roth]

@lebrun78 I haven’t tried this myself yet but just came across this: https://docs.menandmice.com/display/MM/using+the+log+function+for+debugging

You should be able to add specific logging statements to your config to see which one is used depending on VLANs as well as finding the difference between iPXE and Ubuntu booting on VLAN 10.

[lebrun78]

Hello
I take advantage that the university is deserted (covid19) to be able to do my dhcp tests, so I can easely insolate the boot log from a machine.
I have allways this problem on the vlan only with uefi pxe boot

Mar 31 11:37:08 sybille2 dhcpd: DHCPREQUEST for 148.60.10.34 from 00:11:e2:61:00:03 via em2.10
Mar 31 11:37:08 sybille2 dhcpd: DHCPACK on 148.60.10.34 to 00:11:e2:61:00:03 via em2.10
Mar 31 11:37:34 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar 31 11:37:34 sybille2 dhcpd: Lease request from 10:65:30:83:5c:4b in subnet 148.60.10
Mar 31 11:37:34 sybille2 dhcpd: DHCPDISCOVER from 10:65:30:83:5c:4b via em2.10
Mar 31 11:37:35 sybille2 dhcpd: DHCPOFFER on 148.60.10.193 to 10:65:30:83:5c:4b via em2.10
Mar 31 11:37:37 sybille2 dhcpd: PXEClient:Arch:00007:UNDI:003016
Mar 31 11:37:37 sybille2 dhcpd: Lease request from 10:65:30:83:5c:4b in subnet 148.60.10
Mar 31 11:37:37 sybille2 dhcpd: DHCPREQUEST for 148.60.10.193 (148.60.10.252) from 10:65:30:83:5c:4b via em2.10
Mar 31 11:37:37 sybille2 dhcpd: DHCPACK on 148.60.10.193 to 10:65:30:83:5c:4b via em2.10