BMR Linux Client not able to find boot loader from tftp server

[Scott_Chapman]

 Hello everyone, I'm having some trouble getting a linux client to boot from the BMR boot server.  I can boot using the DVD ISO and the machine recovers just fine, but it won't boot from the boot server.

In the attachment you can see that I'm getting an IP, but the client is not able to find the /bmr/pxelinux.0 file.  However, I've tested tftp from another machine and I'm able to find that file.

Here are some details about the config:

Master/Media server:  bckpr2
 Bckpr2 10.6.50.40 (primary NIC)
 Bckpr2-b 10.5.100.40 (backup NIC)

Boot Server:  linuxts3-b
 Linuxts3 10.6.50.159 (primary NIC)
 Linuxts3-b 10.5.101.83 (backup NIC)

Client Server: linuxts4-b
 Linuxts4 10.6.50.204 (primary NIC)
 Linuxts4-b 10.5.103.36 (backup NIC)

/etc/dhcpd.conf file from the boot server:

# DHCP Server Configuration file.
#   see /usr/share/doc/dhcp*/dhcpd.conf.sample
#   see 'man 5 dhcpd.conf'
#
log-facility local7;
ddns-update-style none;
ignore unknown-clients;
subnet 10.5.100.0 netmask 255.255.252.0 {
 default-lease-time 600;
 max-lease-time 7200;
        option domain-name "hidden.domain";
        option broadcast-address 10.5.100.255;
 option domain-name-servers 10.6.34.110,10.6.40.110;
# option routers 10.5.100.5;
}

host linuxts4-b-0A0632CC { hardware ethernet 00:50:56:8A:6E:19; fixed-address 10.6.50.204; next-server linuxts3-b; filename "/bmr/pxelinux.0"; } # added by Bare Metal Restore
host linuxts4-b-0A056724 { hardware ethernet 00:50:56:8A:6E:1F; fixed-address 10.5.103.36; next-server linuxts3-b; filename "/bmr/pxelinux.0"; } # added by Bare Metal Restore

/etc/xinetd.d/tftp file from boot server:

# default: off
# description: The tftp server serves files using the trivial file transfer \
# protocol.  The tftp protocol is often used to boot diskless \
# workstations, download configuration files to network-aware printers, \
# and to start the installation process for some operating systems.
service tftp
{
 socket_type  = dgram
 protocol  = udp
 wait   = yes
 user   = root
 server   = /usr/sbin/in.tftpd
 server_args  = -s /var/lib/tftpboot -v
 disable   = no
 per_source  = 11
 cps   = 100 2
 flags   = IPv4
 #interface  = 10.5.101.83
}
 

[Yasuhisa_Ishika]

Please check following points:

• You have already reloaded  or restarted xinetd after enabling tftpd service.
• xinetd is running.
• /bmr/pxelinux.0 exists under tftp directly (I suppose /var/lib/tftpboot)

BTW, what version of OS and NetBackup are you using on each host?

[Scott_Chapman]

I have checked that tftp is running via the boot server cli.  I've also used tftp from another machine and was able to connect and grab the file.

bckpr2 is solaris 10 running 7.5.0.4

linuxts3 is redhat 6.1 running 7.5.0.4... same for linuxts4.

Thanks, Scott

[Yasuhisa_Ishika]

I tested on VMware with RHEL 6.4 x84_64 and NetBackup 7.5.0.5. It works.
So for the next step, please check following points.

1. tftp requests reached to the boot server by tcpdump​.
Run "tcpdump -i <if_of_10.5.101.83> port tftp" on  linuxts3-b and check if tftp requests from linuxts4-b reached to linuxts3-b.

# tcpdump -i eth1 -w /tmp/tcpdump port tftp
# tcpdump -r /tmp/tcpdump 
reading from file /tmp/tcpdump, link-type EN10MB (Ethernet)
19:37:03.664097 IP bmr1-b.ah-esp-encap > bmr0-b.tftp:  32 RRQ "/bmr/pxelinux.0" octet tsize 0
19:37:03.708570 IP bmr1-b.acp-port > bmr0-b.tftp:  37 RRQ "/bmr/pxelinux.0" octet blksize 1456
19:37:03.769601 IP bmr1-b.57089 > bmr0-b.tftp:  68 RRQ "/bmr/pxelinux.cfg/01-00-0c-29-c8-95-1c" octet tsize 0 blksize 1440
19:37:03.772540 IP bmr1-b.57090 > bmr0-b.tftp:  56 RRQ "/bmr/pxelinux.cfg/0A1400D3" octet tsize 0 blksize 1440 
19:37:03.773256 IP bmr1-b.57091 > bmr0-b.tftp:  43 RRQ "/bmr/boot.msg" octet tsize 0 blksize 1440 
19:37:03.773818 IP bmr1-b.57092 > bmr0-b.tftp:  45 RRQ "/bmr/banner.img" octet tsize 0 blksize 1440 

2. tftp verbose log

I assume you have already enabled verbose logging of tftp as /etc/xinetd.d/tftp you posted shows.
If tftpd handles the requests from BMR client, messages like blow must be logged in /var/log/messages.

Apr  7 19:37:01 bmr0 dhcpd: DHCPDISCOVER from 00:0c:29:c8:95:1c via eth1
Apr  7 19:37:01 bmr0 dhcpd: DHCPOFFER on 10.20.0.211 to 00:0c:29:c8:95:1c via eth1
Apr  7 19:37:03 bmr0 dhcpd: DHCPREQUEST for 10.20.0.211 (10.20.0.210) from 00:0c:29:c8:95:1c via eth1
Apr  7 19:37:03 bmr0 dhcpd: DHCPACK on 10.20.0.211 to 00:0c:29:c8:95:1c via eth1
Apr  7 19:37:03 bmr0 xinetd[14572]: START: tftp pid=23146 from=10.20.0.211
Apr  7 19:37:03 bmr0 in.tftpd[23147]: RRQ from 10.20.0.211 filename /bmr/pxelinux.0
Apr  7 19:37:03 bmr0 in.tftpd[23147]: tftp: client does not accept options
Apr  7 19:37:03 bmr0 in.tftpd[23148]: RRQ from 10.20.0.211 filename /bmr/pxelinux.0
Apr  7 19:37:03 bmr0 in.tftpd[23149]: RRQ from 10.20.0.211 filename /bmr/pxelinux.cfg/01-00-0c-29-c8-95-1c
Apr  7 19:37:03 bmr0 in.tftpd[23150]: RRQ from 10.20.0.211 filename /bmr/pxelinux.cfg/0A1400D3
Apr  7 19:37:03 bmr0 in.tftpd[23151]: RRQ from 10.20.0.211 filename /bmr/boot.msg
Apr  7 19:37:03 bmr0 in.tftpd[23152]: RRQ from 10.20.0.211 filename /bmr/banner.img

3. Files are placed correctly under /var/lib/tftpboot.
As you see above, some files are fetched from the recovery target host. Please check if following files are placed and have read permittions.
Note: RHEL64 is SRT name I have created. 0A0A00D3, 0A1400D3, and C0A801D3 represent IP address in hex format.

# find /var/lib/tftpboot -ls
2230400    4 drwxr-xr-x   3 root     root         4096 Apr  7 08:35 /var/lib/tftpboot
2493274    4 drwxr-xr-x   4 root     root         4096 Apr  7 19:33 /var/lib/tftpboot/bmr
2492766    4 -r--r--r--   1 root     root          375 Apr  7 19:33 /var/lib/tftpboot/bmr/0A1400D3.info
2493278    4 drwxr-xr-x   2 root     root         4096 Apr  7 19:33 /var/lib/tftpboot/bmr/pxelinux.cfg
2492765    4 -r--r--r--   1 root     root          556 Apr  7 19:33 /var/lib/tftpboot/bmr/pxelinux.cfg/0A0A00D3
2493279    4 -r--r--r--   1 root     root          424 Apr  7 08:35 /var/lib/tftpboot/bmr/pxelinux.cfg/default
2492769    4 -r--r--r--   1 root     root          556 Apr  7 19:33 /var/lib/tftpboot/bmr/pxelinux.cfg/C0A801D3
2492767    4 -r--r--r--   1 root     root          556 Apr  7 19:33 /var/lib/tftpboot/bmr/pxelinux.cfg/0A1400D3
2492709    4 -r--r--r--   1 root     root          375 Apr  7 19:33 /var/lib/tftpboot/bmr/0A0A00D3.info
2493277   12 -r--r--r--   1 root     root        11826 Apr  7 19:33 /var/lib/tftpboot/bmr/pxelinux.0
2883873    4 drwxrwxrwx   2 root     root         4096 Apr  7 11:21 /var/lib/tftpboot/bmr/RHEL64
2883875 42092 -r--r--r--   1 root     root     43101607 Apr  7 19:33 /var/lib/tftpboot/bmr/RHEL64/initrd.img
2883874 3952 -r--r--r--   1 root     root      4044560 Apr  7 19:33 /var/lib/tftpboot/bmr/RHEL64/vmlinuz
2492768    4 -r--r--r--   1 root     root          375 Apr  7 19:33 /var/lib/tftpboot/bmr/C0A801D3.info
2493275    8 -r--r--r--   1 root     root         4284 Apr  7 19:33 /var/lib/tftpboot/bmr/banner.img
2493276    4 -r--r--r--   1 root     root           16 Apr  7 19:33 /var/lib/tftpboot/bmr/boot.msg

[Scott_Chapman]

Wow, thanks, that is a lot of good info there!  Thanks so much for your help on this!

I setup a tail of /var/log/messages, started the tcpdump, booted the VM and here is what I got... I'm not seeing any tftp requests anywhere...

==> /var/log/messages <==
Apr  8 08:45:04 linuxts3 dhcpd: DHCPDISCOVER from 00:50:56:8a:6e:1f via eth1
Apr  8 08:45:04 linuxts3 dhcpd: DHCPOFFER on 10.5.103.36 to 00:50:56:8a:6e:1f via eth1
Apr  8 08:45:06 linuxts3 dhcpd: DHCPREQUEST for 10.5.103.36 (10.5.101.83) from 00:50:56:8a:6e:1f via eth1
Apr  8 08:45:06 linuxts3 dhcpd: DHCPACK on 10.5.103.36 to 00:50:56:8a:6e:1f via eth1
^C

[root@linuxts3 bmrd]# tcpdump -i eth1 -w /tmp/tcpdump port tftp
tcpdump: listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes
^C0 packets captured
1 packets received by filter
0 packets dropped by kernel
 

tftp file info:

[root@linuxts3 tftpboot]# find /var/lib/tftpboot/bmr -ls
403435    4 drwxr-xr-x   4 root     root         4096 Apr  5 12:34 /var/lib/tftpboot/bmr
403938    4 -r--r--r--   1 root     root           16 Apr  5 12:34 /var/lib/tftpboot/bmr/boot.msg
398573    4 drwxrwxrwx   2 root     root         4096 Apr  3 13:31 /var/lib/tftpboot/bmr/RH61_7504_64bit
398575 3792 -r--r--r--   1 root     root      3881120 Apr  5 12:34 /var/lib/tftpboot/bmr/RH61_7504_64bit/vmlinuz
398576 39824 -r--r--r--   1 root     root     40777056 Apr  5 12:34 /var/lib/tftpboot/bmr/RH61_7504_64bit/initrd.img
398578    4 -r--r--r--   1 root     root          379 Apr  5 12:34 /var/lib/tftpboot/bmr/0A056724.info
403940    4 drwxr-xr-x   2 root     root         4096 Apr  5 12:34 /var/lib/tftpboot/bmr/pxelinux.cfg
403941    4 -r--r--r--   1 root     root          430 Apr  3 12:53 /var/lib/tftpboot/bmr/pxelinux.cfg/default
398577    4 -r--r--r--   1 root     root          591 Apr  5 12:34 /var/lib/tftpboot/bmr/pxelinux.cfg/0A0632CC
398579    4 -r--r--r--   1 root     root          591 Apr  5 12:34 /var/lib/tftpboot/bmr/pxelinux.cfg/0A056724
403939   12 -r--r--r--   1 root     root        11826 Apr  5 12:34 /var/lib/tftpboot/bmr/pxelinux.0
398568    4 -r--r--r--   1 root     root          379 Apr  5 12:34 /var/lib/tftpboot/bmr/0A0632CC.info
403937    8 -r--r--r--   1 root     root         4284 Apr  5 12:34 /var/lib/tftpboot/bmr/banner.img

[Yasuhisa_Ishika]

Files seems to be OK. Cliest must receive IP address, boot server name, and boot strap file.

I reviewed your dhcpd.conf. Broadcast address for backup LAN seems wrong. Please fix it and try again. routers option is required at later step, please also set "options routers <any ip address in backup LAN>". Both real router and dummy address will be OK.

[Scott_Chapman · Accepted solution]

OK, so I figured out my issue... should have known that windows would be the problem... it's always the problem!  ;)

There was an SMS server that was running Windows Deployment Services and broadcasting on the backup network... I was finally able to trap the tcpdump output showing this machine "getting in the way".  Bound that WDS to the primary interface on that windows box so it wasnt' blasting the backup network and the machine booted from BMR immediately.

Thanks for all the help on this.