bootpd message !

[Ivan Azuara]

Hi !

I was monitoring my rp server (glance), and find a big activity in disk (100%). I detect a lot of processes with the name "bootpd".

This process describe the next:

"The bootpd daemon implements three functions: a Dynamic Host
Configuration Protocol (DHCP) server as defined in RFC1541, an
Internet Boot Protocol (BOOTP) server as defined in RFC951 and
RFC1395, and a DHCP/BOOTP relay agent as defined in RFC1542"

when i invoke the command "auto_parms", receive the next:

# auto_parms

/sbin/auto_parms: DHCP access is disabled (see /etc/auto_parms.log)

How can i eliminate this messages ?.


Thank??s in advance !

[Steven E. Protter · Accepted solution]

What bootp does for you is it allows other HP-UX servers to actually boot off of the server in questions.

This is particularly useful if you are running an Ignite Server. It lets you boot off the Ignite server and load a disaster recovery image (make_net_recovery) and rebuild a machine completely.

Beyond that, it also lets you run a mini DHCP server.

In the file /etc/inetd.conf is a line that looks like this:
bootps dgram udp wait root /usr/lbin/bootpd bootpd

You should be able to comment it out, save the file and ....

ientd -c

To stop this protcol from running.

Don't do this if this server is running Ignite as a server. These messages are perfectly normal under these circumstances.

SEP

[Con O'Kelly]

Hi

I'm not sure of you're question.
auto_parms will give this message if DHCP is disabled. Its perfectly normal.
As regards bootpd, it is also used by ignite-ux during network installs.
If you have bootp processes that are continuously showing very high disk activity then I would investigate them further.

Can you post the output of :
# ps -ef | grep bootp

Cheers
Con

[Michael Steele_2]

/etc/rc.config.d/netconf has a DHCP_ENABLE[index]entry, make sure it's 0. Chances are its '1' when it should be '0'.

Check UDP ports 67 and 68 in /etc/services and 'bootps' in /etc/inetd.conf. Comment them out and run 'inetd -c', but this will be extraneous if netconf is the problem.

You can also 'ps -ef | grep -i bootpd'.

The common file for DHCP and bootp is /etc/bootptab. Does this exist?

[Ivajlo Yanakiev]

Also use kill to stop bootpds process if your server is not Ignite server.
Because inetd will stop only future bootpd session.

[Suresh Patoria]

Hi,

If you want to disable bootpd

then you edit the /etc/rc.config.d/netdaemons file

set the START_RBOTD=0

then check the system

[Ivan Azuara]

That??s very strange because this event appeared suddenly. I was probbing your comments, and i have some points:

a) Constantly appears a core file on the system, and if delete this file, appears again. I??m attaching this file.

b) The exit of the command "ps" is :

# ps -ef | grep -i bootpd
root 8269 26141 0 10:34:21 pts/ta 0:00 grep -i bootpd
root 8259 643 1 10:34:21 ? 0:00 bootpd

If check only appears one process , but in the glance console appears until 25 bootpd processes. All with the PPID 643.

c) If i comment the bootpd line in /etc/inetd.conf, and close the ports that mentioned, will enough for eliminate this processes ?. My server have installed the Ignite/UX product, but not is a ignite server.


Thank??s in advance !

[Michael Steele_2]

A) Can't open your attachment.

B) What was in /etc/rc.config.d/netconf?

C) What is PID 643?

D) Run lsof on 643.

lsof -p 643.

E) Kill all of these bootpd processes:

ps -ef | grep -i bootpd | while read a b c d e f h
do
kill -9 $b
done

F) Analyze core file

file core
strings core | more

G) Regarding "...If i comment the bootpd line in /etc/inetd.conf, and close the ports that mentioned, will enough for eliminate this processes ?..."

No - See E)

[Ivan Azuara]

OK !

a)and f) May be is a problem in the HP site, but basically analizing the core file ,the content show information about the DHCP. An extract of the core file is the next:

# strings core |more

s_inb_02
B.11.11
9000/800
256542636
bootpd
B85H
$Revision: 92453-07 linker linker crt0.o B.11.16 000601 $
/usr/lib/dld.sl
ERROR: mmap failed for dld
ERROR: mmap failed for TSD
reserved
ethernet
ethernet3
ax.25
pronet
chaos
ieee802
arcnet
ethernet
ethernet3
ether
ether3
ieee802
token-ring
pronet
chaos
arcnet
ax.25
bootp
DHCPDISCOVER
DHCPOFFER
DHCPREQUEST
DHCPDECLINE
DHCPACK
DHCPNAK
DHCPRELEASE
DHCPINFORM
dhcppreview
bootprequest
Subnet Mask
Time Offset
Router
Time Server
dhcp_bootrequest(): setup_relay returned %d
offer not sent: reason = %d
request: free_offer returned %d
ACK not sent: reason = %d
dhcp_bootrequest(): do_decline returned %d
dhcp_bootrequest(): do_release returned %d
Server received invalid message type %s
Server received unknown message type %d
assuming infinite lease for client IP address.
DHCP Server received an invalid message type %d.
client %s requested unknown/unservable option.tag = %x
sending NAK: %s
cannot send DHCP msgtype %s to client
File name too long
File name too long

b) The content in my "/etc/rc.config.d/netconf" file for my 2 LANICS is:

# lanscan
Hardware Station Crd Hdw Net-Interface NM MAC HP-DLPI DLPI
Path Address In# State NamePPA ID Type Support Mjr#
1/0/1/0/0/4/0 0x00306E4AD085 0 UP lan0 snap0 1 ETHER Yes
119
1/0/12/0/0 0x00306E4AC5AA 1 UP lan1 snap1 2 ETHER Yes 119

Content of netconf file :

INTERFACE_NAME[0]=lan0
IP_ADDRESS[0]=130.10.17.9
SUBNET_MASK[0]=255.255.255.0
BROADCAST_ADDRESS[0]=""
INTERFACE_STATE[0]=""
DHCP_ENABLE[0]=0

IP_ADDRESS[1]=130.10.17.1
SUBNET_MASK[1]=255.255.255.0
INTERFACE_NAME[1]=lan1
BROADCAST_ADDRESS[1]=130.10.17.255
INTERFACE_STATE[1]=up
DHCP_ENABLE[1]=0

c) Is the PID number of the bootpd process.

d) I don??t have that command.

e) and g) That??s correct!, if i kill the bootpd process, disapear the high activity on the disk, but when restart the server appears the bootpd process again , the high disk consume , and the core file too.

Is obviusly that bootpd process is the response of the high disk percent. I??m going to comment the lines about bootpd for avoiding this event, and restart the inetd configuration "inetd -c".

An other suggestion ?


Thank??s in advance !

[Steven E. Protter]

1/0/1/0/0/4/0 0x00306E4AD085 0 UP lan0 snap0 1 ETHER Yes
119
1/0/12/0/0 0x00306E4AC5AA 1 UP lan1 snap1 2 ETHER Yes 119


INTERFACE_NAME[0]=lan0
IP_ADDRESS[0]=130.10.17.9
SUBNET_MASK[0]=255.255.255.0
BROADCAST_ADDRESS[0]=""
INTERFACE_STATE[0]=""
DHCP_ENABLE[0]=0

IP_ADDRESS[1]=130.10.17.1
SUBNET_MASK[1]=255.255.255.0
INTERFACE_NAME[1]=lan1
BROADCAST_ADDRESS[1]=130.10.17.255
INTERFACE_STATE[1]=up
DHCP_ENABLE[1]=0

This is an unsupported configuration. You can not without APA bring up to different lan cards on the same network and subnet.

I tried this once and it brought the whole network subsystem down. The machine dropped of the network.

The network is 130.10.17

If these are two ports on the same card, you still can't do this without the Auto Port Aggregation(APA) product.

Change one of the cards to a different network and all of your problems will go away. This is a solution. I'm surprised your box stays on the network at all.

SEP

[Ivan Azuara]

I made the changes in my /etc/rc.config.d/netconf file, but the event continuos. The only solution in this moment was comment the bootpd line in the /etc/inetd.conf file , and restart the configuration "inetd -c".

All my problems desappear with this solution. I don??t think that i have problems in the future. What do you think ?.


Thank??s in advance !

[A. Clay Stephenson]

If this host is not serving as a bootp server for Ignite clients, printers, X-terminals, or diskless workstations then you can certainly disable bootpd.

[Michael Steele_2]

Good Deal!

Regarding "...All my problems desappear with this solution. I don??t think that i have problems in the future. What do you think ?...."

I think you'll be find since the bootpd daemon can no longer start.

But with all of those bootp processes running, what was the root cause? Was another node on your network invoking bootp at ports 67 and 68?

Why don't you get a copy of 'tcpdump' and 'lsof' and monitor this?

http://hpux.cict.fr/hppd/hpux/Networking/Admin/tcpdump-3.6.2/

http://hpux.cict.fr/hppd/hpux/Sysadmin/lsof-4.64/