Shutting down and restarting the dhcp server via omapi

Jürgen Dietl juergen.dietl at googlemail.com
Thu Feb 24 15:56:26 UTC 2011


Hello Nils (and all the others :-)

as promised the summary of the trace.
No.     Time        Absolute Time   Source
Destination           New Column Protocol
Info                                                            New Column
      2 0.000097    16:29:13.308980 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [ACK] Seq=1 Ack=13 Win=5530
Len=0         16:29:13.308980
      3 1.992243    16:29:15.301126 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover DHCPFO   Contact xid:
a1                                                 16:29:15.301126
      6 10.008093   16:29:23.316976 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [ACK] Seq=13 Ack=25 Win=5530
Len=0        16:29:23.316976
      7 11.994244   16:29:25.303127 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover DHCPFO   Contact xid:
a2                                                 16:29:25.303127
     10 20.016121   16:29:33.325004 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [ACK] Seq=25 Ack=37 Win=5530
Len=0        16:29:33.325004
     11 21.996225   16:29:35.305108 xServer 2xxx         xServer 1xx
         dhcp-failover DHCPFO   Contact xid:
a3                                                 16:29:35.305108
     14 30.020585   16:29:43.329468 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [ACK] Seq=37 Ack=49 Win=5530
Len=0        16:29:43.329468
     15 36.125682   16:29:49.434565 xxxxxxxxxx          xxxxxxxxxx
14619      TCP      7911 > 14619 [FIN, ACK] Seq=1 Ack=1 Win=730
Len=0               16:29:49.434565
     18 40.030602   16:29:53.339485 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [ACK] Seq=37 Ack=61 Win=5530
Len=0        16:29:53.339485
     19 47.142748   16:30:00.451631 xxxxxxxxxx          xxxxxxxxxx
dhcp-failover TCP      32197 > dhcp-failover [RST, ACK] Seq=37 Ack=61
Win=5530 Len=0   16:30:00.451631
     21 52.144123   16:30:05.453006 xxxxxxxxxx          xxxxxxxxxx
13880      TCP      dhcp-failover > 13880 [RST, ACK] Seq=1 Ack=1 Win=0
Len=0        16:30:05.453006
     23 57.148097   16:30:10.456980 xxxxxxxxxx          xxxxxxxxxx
55083      TCP      dhcp-failover > 55083 [RST, ACK] Seq=1 Ack=1 Win=0
Len=0        16:30:10.456980


Server 2 = the server that shuts down
Server 1 = Failover Primary

I set a filter of source server 2. Because this is the server that should
inform the primary of the shut down. So that the partner-down state occurs.
For me the tcp session is quite ok because it finnish with a FIN. But the
DHCP-FO session suddenly breaks with a reset.

Is this normal? Or is this the reason why I cant see any shut down
information for the primary?

What Version of ISC-DHCP do you run?
What OS and Version do you use?

thanx a lot for all your help,
cheers, Juergen


2011/2/24 Jürgen Dietl <juergen.dietl at googlemail.com>

> Hello,
>
> thanx a lot for your help. I tried it with the following script:
>
> #!/bin/bash
>
> /usr/local/bin/omshell << END
> server localhost
> port 7911
> key xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
> connect
> new control
> open
> set state=2
> update
> END
>
> The log file:
>
> Feb 24 16:03:14 xxxxxxx dhcpd: peer xxxxxxxx: disconnected
> Feb 24 16:03:14 xxxxxxx dhcpd: failover peer xxxxxxx: I move from normal to
> communications-interrupted
>
>
> Feb 24 16:03:09 xxxxxxx dhcpd: Disabling output on
> LPF/eth0/00:00:00:00:00:00/10.2.220.0/24
> Feb 24 16:03:09 xxxxxxx dhcpd: Disabling input on
> LPF/eth0/00:00:00:00:00:00/10.2.220.0/24
> Feb 24 16:03:09 xxxxxxx dhcpd: failover peer xxxxxxx: I move from normal to
> shutdown
> Feb 24 16:03:14 xxxxxxx dhcpd: failover peer xxxxxxx: I move from shutdown
> to recover
>
> I will do some trace soon and try to see how the 2 servers communicate.
> Till now it looks to me that the server that shuts down dont tell it to the
> other server.
>
> cu later,
> cheers,
> Juergen
>
>
>
>
> 2011/2/24 Nils-Henner Krueger <nhk at snhc-krueger.de>
>
> Am 24.02.2011 14:56, schrieb Jürgen Dietl:
>> > My main problem is that when I stop the dhcp server I always will get a
>> > "communication-interrupted". What I would like to have is a
>> "partner-down".
>> >
>> > So do any of you have a script or an idea how I can make the server to
>> > make a graceful shutdown of dhcp service?
>>
>> 'man dhcpd' of version 4.11 says "To shut the server down, open its
>> control object and set the state attribute to 2."
>>
>> I do it like this in a start/stop-Script:
>>
>> ${PRODPATH}/bin/omshell << END
>> server localhost
>> port xxxx
>> connect
>> new control
>> open
>> set state=2
>> update
>> END
>>
>> This works fine, logging on the shutdown server:
>>
>> dhcpd: failover peer ha: I move from normal to shutdown
>> dhcpd: failover peer ha: peer moves from normal to partner-down
>> dhcpd: failover peer ha: I move from shutdown to recover
>>
>> Logging on the partner cluster member:
>>
>> dhcpd: failover peer ha: peer moves from normal to shutdown
>> dhcpd: failover peer ha: I move from normal to partner-down
>> dhcpd: peer ha: disconnected
>>
>>
>> nils-henner
>> _______________________________________________
>> dhcp-users mailing list
>> dhcp-users at lists.isc.org
>> https://lists.isc.org/mailman/listinfo/dhcp-users
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20110224/2d94bf84/attachment.html>


More information about the dhcp-users mailing list