Send kea-dev mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.isc.org/mailman/listinfo/kea-dev
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of kea-dev digest..."
Today's Topics:
1. perfdhcp bugfix (Angelo Failla)
2. Re: perfdhcp bugfix (Francis Dupont)
3. Re: perfdhcp bugfix (Angelo Failla)
4. Re: perfdhcp bugfix (Marcin Siodelski)
5. Re: perfdhcp bugfix (Angelo Failla)
----------------------------------------------------------------------
Message: 1
Date: Thu, 7 Apr 2016 19:25:17 +0000
From: Angelo Failla <[email protected]>
To: "[email protected]" <[email protected]>
Subject: [kea-dev] perfdhcp bugfix
Message-ID: <[email protected]>
Content-Type: text/plain; charset="utf-8"
Found a problem when making some benchmark testing. The XID is incremented
wrongly.
A XID should last for the duration of a SORA 4-way handshake...
However it doesn?t seem to be working properly:
$ perfdhcp -4 -r 1 -n 2 -p 60 -t 1 -b mac=90:e2:ba:95:b9:28 10.88.45.33
$ sudo tshark udp port 67
Running as user "root" and group "root". This could be dangerous.
Capturing on 'eth0'
1 0.000000 10.35.166.47 -> 10.88.45.33 DHCP 295 DHCP Discover -
Transaction ID 0x0
2 0.030098 10.88.45.33 -> 10.35.166.47 DHCP 416 DHCP Offer -
Transaction ID 0x0
3 0.030426 10.35.166.47 -> 10.88.45.33 DHCP 307 DHCP Request -
Transaction ID 0x1
4 0.057158 10.88.45.33 -> 10.35.166.47 DHCP 416 DHCP ACK -
Transaction ID 0x1
5 1.000689 10.35.166.47 -> 10.88.45.33 DHCP 295 DHCP Discover -
Transaction ID 0x2
6 1.027757 10.88.45.33 -> 10.35.166.47 DHCP 416 DHCP Offer -
Transaction ID 0x2
7 1.027866 10.35.166.47 -> 10.88.45.33 DHCP 307 DHCP Request -
Transaction ID 0x3
8 1.047429 10.88.45.33 -> 10.35.166.47 DHCP 416 DHCP ACK -
Transaction ID 0x3
I made a PR on github for this in case you want to take a look:
https://github.com/isc-projects/kea/pull/19
--
Angelo Failla
Cluster Infrastructure - Dublin
[email protected]
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://lists.isc.org/pipermail/kea-dev/attachments/20160407/c5e0dc7c/attachment-0001.html>
------------------------------
Message: 2
Date: Thu, 07 Apr 2016 21:44:24 +0000
From: Francis Dupont <[email protected]>
To: Angelo Failla <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [kea-dev] perfdhcp bugfix
Message-ID: <[email protected]>
Content-Type: text/plain; charset="us-ascii"
Even I cannot see how it could matter I agree the spec says the XID
is copied from the server OFFER so it is a bug.
Regards
Francis Dupont <[email protected]>
------------------------------
Message: 3
Date: Thu, 7 Apr 2016 21:50:58 +0000
From: Angelo Failla <[email protected]>
To: Francis Dupont <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [kea-dev] perfdhcp bugfix
Message-ID: <[email protected]>
Content-Type: text/plain; charset="utf-8"
Ok can you merge my pull request? Are you happy with it? Any feedback? I ran
`make test` and it seems that it is ok. There are some flaky unit tests
happening but I don?t think they are introduced by my commits.
--
Angelo Failla
Cluster Infrastructure - Dublin
[email protected]
On 4/7/16, 10:44 PM, "Francis Dupont" <[email protected]> wrote:
>Even I cannot see how it could matter I agree the spec says the XID
>is copied from the server OFFER so it is a bug.
>
>Regards
>
>Francis Dupont <[email protected]>
------------------------------
Message: 4
Date: Fri, 8 Apr 2016 13:31:37 +0200
From: Marcin Siodelski <[email protected]>
To: Angelo Failla <[email protected]>, Francis Dupont <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [kea-dev] perfdhcp bugfix
Message-ID: <[email protected]>
Content-Type: text/plain; charset=utf-8
Angelo,
Thanks for the pull request.
I left a comment in your pull request
(https://github.com/isc-projects/kea/pull/19)
>From my reading of RFC3315, the transaction id should not be copied
between Advertise and Request, but generated for each message sent.
Do you have any issues in your testbed for DHCPv6, related to
transaction id being regenerated for each message sent? Or you have
included the change for DHCPv6 for symmetry with DHCPv4?
Marcin
On 07.04.2016 23:50, Angelo Failla wrote:
> Ok can you merge my pull request? Are you happy with it? Any feedback? I ran
> `make test` and it seems that it is ok. There are some flaky unit tests
> happening but I don?t think they are introduced by my commits.
>
------------------------------
Message: 5
Date: Fri, 8 Apr 2016 11:38:00 +0000
From: Angelo Failla <[email protected]>
To: Marcin Siodelski <[email protected]>, Francis Dupont
<[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [kea-dev] perfdhcp bugfix
Message-ID: <[email protected]>
Content-Type: text/plain; charset="utf-8"
On 4/8/16, 12:31 PM, "Marcin Siodelski" <[email protected]> wrote:
>Angelo,
>
>Thanks for the pull request.
>
>I left a comment in your pull request
>(https://github.com/isc-projects/kea/pull/19)
>
>From my reading of RFC3315, the transaction id should not be copied
>between Advertise and Request, but generated for each message sent.
In theory (from what I can see in pcap files I take regularly in our prod env)
the transaction ID should be generated by the client when sending DISCOVER, and
should be used for the duration of the 4-way handshake. Which perfdhcp is not
doing (it generates a different ID when sending REQUEST) according to my tests.
I now that in theory it?s against RFCs to use the TXID provided by the server.
But this is a test tool, I guess you can make an exception, I don?t know
perfdhcp code well enough? I think generateTransid() should be only called once
per session and used for the duration of the session.
>
>Do you have any issues in your testbed for DHCPv6, related to
>transaction id being regenerated for each message sent? Or you have
>included the change for DHCPv6 for symmetry with DHCPv4?
No issues in prod per se, we are writing a DHCP relayer/loadbalancer internally
in Go to solve some issues we have, we balance things hashing on GIADDR + TXID,
and we were using perfdhcp to test the solution when I found out about the
current behavior.
And yes I included the changes for DHCPv6 for symmetry.
------------------------------
_______________________________________________
kea-dev mailing list
[email protected]
https://lists.isc.org/mailman/listinfo/kea-dev
End of kea-dev Digest, Vol 25, Issue 1
**************************************
