Date: Wed, 12 Dec 2007 06:07:28 -0800 (PST)
From: Dan Norris <dannorris@dannorris.com>
Subject: Re: Linux NIC bonding
To: Oracle L <oracle-l@freelists.org>
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="0-109342015-1197468448=:90031"
Message-ID: <392142.90031.qm@web35408.mail.mud.yahoo.com>
Sender: oracle-l-bounce@freelists.org
Errors-to: oracle-l-bounce@freelists.org
Precedence: normal
Reply-to: dannorris@dannorris.com
--0-109342015-1197468448=:90031
Content-Type: text/plain; charset=us-ascii

It was my understanding that active-active required trunking capability on the switches so that the switches know that both of the network ports are going to the same host. I don't know if that capability is available on these switches. That's why I chose active-passive. 

As for Oracle support, I know that Oracle's biggest issue is supporting kernels that are not tainted. I was thinking of the 3rd party intel or broadcom software that the customer is considering putting on these nodes--not sure if or what Oracle would say about that. If anyone has experience with NIC teaming software from Intel or Broadcom on Linux, please let me know your experiences with it.

Thanks for all the replies. Please let me know if I'm wrong about my assumption above that special switch trunking configuration is needed to enable active-active. I believe it is required per the bonding.txt file in the kernel source that says this in the sections about active-active: "This mode requires the switch to have the appropriate ports configured for "etherchannel" or "trunking."" We're trying to do all this stuff without doing any real configuration on the switches.

Dan

----- Original Message ----
From: goran bogdanovic <goran00@gmail.com>
To: Oracle L <oracle-l@freelists.org>
Sent: Wednesday, December 12, 2007 5:47:51 AM
Subject: Re: Linux NIC bonding


I have to correct myself...
I have recalled my memories for Linux bonding and 'active-backup' is possible (I haven't used it myself)...when you set mode parameter when defining bonding drivers here are the possibilities:


mode 0 = active-active (round-robin rule when transmitting packets)
mode 1 = active-passive
mode 2 = indicates either of them can be used


On Dec 12, 2007 10:28 AM, goran bogdanovic <
goran00@gmail.com> wrote:
Hi Dan,

I am not sure if 'active-passive' bonding under Linux is possible but why would you use 'active-passive' configuration? With 'active-active' you get load-balancing plus failover capability.


>It's questionable how Oracle would view that from a support perspective (if they care at all).

As long as bonding driver does not taint the kernel, Oracle will support it.


cheers,
goran


On Dec 12, 2007 6:22 AM, Dan Norris <dannorris@dannorris.com> wrote:


I'm building a cluster using Oracle Clusterware and have configured redundant pairs of NICs on the Linux servers using the Linux bonding module (active-passive). I know how to configure the Linux end of things and that's working fine. However, the switches used by this customer (they're Dell branded--not sure who really makes the guts) have been getting "confused" by bonding and our simple ping tests lose about 50% of the packets when we plug the redundant pairs into the two separate (trunked) switches. 


This is only peripherally related to Oracle--it's really a networking and server config issue, but I'm hoping that someone with more networking background than I have can explain what switch configuration will enable this to work properly. The customer has stated that this will never work (though I
 have done it before at other sites and it worked fine). The customer has stated that the Linux configuration must be two active NICs
 with a virtual IP and virtual MAC. I imagine that will work, but I'm not familiar with that configuration and I don't think Linux has native support for it, so it'd probably be a Broadcom or Intel software package. It's questionable how Oracle would view that from a support perspective (if they care at all). 


So, the question: Is there some "magical" switch setting that would possibly be missing in a fairly "default" configuration that might enable this to work properly? I figure it's a long shot asking a networking question on the Oracle list, but thanks for any pointers.


Dan


--0-109342015-1197468448=:90031
Content-Type: text/html; charset=us-ascii

<html><head><style type="text/css"><!-- DIV {margin:0px;} --></style></head><body><div style="font-family:arial,helvetica,sans-serif;font-size:10pt"><div style="font-family: arial,helvetica,sans-serif; font-size: 10pt;">It was my understanding that active-active required trunking capability on the switches so that the switches know that both of the network ports are going to the same host. I don't know if that capability is available on these switches. That's why I chose active-passive. <br><br>As for Oracle support, I know that Oracle's biggest issue is supporting kernels that are not tainted. I was thinking of the 3rd party intel or broadcom software that the customer is considering putting on these nodes--not sure if or what Oracle would say about that. If anyone has experience with NIC teaming software from Intel or Broadcom on Linux, please let me know your experiences with it.<br><br>Thanks for all the replies. Please let me know if I'm wrong about
 my assumption above that special switch trunking configuration is needed to enable active-active. I believe it is required per the bonding.txt file in the kernel source that says this in the sections about active-active: "This mode requires the switch to have the appropriate ports configured for "etherchannel" or "trunking."" We're trying to do all this stuff without doing any real configuration on the switches.<br><br>Dan<br><br><div style="font-family: times new roman,new york,times,serif; font-size: 12pt;">----- Original Message ----<br>From: goran bogdanovic &lt;goran00@gmail.com&gt;<br>To: Oracle L &lt;oracle-l@freelists.org&gt;<br>Sent: Wednesday, December 12, 2007 5:47:51 AM<br>Subject: Re: Linux NIC bonding<br><br>
I have to correct myself...<br>I have recalled my memories for Linux bonding and 'active-backup' is possible (I haven't used it myself)...when you set mode parameter when defining bonding drivers here are the possibilities:
<br><br>mode 0 = active-active (round-robin rule when transmitting packets)<br>mode 1 = active-passive<br>mode 2 = indicates either of them can be used<br><br><br><div class="gmail_quote">On Dec 12, 2007 10:28 AM, goran bogdanovic &lt;
<a rel="nofollow" ymailto="mailto:goran00@gmail.com" target="_blank" href="mailto:goran00@gmail.com">goran00@gmail.com</a>&gt; wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi Dan,<br><br>I am not sure if 'active-passive' bonding under Linux is possible but why would you use 'active-passive' configuration? With 'active-active' you get load-balancing plus failover capability.
<div class="Ih2E3d"><br><br>&gt;It's questionable how Oracle would view that from a support perspective (if they care at all).<br></div>As long as bonding driver does not taint the kernel, Oracle will support it.<br>
<br>cheers,<br><font color="#888888">goran</font><div><div></div><div class="Wj3C7c"><br><br>
<br><div class="gmail_quote">On Dec 12, 2007 6:22 AM, Dan Norris &lt;<a rel="nofollow" ymailto="mailto:dannorris@dannorris.com" target="_blank" href="mailto:dannorris@dannorris.com">dannorris@dannorris.com</a>&gt; wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">

<div><div style="font-family: arial,helvetica,sans-serif; font-size: 10pt;">I'm building a cluster using Oracle Clusterware and have configured redundant pairs of NICs on the Linux servers using the Linux bonding module (active-passive). I know how to configure the Linux end of things and that's working fine. However, the switches used by this customer (they're Dell branded--not sure who really makes the guts) have been getting "confused" by bonding and our simple ping tests lose about 50% of the packets when we plug the redundant pairs into the two separate (trunked) switches. 
<br><br>This is only peripherally related to Oracle--it's really a networking and server config issue, but I'm hoping that someone with more networking background than I have can explain what switch configuration will enable this to work properly. The customer has stated that this will never work (though I
 have done it before at other sites and it worked fine). The customer has stated that the Linux configuration must be two active NICs<br>&nbsp;with a virtual IP and virtual MAC. I imagine that will work, but I'm not familiar with that configuration and I don't think Linux has native support for it, so it'd probably be a Broadcom or Intel software package. It's questionable how Oracle would view that from a support perspective (if they care at all). 
<br><br>So, the question: Is there some "magical" switch setting that would possibly be missing in a fairly "default" configuration that might enable this to work properly? I figure it's a long shot asking a networking question on the Oracle list, but thanks for any pointers.
<br><font color="#888888"><br>Dan<br></font></div></div></blockquote></div><br>
</div></div></blockquote></div><br>
</div><br></div></div></body></html>
--0-109342015-1197468448=:90031--
--
http://www.freelists.org/webpage/oracle-l