[prev in list] [next in list] [prev in thread] [next in thread]
List: loadbalancing-l
Subject: RE: [load balancing] Alteon Ad3 - Cookie LB Persistence Question
From: "Kenneth Thurman" <kthurm1 () nortelnetworks ! com>
Date: 2002-05-29 19:13:52
[Download RAW message or body]
Allan,
I wasn't aware that all versions prior to version 9.0.x had issues with
cookie persistency, I do know that the functionality was limited. The Cookie
support that went into version 9.0+ Supports cookie Insert mode, and in
version 10 this was enhanced to allow the cookie that is inserted to be
either a permanent or temporary, in version 9 it could only be temporary.
IMHO using cookie Insert mode is much more desirable as it doesn't require
any interaction with the Server application. Where as the Passive cookie and
cookie rewrite do, (below I suggested rewrite as alternative, but really
meant to say insert).
In any case the bug that I mentioned is only problem if you are doing
cookie persistency and are using passive cookie and the cookie value you are
using is NOT the hex encoded RIP. There are no know issues using Cookie
Insert mode or Cookie rewrite, or if you are using passive cookie with the
cookie value being the Hex encoded RIP. And even then the bug can be
avoided by have a real server timeout set to a value that is longer than the
Web Transaction is expected to be, or adjusting the Slowage timer to achieve
the same thing. But the best way to avoid it is to use Cookie Insert mode.
Regards,
Ken
-----Original Message-----
From: Allan Liska [mailto:allan@allan.org]
Sent: Wednesday, May 29, 2002 1:46 PM
To: 'lb-l@vegan.net'
Subject: RE: [load balancing] Alteon Ad3 - Cookie LB Persistence Question
On Wed, 29 May 2002, Kenneth Thurman wrote:
>
> What you have discovered may be a known bug in the passive cookie
> mode of operation on the Alteon Web Switch in the Version 9.0.x WebOS
> version (and version 10.0.x as well). The symptom of the bug is that a
> passive cookie session will be persistent for 10mins then it will be
> loadbalanced to another real server based on the group metric. If this
> is what you are seeing then I would suggest that you use a workaround
> for this bug, and use cookie rewrite mode, instead of passive cookie.
> In this mode the Switch will create a cookie named Alteonp with an
> encoded value of the VIP/RIP and send it to the client, it is a
> temporary cookie only.
>
Are you saying there is another bug in the Alteon cookie processing
engine? Part of the impetus for us to upgrade to 9.0.41.5 was that we
were told there was bug with cookie persistency in all versions of WebOS
up to 9.0.41.3. But now there is another bug that affects version 9.0.x
and 10.0.x?
That seems a little ridiculous. In my opinion, WebOS has always had
shoddy cookie support when compared to F5 and Arrowpoint. If what you are
saying is true, then this trend appears to be continuing.
I don't want this to sound like an attack, because I have been extremely
happy with Alteon products overall. But I think cookie support is an
important functionality, and something that customers do insist on having,
so I don't understand why Alteon has had such a problem with implementing
it.
allan
--
Allan Liska
allan@allan.org
http://www.allan.org
____________________
The Load Balancing Mailing List
Unsubscribe: mailto:majordomo@vegan.net?body=unsubscribe%20lb-l
Archive: http://vegan.net/lb/archive
LBDigest: http://lbdigest.com
MRTG with SLB: http://vegan.net/MRTG
Hosted by: http://www.tokkisystems.com
[Attachment #3 (text/html)]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2654.89">
<TITLE>RE: [load balancing] Alteon Ad3 - Cookie LB Persistence Question</TITLE>
</HEAD>
<BODY>
<P><FONT SIZE=2>Allan,</FONT>
</P>
<P><FONT SIZE=2> I wasn't aware that all versions prior to version 9.0.x \
had issues with cookie persistency, I do know that the functionality was limited. The \
Cookie support that went into version 9.0+ Supports cookie Insert mode, and in \
version 10 this was enhanced to allow the cookie that is inserted to be either a \
permanent or temporary, in version 9 it could only be temporary. IMHO using cookie \
Insert mode is much more desirable as it doesn't require any interaction with the \
Server application. Where as the Passive cookie and cookie rewrite do, (below I \
suggested rewrite as alternative, but really meant to say insert). </FONT></P>
<P><FONT SIZE=2> In any case the bug that I mentioned is only problem if you \
are doing cookie persistency and are using passive cookie and the cookie value you \
are using is NOT the hex encoded RIP. There are no know issues using Cookie Insert \
mode or Cookie rewrite, or if you are using passive cookie with the cookie value \
being the Hex encoded RIP. And even then the bug can be avoided by have a real \
server timeout set to a value that is longer than the Web Transaction is expected to \
be, or adjusting the Slowage timer to achieve the same thing. But the best way to \
avoid it is to use Cookie Insert mode.</FONT></P>
<P><FONT SIZE=2>Regards,</FONT>
</P>
<P><FONT SIZE=2>Ken</FONT>
<BR><FONT SIZE=2> </FONT>
</P>
<P><FONT SIZE=2>-----Original Message-----</FONT>
<BR><FONT SIZE=2>From: Allan Liska [<A \
HREF="mailto:allan@allan.org">mailto:allan@allan.org</A>] </FONT> <BR><FONT \
SIZE=2>Sent: Wednesday, May 29, 2002 1:46 PM</FONT> <BR><FONT SIZE=2>To: \
'lb-l@vegan.net'</FONT> <BR><FONT SIZE=2>Subject: RE: [load balancing] Alteon Ad3 - \
Cookie LB Persistence Question</FONT> </P>
<BR>
<P><FONT SIZE=2>On Wed, 29 May 2002, Kenneth Thurman wrote:</FONT>
<BR><FONT SIZE=2>> </FONT>
<BR><FONT SIZE=2>> What you have discovered may be a known bug \
in the passive cookie </FONT> <BR><FONT SIZE=2>> mode of operation on the Alteon \
Web Switch in the Version 9.0.x WebOS </FONT> <BR><FONT SIZE=2>> version (and \
version 10.0.x as well). The symptom of the bug is that a </FONT> <BR><FONT \
SIZE=2>> passive cookie session will be persistent for 10mins then it will be \
</FONT> <BR><FONT SIZE=2>> loadbalanced to another real server based on the group \
metric. If this </FONT> <BR><FONT SIZE=2>> is what you are seeing then I would \
suggest that you use a workaround </FONT> <BR><FONT SIZE=2>> for this bug, and use \
cookie rewrite mode, instead of passive cookie. </FONT> <BR><FONT SIZE=2>> In this \
mode the Switch will create a cookie named Alteonp with an </FONT> <BR><FONT \
SIZE=2>> encoded value of the VIP/RIP and send it to the client, it is a </FONT> \
<BR><FONT SIZE=2>> temporary cookie only.</FONT> <BR><FONT SIZE=2>> \
</FONT> </P>
<P><FONT SIZE=2>Are you saying there is another bug in the Alteon cookie processing \
</FONT> <BR><FONT SIZE=2>engine? Part of the impetus for us to upgrade to \
9.0.41.5 was that we </FONT> <BR><FONT SIZE=2>were told there was bug with cookie \
persistency in all versions of WebOS </FONT> <BR><FONT SIZE=2>up to 9.0.41.3. \
But now there is another bug that affects version 9.0.x </FONT> <BR><FONT SIZE=2>and \
10.0.x? </FONT> </P>
<P><FONT SIZE=2>That seems a little ridiculous. In my opinion, WebOS has always \
had </FONT> <BR><FONT SIZE=2>shoddy cookie support when compared to F5 and \
Arrowpoint. If what you are </FONT> <BR><FONT SIZE=2>saying is true, then this \
trend appears to be continuing. </FONT> </P>
<P><FONT SIZE=2>I don't want this to sound like an attack, because I have been \
extremely </FONT> <BR><FONT SIZE=2>happy with Alteon products overall. But I \
think cookie support is an </FONT> <BR><FONT SIZE=2>important functionality, and \
something that customers do insist on having, </FONT> <BR><FONT SIZE=2>so I don't \
understand why Alteon has had such a problem with implementing </FONT> <BR><FONT \
SIZE=2>it.</FONT> </P>
<BR>
<P><FONT SIZE=2>allan</FONT>
<BR><FONT SIZE=2>-- </FONT>
<BR><FONT SIZE=2>Allan Liska</FONT>
<BR><FONT SIZE=2>allan@allan.org</FONT>
<BR><FONT SIZE=2><A HREF="http://www.allan.org" \
TARGET="_blank">http://www.allan.org</A></FONT> </P>
<P><FONT SIZE=2>____________________</FONT>
<BR><FONT SIZE=2>The Load Balancing Mailing List</FONT>
<BR><FONT SIZE=2>Unsubscribe: <A \
HREF="mailto:majordomo@vegan.net?body=unsubscribe%20lb-l">mailto:majordomo@vegan.net?body=unsubscribe%20lb-l</A></FONT>
<BR><FONT SIZE=2>Archive: <A \
HREF="http://vegan.net/lb/archive" \
TARGET="_blank">http://vegan.net/lb/archive</A></FONT> <BR><FONT \
SIZE=2>LBDigest: <A HREF="http://lbdigest.com" \
TARGET="_blank">http://lbdigest.com</A></FONT> <BR><FONT SIZE=2>MRTG with SLB: \
<A HREF="http://vegan.net/MRTG" TARGET="_blank">http://vegan.net/MRTG</A></FONT> \
<BR><FONT SIZE=2>Hosted by: <A \
HREF="http://www.tokkisystems.com" \
TARGET="_blank">http://www.tokkisystems.com</A></FONT> </P>
</BODY>
</HTML>
____________________
The Load Balancing Mailing List
Unsubscribe: mailto:majordomo@vegan.net?body=unsubscribe%20lb-l
Archive: http://vegan.net/lb/archive
LBDigest: http://lbdigest.com
MRTG with SLB: http://vegan.net/MRTG
Hosted by: http://www.tokkisystems.com
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic