<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML xmlns="http://www.w3.org/TR/REC-html40" xmlns:m =
"http://schemas.microsoft.com/office/2004/12/omml" xmlns:w =
"urn:schemas-microsoft-com:office:word" xmlns:o =
"urn:schemas-microsoft-com:office:office" xmlns:v =
"urn:schemas-microsoft-com:vml"><HEAD>
<META content="text/html; charset=us-ascii" http-equiv=Content-Type>
<STYLE>@font-face {
font-family: Cambria Math;
}
@font-face {
font-family: Calibri;
}
@font-face {
font-family: Tahoma;
}
@page Section1 {size: 8.5in 11.0in; margin: 1.0in 1.0in 1.0in 1.0in; }
P.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt
}
LI.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt
}
DIV.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt
}
A:link {
COLOR: blue; TEXT-DECORATION: underline; mso-style-priority: 99
}
SPAN.MsoHyperlink {
COLOR: blue; TEXT-DECORATION: underline; mso-style-priority: 99
}
A:visited {
COLOR: purple; TEXT-DECORATION: underline; mso-style-priority: 99
}
SPAN.MsoHyperlinkFollowed {
COLOR: purple; TEXT-DECORATION: underline; mso-style-priority: 99
}
P.MsoPlainText {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt; mso-style-priority: 99; mso-style-link: "Plain Text Char"
}
LI.MsoPlainText {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt; mso-style-priority: 99; mso-style-link: "Plain Text Char"
}
DIV.MsoPlainText {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Times New Roman","serif"; COLOR: black; FONT-SIZE: 12pt; mso-style-priority: 99; mso-style-link: "Plain Text Char"
}
PRE {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Courier New"; FONT-SIZE: 10pt; mso-style-priority: 99; mso-style-link: "HTML Preformatted Char"
}
SPAN.EmailStyle17 {
FONT-FAMILY: "Times New Roman","serif"; COLOR: #1f497d; mso-style-type: personal-reply
}
SPAN.HTMLPreformattedChar {
FONT-FAMILY: "Courier New"; mso-style-priority: 99; mso-style-link: "HTML Preformatted"; mso-style-name: "HTML Preformatted Char"
}
SPAN.PlainTextChar {
FONT-FAMILY: "Calibri","sans-serif"; COLOR: black; mso-style-priority: 99; mso-style-link: "Plain Text"; mso-style-name: "Plain Text Char"
}
.MsoChpDefault {
FONT-SIZE: 10pt; mso-style-type: export-only
}
DIV.Section1 {
page: Section1
}
</STYLE>
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
<META name=GENERATOR content="MSHTML 8.00.6001.18702"></HEAD>
<BODY lang=EN-US link=blue bgColor=white vLink=purple>
<DIV><FONT color=#0000ff size=2 face=Arial><SPAN class=529514220-01072009>Thanks
for your info - <FONT color=#000000><FONT face=Tahoma><STRONG>
</STRONG>Rupert Dance.</FONT></FONT></SPAN></FONT></DIV>
<DIV><FONT color=#000000 size=2 face=Tahoma><SPAN
class=529514220-01072009></SPAN></FONT> </DIV>
<DIV><FONT color=#000000 size=2 face=Tahoma><SPAN
class=529514220-01072009>However it seems you are talking about Linux, while we
were talking about windows, o r am I missing something?</SPAN></FONT></DIV>
<DIV><FONT color=#000000 size=2 face=Tahoma><SPAN
class=529514220-01072009></SPAN></FONT> </DIV>
<DIV><FONT color=#000000 size=2 face=Tahoma><SPAN
class=529514220-01072009>Thanks</SPAN></FONT></DIV>
<DIV><FONT color=#000000 size=2 face=Tahoma><SPAN
class=529514220-01072009>Tzachi</SPAN></FONT></DIV><BR>
<BLOCKQUOTE
style="BORDER-LEFT: #0000ff 2px solid; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px"
dir=ltr>
<DIV dir=ltr lang=en-us class=OutlookMessageHeader align=left>
<HR tabIndex=-1>
<FONT size=2 face=Tahoma><B>From:</B> ofw-bounces@lists.openfabrics.org
[mailto:ofw-bounces@lists.openfabrics.org] <B>On Behalf Of </B>Rupert
Dance<BR><B>Sent:</B> Wednesday, July 01, 2009 7:27 PM<BR><B>To:</B> 'David
Brean'; ofw@lists.openfabrics.org<BR><B>Subject:</B> RE: [ofw] ping on
WinOF<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV class=Section1>
<P class=MsoNormal><SPAN style="COLOR: #1f497d">We have seen issues with IPoIB
in datagram mode particularly when you use a large size (8192 and greater).
This was reported to the OFA <A
href="https://bugs.openfabrics.org/show_bug.cgi?id=1287">Bugzilla Bug #
1287</A>. Yosef Etigin looked into this and suggested a workaround that did
affect the first packet drop. Here is his comment:<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: ; FONT-SIZE: 10pt" New??
Courier>It is a network stack limitation and not related ipoib in
particular.<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: ; FONT-SIZE: 10pt" New??
Courier>There's a limit (default = 3) on number of pending skb's before a
neighbour is<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: ; FONT-SIZE: 10pt" New??
Courier>resolved. You can increase it with sysctl
net.ipv4.neigh.ib0.unres_qlen.<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: ; FONT-SIZE: 10pt" New??
Courier>Obviously, same thing happens with Ethernet
interface.<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d">When testing at UNH-IOL for
the Logo program, this is what we did:<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<P class=MsoPlainText>After working with Sasha Khapyorsky on this issue we
have a working fix. To further explain the situation, the large packet sizes
we are using are overflowing the buffers so there is no room to append the arp
request on to the beginning of the cmd. This results in a dropped packet
because the system doesn't know how to get to the destination due to an empty
arp table. The fix, increase the buffer size via:<o:p></o:p></P>
<P class=MsoPlainText><o:p> </o:p></P>
<P class=MsoPlainText>sysctl net.ipv4.neigh.ib0.unres_qlen=17 # default is the
value 3<o:p></o:p></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d">Thanks<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d">Rupert
Dance<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="COLOR: #1f497d"><o:p> </o:p></SPAN></P>
<DIV>
<DIV
style="BORDER-BOTTOM: medium none; BORDER-LEFT: medium none; PADDING-BOTTOM: 0in; PADDING-LEFT: 0in; PADDING-RIGHT: 0in; BORDER-TOP: #b5c4df 1pt solid; BORDER-RIGHT: medium none; PADDING-TOP: 3pt">
<P class=MsoNormal><B><SPAN style="FONT-FAMILY: ; FONT-SIZE: 10pt"
color:windowtext? Tahoma?,?sans-serif?;>From:</SPAN></B><SPAN
style="FONT-FAMILY: ; FONT-SIZE: 10pt" Tahoma?,?sans-serif?;color:windowtext?>
ofw-bounces@lists.openfabrics.org [mailto:ofw-bounces@lists.openfabrics.org]
<B>On Behalf Of </B>David Brean<BR><B>Sent:</B> Wednesday, July 01, 2009 11:39
AM<BR><B>To:</B> ofw@lists.openfabrics.org<BR><B>Subject:</B> [ofw] ping on
WinOF<o:p></o:p></SPAN></P></DIV></DIV>
<P class=MsoNormal><o:p> </o:p></P>
<P style="MARGIN-BOTTOM: 12pt" class=MsoNormal>Hello,<BR><BR>An internal
customer is using WinOF 2.0.X and has reported to me the following behavior
related to IPoIB and ping:<BR><BR><I>Do you have any ideas on why windows 2008
client with HCA may first timeout ping to other clients on the
fabric?<BR><BR>Initially ping fails but then starts working.<BR><BR>Example
: Ping is invoked three times successfully.<BR><BR>C:\GRITS>ping -a
192.168.100.235<BR><BR>Pinging 192.168.100.235 with 32 bytes of
data:<BR>Request timed out.<BR>Request timed out.<BR>Request timed
out.<BR>Request timed out.<BR><BR>Ping statistics for
192.168.100.235:<BR> Packets: Sent = 4, Received = 0, Lost = 4
(100% loss),<BR><BR>C:\GRITS>ping -a 192.168.100.235<BR><BR>Pinging
192.168.100.235 with 32 bytes of data:<BR>Request timed out.<BR>Request timed
out.<BR>Reply from 192.168.100.235: bytes=32 time<1ms TTL=255<BR>Reply from
192.168.100.235: bytes=32 time<1ms TTL=255<BR><BR>Ping statistics for
192.168.100.235:<BR> Packets: Sent = 4, Received = 2, Lost = 2
(50% loss),<BR>Approximate round trip times in milli-seconds:<BR>
Minimum = 0ms, Maximum = 0ms, Average = 0ms<BR><BR>C:\GRITS>ping -a
192.168.100.235<BR><BR>Pinging 192.168.100.235 with 32 bytes of data:<BR>Reply
from 192.168.100.235: bytes=32 time<1ms TTL=255<BR>Reply from
192.168.100.235: bytes=32 time<1ms TTL=255<BR>Reply from 192.168.100.235:
bytes=32 time<1ms TTL=255<BR>Reply from 192.168.100.235: bytes=32
time<1ms TTL=255<BR><BR>Ping statistics for
192.168.100.235:<BR> Packets: Sent = 4, Received = 4, Lost = 0 (0%
loss),<BR>Approximate round trip times in milli-seconds:<BR>
Minimum = 0ms, Maximum = 0ms, Average = 0ms<BR><BR>Then we are good for
sometime before this starts again if network is idle on the
fabric.<BR></I><BR>Has this sort of behavior been observed before? The
Linux and Solaris nodes sharing the same IP subnet appear to be behaving
normally. Windows server is the "out-of-the-box" configuration with
Voltaire switch configured with only the default partition
(0xFFFF).<BR><BR>-David<o:p></o:p></P></DIV></BLOCKQUOTE></BODY></HTML>