Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Mar 2013 11:40:01 GMT
From:      david@dr.eclipse.co.uk
To:        freebsd-net@FreeBSD.org
Subject:   Re: kern/177139: [igb] igb drops ethernet ports 2 and 3
Message-ID:  <201303211140.r2LBe1wP059908@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/177139; it has been noted by GNATS.

From: david@dr.eclipse.co.uk
To: "Vogel Jack" <jack.vogel@intel.com>,
 bug-followup@FreeBSD.org
Cc:  
Subject: Re: kern/177139: [igb] igb drops ethernet ports 2 and 3
Date: Thu, 21 Mar 2013 11:02:49 +0000

 --=_839ebb776a4668cec96eeb4f50d4c37a
 Content-Type: text/plain; charset=UTF-8
 Content-Transfer-Encoding: quoted-printable
 
 =C2=A0=0A=0A Hi Jack,=0A=0AIn dmesg and /var/log/messages we see=0A=0Aig=
 b2: link state changed to UP=0Aigb2: link state changed to DOWN=0A=C2=A0=
 =0Aifconfig reports=0A=0A=C2=A0igb2: flags=3D8c02 metric 0 mtu 1500=0A=
 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 options=3D401bb=0A=C2=A0=C2=
 =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ether bc:30:5b:f6:12:de=0A=C2=A0=C2=A0=
 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 inet6 fe80::be30:5bff:fef6:12de%igb2 pref=
 ixlen 64=0Ascopeid 0x3=0A=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 inet=
  192.168.179.134 netmask 0xffffff00 broadcast=0A192.168.179.255=0A=C2=A0=
 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 nd6 options=3D29=0A=C2=A0=C2=A0=C2=
 =A0=C2=A0=C2=A0=C2=A0=C2=A0 media: Ethernet 1000baseT  (autoselect)=0A=
 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 status: no carrier=0A=0AThis=
  interface is being used exclusively for Postgresql database=0Areplicati=
 on from one server to another.=0ABoth servers do the same.=C2=A0 They ca=
 n stay up for anything from a few=0Aminutes to a few hours but always di=
 e.=0AOne major difference is that the two interfaces are connected with=
  a=0Acrossover (also tried straight through) cable.=0AThis is a configur=
 ation we have been using with great success for a=0Anumber of years with=
  broadcom adapters.=0AOpted for the Intel card on these servers as we ha=
 d trouble with the=0Alatest broadcom drivers. =0AI have tried disabling=
  msix as I read this could possibly help and=0Aalso tried disabling auto=
 select and manually configuring the=0Ainterfaces.=0ANone of these have h=
 elped.=C2=A0 igb0 and igb1 are connect to switches=0Aand have been stabl=
 e.=0ASo at the moment I have created a vlan on one of our switches and=
 =0Ahave plugged both servers into these vlan ports and they have=0Aremai=
 ned up over night.=0ACould it be that this network card does not like di=
 rect connections=0Aand is looking for a switch?=0A=0ADavid=0A=0A----- Or=
 iginal Message -----=0AFrom: "Vogel Jack" =0ATo:"bug-followup@FreeBSD.or=
 g" , "david@dr.eclipse.co.uk" =0ACc:=0ASent:Thu, 21 Mar 2013 00:17:03 +0=
 000=0ASubject:Re: kern/177139: [igb] igb drops ethernet ports 2 and 3=0A=
 =0A=09Please provide more details, are the ports described as=0A=E2=80=
 =98dropped=E2=80=99 just losing link, or some other symptoms as well.=C2=
 =A0=0AAlso, is there any different in the link-partner between the sets=
  of=0Aports? Is the activity different, etc=E2=80=A6=0A=0A=09=C2=A0=0A=
 =0A=09Jack=0A=0A=09=C2=A0=0A=0A=09
 
 --=_839ebb776a4668cec96eeb4f50d4c37a
 Content-Type: text/html; charset=UTF-8
 Content-Transfer-Encoding: quoted-printable
 
 <html><body>=C2=A0<br /><br />=09=09Hi Jack,<br /><br />In dmesg and /va=
 r/log/messages we see<br /><br />igb2: link state changed to UP<br />igb=
 2: link state changed to DOWN<br />=C2=A0<br />ifconfig reports<br /><br=
  />=C2=A0igb2: flags=3D8c02&lt;BROADCAST,OACTIVE,SIMPLEX,MULTICAST&gt; m=
 etric 0 mtu 1500<br />=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 options=
 =3D401bb&lt;RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,=
 TSO4,VLAN_HWTSO&gt;<br />=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ethe=
 r bc:30:5b:f6:12:de<br />=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 inet=
 6 fe80::be30:5bff:fef6:12de%igb2 prefixlen 64 scopeid 0x3<br />=C2=A0=C2=
 =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 inet 192.168.179.134 netmask 0xffffff0=
 0 broadcast 192.168.179.255<br />=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=
 =A0 nd6 options=3D29&lt;PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL&gt;<br />=
 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 media: Ethernet 1000baseT &lt=
 ;full-duplex&gt; (autoselect)<br />=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
 =C2=A0 status: no carrier<br /><br />This interface is being used exclus=
 ively for Postgresql database replication from one server to another.<br=
  />Both servers do the same.=C2=A0 They can stay up for anything from a=
  few minutes to a few hours but always die.<br />One major difference is=
  that the two interfaces are connected with a crossover (also tried stra=
 ight through) cable.<br />This is a configuration we have been using wit=
 h great success for a number of years with broadcom adapters.<br />Opted=
  for the Intel card on these servers as we had trouble with the latest b=
 roadcom drivers. <br />I have tried disabling msix as I read this could=
  possibly help and also tried disabling autoselect and manually configur=
 ing the interfaces.<br />None of these have helped.=C2=A0 igb0 and igb1=
  are connect to switches and have been stable.<br />So at the moment I h=
 ave created a vlan on one of our switches and have plugged both servers=
  into these vlan ports and they have remained up over night.<br />Could=
  it be that this network card does not like direct connections and is lo=
 oking for a switch?<br /><br />David<br /><blockquote><br />----- Origin=
 al Message -----<br /><div style=3D"width:100%;background:rgb(228,228,22=
 8);"><div style=3D"font-weight:bold;">From:</div> "Vogel Jack" &lt;jack.=
 vogel@intel.com&gt;</div><br /><div style=3D"font-weight:bold;">To:</div=
 >"bug-followup@FreeBSD.org" &lt;bug-followup@FreeBSD.org&gt;, "david@dr.=
 eclipse.co.uk" &lt;david@dr.eclipse.co.uk&gt;<br /><div style=3D"font-we=
 ight:bold;">Cc:</div><br /><div style=3D"font-weight:bold;">Sent:</div>T=
 hu, 21 Mar 2013 00:17:03 +0000<br /><div style=3D"font-weight:bold;">Sub=
 ject:</div>Re: kern/177139: [igb] igb drops ethernet ports 2 and 3<br />=
 <br /><br /><div class=3D"WordSection1">=0A<p class=3D"MsoNormal">Please=
  provide more details, are the ports described as =E2=80=98dropped=E2=80=
 =99 just losing link, or some other symptoms as well.=C2=A0 Also, is the=
 re any different in the link-partner between the sets of ports? Is the a=
 ctivity different, etc=E2=80=A6</p><p></p><p></p>=0A<p class=3D"MsoNorma=
 l"></p><p>=C2=A0</p><p></p>=0A<p class=3D"MsoNormal">Jack</p><p></p><p><=
 /p>=0A<p class=3D"MsoNormal"></p><p>=C2=A0</p><p></p>=0A</div>=0A=0A=0A=
 =0A</blockquote></body></html>
 
 --=_839ebb776a4668cec96eeb4f50d4c37a--
 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201303211140.r2LBe1wP059908>