From vivek@m1e.net  Fri Jan 31 16:32:47 2014
Return-Path: <vivek@m1e.net>
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1])
	(using TLSv1 with cipher ADH-AES256-SHA (256/256 bits))
	(No client certificate requested)
	by hub.freebsd.org (Postfix) with ESMTPS id 9E92F668
	for <FreeBSD-gnats-submit@freebsd.org>; Fri, 31 Jan 2014 16:32:47 +0000 (UTC)
Received: from yertle.kcilink.com (yertle-comcast.kcilink.com [74.92.149.58])
	(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))
	(No client certificate requested)
	by mx1.freebsd.org (Postfix) with ESMTPS id 735721159
	for <FreeBSD-gnats-submit@freebsd.org>; Fri, 31 Jan 2014 16:32:46 +0000 (UTC)
Received: from d06.m1e.net (neo.int.kcilink.com [192.168.7.50])
	by yertle.kcilink.com (Postfix) with SMTP id 785998A033
	for <FreeBSD-gnats-submit@freebsd.org>; Fri, 31 Jan 2014 11:32:40 -0500 (EST)
Received: (nullmailer pid 70800 invoked by uid 120);
	Fri, 31 Jan 2014 16:32:40 -0000
Message-Id: <1391185960.453017.70244.nullmailer@d06.m1e.net>
Date: Fri, 31 Jan 2014 11:32:40 -0500
From: Vivek Khera <vivek@khera.org>
Reply-To: Vivek Khera <vivek@khera.org>
To: FreeBSD-gnats-submit@freebsd.org
Subject: watchdog service causes BMC controller reset every 20-30 minutes on Supermicro X9DRW-3F
X-Send-Pr-Version: 3.114
X-GNATS-Notify:

>Number:         186304
>Category:       kern
>Synopsis:       [bmc] watchdog service causes BMC controller reset every 20-30 minutes on Supermicro X9DRW-3F
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    freebsd-net
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Jan 31 16:40:00 UTC 2014
>Closed-Date:    
>Last-Modified:  Sun May 04 04:56:04 UTC 2014
>Originator:     Vivek Khera
>Release:        FreeBSD 10.0-RELEASE amd64
>Organization:
>Environment:
System: FreeBSD d06.m1e.net 10.0-RELEASE FreeBSD 10.0-RELEASE #1 r261210: Tue Jan 28 10:36:57 EST 2014 vivek@devbox.int.kcilink.com:/n/yertle1/sources/usr10/obj/n/yertle1/sources/usr10/src/sys/KCI64 amd64

I have a pair of brand new servers on SuperMicro X9DRW-3F motherboards with
256GB of RAM. They have the latest BIOS (3.0a) and BMC controller (2.19.0)
firmware.

I run a custom kernel with the IPMI driver compiled in, detected as follows:

 kernel: ipmi0: <IPMI System Interface> port 0xca2,0xca3 on acpi0
 kernel: ipmi0: KCS mode found at io 0xca2 on acpi
 kernel: ipmi0: IPMI device rev. 1, firmware rev. 2.19, version 2.0
 kernel: ipmi0: Number of channels 3
 kernel: ipmi0: Attached watchdog

	
>Description:
	

The FreeBSD watchdog causes the BMC controller to reboot every 20-30 minutes,
logging this in the system log:

 kernel: ipmi0: KCS: Failed to read command
 kernel: ipmi0: KCS error: 09

where the error number varies. At this time, the virtual USB keyboard that the
BMC provides for the remote console disappears and reappears as well. There
are otherwise no ill effects on the running FreeBSD system.

>How-To-Repeat:
	

Start the watchdog service (/etc/rc.d/watchdog start) and wait 30 minutes.
System load has no effect. It can be an idle server or running heavy duty.


>Fix:

	


Do not run watchdog, or disable IPMI driver, or both.

As a side not, I notice that watchdog does not run on any of my systems that
do not have IPMI. This probably should be mentioned in the watchdog(8) man
page.
>Release-Note:
>Audit-Trail:

From: Doug White <dwhite@ixsystems.com>
To: bug-followup@FreeBSD.org,
 vivek@khera.org
Cc:  
Subject: Re: kern/186304: watchdog service causes BMC controller reset every 20-30 minutes on Supermicro X9DRW-3F
Date: Tue, 25 Mar 2014 18:55:01 -0700

 Vivek,
 
 Contact your systems vendor to see if you can upgrade to IPMI ver 2.35. =
 The upgrade from 2.19 is non-trivial however, thus the suggestion to =
 contact your vendor, but it might address this issue.=

From: Vick Khera <vivek@khera.org>
To: Doug White <dwhite@ixsystems.com>
Cc: "bug-followup@FreeBSD.org" <bug-followup@freebsd.org>
Subject: Re: kern/186304: watchdog service causes BMC controller reset every
 20-30 minutes on Supermicro X9DRW-3F
Date: Wed, 26 Mar 2014 10:41:20 -0400

 --001a11332d16469fe004f5837627
 Content-Type: text/plain; charset=ISO-8859-1
 
 Thanks. Looks like they released this newer IPMI firmware since I filed the
 PR. This will have to go down low on the to-do list since it requires a
 power cycle, and the servers in question are now in production.
 
 --001a11332d16469fe004f5837627--
Responsible-Changed-From-To: freebsd-bugs->freebsd-net 
Responsible-Changed-By: linimon 
Responsible-Changed-When: Sun May 4 04:55:48 UTC 2014 
Responsible-Changed-Why:  
Over to maintainer(s). 

http://www.freebsd.org/cgi/query-pr.cgi?pr=186304 
>Unformatted:
