Re: [Freeipmi-devel] ipmi_ctx_find_inband: device already open

[Albert Chu]

Hi Kenneth,

On Tue, 2010-09-14 at 08:06 -0700, Kenneth Stailey wrote:
> Hi Al,
> 
> We had two HP ProLiant DL145 G1 servers with this issue.
> 
> One of them happens to be our primary OpenLDAP server so I cannot reboot it 
> without change control, etc.  It is running Ubuntu 8.04.4.  The kipmi0 kernel 
> thread is getting around 92% CPU continuously.  This has been the case for a 
> long time and I have not been able to determine the cause of this.  The 
> server is functioning normally otherwise.
> 
> The other one with the IPMI issue was rebooted since it is a secondary 
> server.  kipmi0 is not "spinning" the CPU on that one.  The patch
> 
> When I reboot the one with the kimpi0 thread issue I'll retest it and let you 
> know.
> 
> When I run without the modules loaded via "rmmod ipmi_si ipmi_devintf 
> ipmi_msghandler" and with the patch on the other one I see:
> 
> $ sudo /opt/freeipmi/sbin/ipmi-sensors
> ipmi-openipmi-driver.c: 379: ipmi_openipmi_ctx_io_init: errno '' (2)
> ipmi-api.c: 1004: ipmi_ctx_open_inband: error 'device not found' (5)
> ipmi-sunbmc-driver.c: 330: ipmi_sunbmc_ctx_io_init: errno '' (2)
> ipmi-api.c: 1031: ipmi_ctx_open_inband: error 'device not found' (5)
> ipmi-api.c: 849: ipmi_ctx_open_inband: error 'device not supported' (17)
> ipmi-locate-dmidecode.c: 562: ipmi_locate_dmidecode_get_device_info: error 
> 'internal system error' (6)
> ipmi-locate-smbios.c: 411: _copy_ipmi_dev_info: error 'internal system error' 
> (6)
> ipmi-locate-pci.c: 301: ipmi_locate_pci_get_device_info: error 'internal 
> system error' (6)
> ipmi-locate.c: 161: _ipmi_locate_get_device_info: error 'internal system 
> error' (6)
> ipmi-api.c: 849: ipmi_ctx_open_inband: error 'device not supported' (17)
> ipmi-locate-dmidecode.c: 562: ipmi_locate_dmidecode_get_device_info: error 
> 'internal system error' (6)
> ipmi-locate-smbios.c: 411: _copy_ipmi_dev_info: error 'internal system error' 
> (6)
> ipmi-locate-pci.c: 301: ipmi_locate_pci_get_device_info: error 'internal 
> system error' (6)
> ipmi-ssif-driver.c: 683: ipmi_ssif_ctx_io_init: errno '' (2)
> ipmi-api.c: 977: ipmi_ctx_open_inband: error 'device not found' (9)
> ipmi-api.c: 1994: ipmi_ctx_close: error 'device not open' (16)
> could not find inband device

This trace is the same as when the patch wasn't applied.  Are you sure
you applied the patch on this particular version?

On the one w/ the patch applied, could you send a --debug output?  Maybe
there is something w/n the IPMI protocol I'm missing.

Thanks,

Al

> address@hidden:~/src/freeipmi-0.8.9$ /opt/freeipmi/sbin/ipmi-sensors --version
> ipmi-sensors - 0.8.9
> Copyright (C) 2003-2010 FreeIPMI Core Team
> This program is free software; you may redistribute it under the terms of
> the GNU General Public License.  This program has absolutely no warranty.
> 
> I reload the modules and it works again.
> 
> Thanks,
> Ken
> 
> --- On Mon, 9/13/10, Albert Chu <address@hidden> wrote:
> 
> > From: Albert Chu <address@hidden>
> > Subject: Re: [Freeipmi-devel] ipmi_ctx_find_inband: device already open
> > To: "Kenneth Stailey" <address@hidden>
> > Cc: "address@hidden" <address@hidden>
> > Date: Monday, September 13, 2010, 1:57 PM
> > Hey Kenneth,
> > 
> > This is a different problem, so the other problem was
> > fixed.
> > 
> > As for this one, it appears the "Get SDR Repository Info"
> > command just
> > isn't supported on your motherboard.  But this is
> > working when the
> > kernel drivers are loaded?
> > 
> > Al
> > 
> > On Mon, 2010-09-13 at 10:30 -0700, Kenneth Stailey wrote:
> > > Patch applied and modules not loaded
> > > 
> > > $ sudo /opt/freeipmi/sbin/ipmi-sensors
> > > ipmi-api-util.c: 253: api_ipmi_cmd: error 'Cannot
> > execute command. Command, or request parameter(s), not
> > supported in present state.' (213)
> > > ipmi-sdr-repository-cmds-api.c: 89:
> > ipmi_cmd_get_sdr_repository_info: error 'bad completion
> > code' (22)
> > > ipmi-sdr-cache-common.c: 100: ipmi_sdr_cache_info:
> > error 'internal IPMI error' (20)
> > > ipmi_sdr_cache_open:
> > /root/.freeipmi/sdr-cache/sdr-cache-molybdenum.localhost:
> > internal IPMI error
> > > 
> > > --- On Mon, 9/13/10, Albert Chu <address@hidden>
> > wrote:
> > > 
> > > > From: Albert Chu <address@hidden>
> > > > Subject: Re: [Freeipmi-devel]
> > ipmi_ctx_find_inband: device already open
> > > > To: "Kenneth Stailey" <address@hidden>
> > > > Cc: "address@hidden"
> > <address@hidden>
> > > > Date: Monday, September 13, 2010, 1:14 PM
> > > > Hi Kenneth,
> > > > 
> > > > If you could try the patch w/o those modules
> > installed I'd
> > > > appreciate
> > > > it.  In the function, I "probe" looking for
> > an
> > > > interface to do inband
> > > > IPMI with.  The below gets around the
> > problem by
> > > > finding an interface
> > > > earlier in the probe, so the bug is never
> > reached, but it's
> > > > still there.
> > > > 
> > > > Thanks,
> > > > Al
> > > > 
> > > > On Mon, 2010-09-13 at 10:01 -0700, Kenneth
> > Stailey wrote:
> > > > > I got it working without the patch.  I
> > had to run
> > > > modprobe to install these:
> > > > > 
> > > > > ipmi_si       
> >    
> > > >     57164  0 
> > > > > ipmi_devintf   
> >    
> > > >    20624  0 
> > > > > ipmi_msghandler       
> > 50680 
> > > > 2 ipmi_si,ipmi_devintf
> > > > > 
> > > > > 
> > > > > --- On Mon, 9/13/10, Albert Chu <address@hidden>
> > > > wrote:
> > > > > 
> > > > > > From: Albert Chu <address@hidden>
> > > > > > Subject: Re: [Freeipmi-devel]
> > > > ipmi_ctx_find_inband: device already open
> > > > > > To: "Kenneth Stailey" <address@hidden>
> > > > > > Cc: "address@hidden"
> > > > <address@hidden>
> > > > > > Date: Monday, September 13, 2010, 12:46
> > PM
> > > > > > Hey Kenneth,
> > > > > > 
> > > > > > Doh!  I didn't fall through on a
> > error case
> > > > > > correctly.  Does the
> > > > > > attached patch fix the problem?
> > > > > > 
> > > > > > cd freeipmi-0-8-9/
> > > > > > patch -p0 < fallthrough.patch
> > > > > > 
> > > > > > Al
> > > > > > 
> > > > > > On Mon, 2010-09-13 at 05:09 -0700,
> > Kenneth
> > > > Stailey wrote:
> > > > > > > 
> > > > > > > $ sudo
> > /opt/freeipmi/sbin/ipmi-sensors
> > > > > > > ipmi-openipmi-driver.c: 379:
> > > > > > ipmi_openipmi_ctx_io_init: errno ''
> > (2)
> > > > > > > ipmi-api.c: 1004:
> > ipmi_ctx_open_inband:
> > > > error 'device
> > > > > > not found' (5)
> > > > > > > ipmi-sunbmc-driver.c: 330:
> > > > ipmi_sunbmc_ctx_io_init:
> > > > > > errno '' (2)
> > > > > > > ipmi-api.c: 1031:
> > ipmi_ctx_open_inband:
> > > > error 'device
> > > > > > not found' (5)
> > > > > > > ipmi-api.c: 849:
> > ipmi_ctx_open_inband: error
> > > > 'device
> > > > > > not supported' (17)
> > > > > > > ipmi-locate-dmidecode.c: 562:
> > > > > > ipmi_locate_dmidecode_get_device_info:
> > error
> > > > 'internal
> > > > > > system error' (6)
> > > > > > > ipmi-locate-smbios.c: 411:
> > > > _copy_ipmi_dev_info: error
> > > > > > 'internal system error' (6)
> > > > > > > ipmi-locate-pci.c: 301:
> > > > > > ipmi_locate_pci_get_device_info: error
> > 'internal
> > > > system
> > > > > > error' (6)
> > > > > > > ipmi-locate.c: 161:
> > > > _ipmi_locate_get_device_info:
> > > > > > error 'internal system error' (6)
> > > > > > > ipmi-api.c: 778:
> > ipmi_ctx_open_inband: error
> > > > 'device
> > > > > > already open' (15)
> > > > > > > ipmi_ctx_find_inband: device
> > already open
> > > > > > > 
> > > > > > > --- On Fri, 9/10/10, Albert Chu
> > <address@hidden>
> > > > > > wrote:
> > > > > > > 
> > > > > > > > From: Albert Chu <address@hidden>
> > > > > > > > Subject: Re:
> > [Freeipmi-devel]
> > > > > > ipmi_ctx_find_inband: device already
> > open
> > > > > > > > To: "Kenneth Stailey" <address@hidden>
> > > > > > > > Cc: "address@hidden"
> > > > > > <address@hidden>
> > > > > > > > Date: Friday, September 10,
> > 2010, 5:28
> > > > PM
> > > > > > > > Hey Kenneth,
> > > > > > > > 
> > > > > > > > I looked through to code to
> > see what
> > > > might
> > > > > > possibly cause
> > > > > > > > this, but I
> > > > > > > > can't find anything.  Do
> > you think
> > > > you could
> > > > > > download
> > > > > > > > the .tar.gz and
> > > > > > > > compile with debugging to get
> > me some
> > > > extra
> > > > > > info?  To
> > > > > > > > compile w/
> > > > > > > > debugging, specify
> > --enable-debug and
> > > > > > --enable-trace in
> > > > > > > > ./configure.
> > > > > > > > 
> > > > > > > > Thanks,
> > > > > > > > Al
> > > > > > > > 
> > > > > > > > On Fri, 2010-09-10 at 12:57
> > -0700,
> > > > Kenneth
> > > > > > Stailey wrote:
> > > > > > > > > $ lsb_release -d
> > > > > > > > > Description: 
> >   Ubuntu
> > > > 8.04.4 LTS
> > > > > > > > > 
> > > > > > > > > $ uname -srv
> > > > > > > > > Linux 2.6.24-28-server
> > #1 SMP Wed
> > > > Aug 25
> > > > > > 14:46:03 UTC
> > > > > > > > 2010
> > > > > > > > > 
> > > > > > > > > $ echo `sudo dmidecode
> > -s
> > > > > > system-manufacturer ; sudo
> > > > > > > > dmidecode -s
> > system-product-name`
> > > > > > > > > HP ProLiant DL145 G1
> > > > > > > > > 
> > > > > > > > > $ sudo ipmi-sensors
> > --version
> > > > > > > > > ipmi-sensors - 0.8.9
> > > > > > > > > Copyright (C) 2003-2010
> > FreeIPMI
> > > > Core Team
> > > > > > > > > This program is free
> > software; you
> > > > may
> > > > > > redistribute it
> > > > > > > > under the terms of
> > > > > > > > > the GNU General Public
> > > > License.  This
> > > > > > program has
> > > > > > > > absolutely no warranty.
> > > > > > > > > 
> > > > > > > > > $ sudo ipmi-sensors
> > > > > > > > > ipmi_ctx_find_inband:
> > device
> > > > already open
> > > > > > > > > 
> > > > > > > > > Please help, thanks.
> > > > > > > > > 
> > > > > > > > > 
> > > > > > > > >
> > > > > >
> > _______________________________________________
> > > > > > > > > Freeipmi-devel mailing
> > list
> > > > > > > > > address@hidden
> > > > > > > > > http://*****lists.gnu.org/mailman/listinfo/freeipmi-devel
> > > > > > > > > 
> > > > > > > > -- 
> > > > > > > > Albert Chu
> > > > > > > > address@hidden
> > > > > > > > Computer Scientist
> > > > > > > > High Performance Systems
> > Division
> > > > > > > > Lawrence Livermore National
> > Laboratory
> > > > > > > > 
> > > > > > > > 
> > > > > > > 
> > > > > > -- 
> > > > > > Albert Chu
> > > > > > address@hidden
> > > > > > Computer Scientist
> > > > > > High Performance Systems Division
> > > > > > Lawrence Livermore National Laboratory
> > > > > > 
> > > > > 
> > > > -- 
> > > > Albert Chu
> > > > address@hidden
> > > > Computer Scientist
> > > > High Performance Systems Division
> > > > Lawrence Livermore National Laboratory
> > > > 
> > > > 
> > > 
> > -- 
> > Albert Chu
> > address@hidden
> > Computer Scientist
> > High Performance Systems Division
> > Lawrence Livermore National Laboratory
> > 
> > 
> 
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory