Re: [Freeipmi-users] Help | Cannot Detect Nodes with IPMI

[Jeff Van]

Hi Al,

After tinkering around for a bit, we've successfully solved the problem! It
turned out to be a cabling issue (ie. our cables needed to all be connected
to an external switch, as opposed to a management switch).

Thank you again for all of your patience and support!

On Wed, Jan 10, 2018 at 12:03 PM, Albert Chu <address@hidden> wrote:

> Hi Jeff,
>
> On Wed, 2018-01-10 at 11:30 -0800, Jeff Van wrote:
> > Hi Al,
> >
> > Hmm, it seems as though I've configured everything correctly then.
> >
> > I also tried checking the ARP cache like you suggested, but in the
> > output, there were no IP/MAC addresses that matched the IP/MAC
> > address of the node I am trying to communicate with. However, if my
> > understanding of ARP is correct, that shouldn't be unusual, since my
> > head node has not been able to receive any kind of communications
> > from the node I am trying to communicate with.
>
> This is a good piece of data.  If the problem was on the server side
> (i.e. the IPMI chip it isn't responding to your client packets) then
> there should still be a cache entry on the client side. (This is
> assuming you have a typical setup where recent ARP entries are cached.)
>
> But the lack of an entry suggests this a client side problem.  It can't
> send packets at all because it cannot resolve the IP to a MAC address.
>
> You may want to check the ARP configuration on the BMC (on the server)
> itself.  You can see it with "ipmi-config --checkout".
>
> Al
>
> > I've also checked the ipmi-oem manpage, and I wasn't able to find
> > anything particularly useful.
> >
> > However, I noticed that when I use the command on the head node:
> > `sudo bmc-device --get-bmc-global-enables`, I receive the following
> > output:
> >
> > Receive Message Queue Interrupt : disabled
> > Event Message Buffer Full       : disabled
> > Event Message Buffer            : enabled
> > System Event Logging            : enabled
> > OEM 0                           : disabled
> > OEM 1                           : disabled
> > OEM 2                           : disabled
> >
> > Perhaps the fact that OEM 0-2 are disabled might provide a clue as to
> > what my problem could be?
> >
> > If you happen to have any other ideas on how I might go about
> > debugging this, or if there are other places to which I can direct my
> > questions, please let me know. Thank you again for taking the time to
> > help me with this!
> >
> > On Tue, Jan 9, 2018 at 5:32 PM, Albert Chu <address@hidden> wrote:
> > > Hi Jeff,
> > >
> > > On Tue, 2018-01-09 at 17:14 -0800, Jeff Van wrote:
> > > > Hi Al,
> > > >
> > > > Thank you for the response!
> > > >
> > > > As a follow up question, would you happen to be able to describe
> > > how
> > > > one might do BIOS configs to determine which port allows IPMI?
> > >
> > > Every motherboard is different, so you'll ultimately have to look
> > > into
> > > the product info for your board.
> > >
> > > > After searching on the internet on how to do it, I've tried the
> > > > following on the head node:
> > > > Pressed F2 to enter system setup after system Power-On Self-Test
> > > > (POST, or while it's booting up). However, after combing through
> > > the
> > > > settings options, I could not find any information about IPMI and
> > > > associated ports.  The closest thing I could find was under
> > > > "Integrated Devices", where it displayed the capabilities of the
> > > NIC
> > > > ports, which I've included a screenshot of, in this email.
> > > However,
> > > > none of the NIC ports display IPMI or idrac under the
> > > "capabilities"
> > > > section. Additionally, in my configurations, I have NIC Selection
> > > set
> > > > to Shared on both the head node and the node I am trying to
> > > > communicate with, in case that makes a difference.
> > >
> > > I haven't worked on a poweredge for years, so I can only go off
> > > recollection.  My recollection is somewhere in the BIOS is some
> > > type of
> > > "baseboard management" config and one could pick things like
> > > "dedicated", "shared" in the BIOS.  Take this with a grain of salt,
> > > as
> > > I may be confusing this other motherboards.
> > >
> > > In the ipmi-oem manpage, "shared" indicates IPMI only works on
> > > NIC1.
> > > So that seems like a reasonable config if you're using NIC1.
> > >
> > > Al
> > >
> > > > I've also tried pressing ctrl + e after POST, but that takes me
> > > to
> > > > configurations that are very similar to pressing F10.
> > > >
> > > > I've also tried pressing F10 to enter hardware configuration
> > > > settings, but I was not able to find any useful information there
> > > > either.
> > > >
> > > > If there is any other useful information or advice that you could
> > > > give me, I would very much appreciate it! I understand that you
> > > must
> > > > be busy, so thank you again for taking the time to help me with
> > > > this.
> > > >
> > > > Additionally, I have included a diagram of the physical ports on
> > > the
> > > > back of my Dell Poweredge R710. Currently, the cable connecting
> > > the
> > > > two nodes are plugged in the left-most port in at number 10 of
> > > the
> > > > diagram.
> > > >
> > > > On Mon, Jan 8, 2018 at 5:22 PM, Albert Chu <address@hidden>
> > > wrote:
> > > > > Hi Jeff,
> > > > >
> > > > > > I have double checked the BMC configuration for these nodes,
> > > and
> > > > > they
> > > > > > seem to be correct (I can provide screenshots if necessary),
> > > > > > however, I am still unable to communicate with these servers
> > > > > using
> > > > > > IPMI. (eg. ipmipower, ipmiping, etc. all do not work.) and I
> > > am
> > > > > not
> > > > > > sure what else I can do. If anybody could help me or point me
> > > in
> > > > > the
> > > > > > right direction, I would very much appreciate it!
> > > > >
> > > > > It's hard to know for sure what the problem could or could not
> > > be.
> > > > > But
> > > > > if ipmiping isn't even working, that means something very basic
> > > is
> > > > > not
> > > > > configured correctly.  Here's some ideas of what to look at:
> > > > >
> > > > > A) ensure that basic networking is setup correctly on both
> > > client
> > > > > and
> > > > > server.  i.e. valid IP, subnet, MAC, etc.  Most use a different
> > > IP
> > > > > address from their "main" IP address for the node.  Ensure that
> > > > > this IP
> > > > > address can be resolved on the client (i.e. IP converted into
> > > MAC
> > > > > address).  You can check the ARP cache to see if it's been done
> > > > > correctly.
> > > > >
> > > > > B) ensure IPMI is enabled (most notably LAN_Channel
> > > configuration).
> > > > >
> > > > > C) ensure you're using the right networking port.  Many
> > > > > motherboards
> > > > > only support IPMI on one of their networking ports if there is
> > > >
> > > > > 1.  Or
> > > > > sometimes BIOS configs need to be done to determine which port
> > > > > allows
> > > > > IPMI.
> > > > >
> > > > > I've personally worked with Poweredges before, and suspect C
> > > above
> > > > > is
> > > > > your problem.  You may want to look into the ipmi-oem command
> > > > > manpage
> > > > > and read through some of the extensions in there for "Dell" and
> > > any
> > > > > commands that apply to Poweredges.
> > > > >
> > > > > Al
> > > > >
> > > > > On Mon, 2018-01-08 at 17:04 -0800, Jeff Van wrote:
> > > > > > Hello! I'm not sure if this is the right place to ask this
> > > > > question,
> > > > > > but if
> > > > > > not, please let me know where I can direct my question.
> > > > > >
> > > > > > I am currently using MAAS Version 1.9.5+bzr4599-0ubuntu1 with
> > > > > Ubuntu
> > > > > > 14.04.5, with three nodes (all Dell PowerEdge R710). I am
> > > > > currently
> > > > > > trying
> > > > > > to use IPMI version 2.0.
> > > > > >
> > > > > > However, I am unable to power-up these nodes with IPMI, due
> > > to an
> > > > > > error:
> > > > > > Failed to query node's BMC - Power state could not be
> > > queried:
> > > > > > 10.1.10.134:
> > > > > > connection timeout
> > > > > >
> > > > > > In /var/log/maas/maas.log, an identical error message
> > > appears.
> > > > > >
> > > > > > I have double checked the BMC configuration for these nodes,
> > > and
> > > > > they
> > > > > > seem
> > > > > > to be correct (I can provide screenshots if necessary),
> > > however,
> > > > > I am
> > > > > > still
> > > > > > unable to communicate with these servers using IPMI. (eg.
> > > > > ipmipower,
> > > > > > ipmiping, etc. all do not work.) and I am not sure what else
> > > I
> > > > > can
> > > > > > do. If
> > > > > > anybody could help me or point me in the right direction, I
> > > would
> > > > > > very much
> > > > > > appreciate it!
> > > > > >
> > > > > > If there's any more useful information that I can provide,
> > > please
> > > > > let
> > > > > > me
> > > > > > know.
> > > > > > _______________________________________________
> > > > > > Freeipmi-users mailing list
> > > > > > address@hidden
> > > > > > https://lists.gnu.org/mailman/listinfo/freeipmi-users
> > > > > --
> > > > > 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
>
>