Open Stack

Sigh.

And the answer is: user error. Adminstrator != Administrator.

Well this was fun. Sorry for the bother. All is well. :)

Dan

On Thu, Apr 5, 2018 at 8:13 AM, Dan Prince <dprince at redhat.com> wrote:
> On Wed, Apr 4, 2018 at 1:27 PM, Jim Rollenhagen <jim at jimrollenhagen.com> wrote:
>> On Wed, Apr 4, 2018 at 1:18 PM, Jim Rollenhagen <jim at jimrollenhagen.com>
>> wrote:
>>>
>>> On Wed, Apr 4, 2018 at 8:39 AM, Dan Prince <dprince at redhat.com> wrote:
>>>>
>>>> Kind of a support question but figured I'd ask here in case there are
>>>> suggestions for workarounds for specific machines.
>>>>
>>>> Setting up a new rack of mixed machines this week and hit this issue
>>>> with HP machines using the ipmi power driver for Ironic. Curious if
>>>> anyone else has seen this before? The same commands work great with my
>>>> Dell boxes!
>>>>
>>>> -----
>>>>
>>>> [root at localhost ~]# cat x.sh
>>>> set -x
>>>> # this is how Ironic sends its IPMI commands it fails
>>>> echo -n password > /tmp/tmprmdOOv
>>>> ipmitool -I lanplus -H 172.31.0.19 -U Adminstrator -f /tmp/tmprmdOOv
>>>> power status
>>>>
>>>> # this works great
>>>> ipmitool -I lanplus -H 172.31.0.19 -U Administrator -P password power
>>>> status
>>>>
>>>> [root at localhost ~]# bash x.sh
>>>> + echo -n password
>>>> + ipmitool -I lanplus -H 172.31.0.19 -U Adminstrator -f /tmp/tmprmdOOv
>>>> power status
>>>> Error: Unable to establish IPMI v2 / RMCP+ session
>>>> + ipmitool -I lanplus -H 172.31.0.19 -U Administrator -P password power
>>>> status
>>>> Chassis Power is on
>>>
>>>
>>> Very strange. A tcpdump of both would probably be enlightening. :)
>>>
>>> Also curious what version of ipmitool this is, maybe you're hitting an old
>>> bug.
>>
>>
>> https://sourceforge.net/p/ipmitool/bugs/90/ would seem like a prime suspect
>> here.
>
> Thanks for the suggestion Jim! So I tried a few very short passwords
> and no dice so far. Looking into the tcpdump info a bit now.
>
> I'm in a bit of a rush so I may hack in a quick patch Ironic to make
> ipmitool to use the -P option to proceed and loop back to fix this a
> bit later.
>
> Dan
>
>>
>> // jim
>>
>> __________________________________________________________________________
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>