[chef] Re: Re: IPMI / Ohai

[John Dewey < >]

Not to mention it appears dmidecode is run via Ohai regardless.

https://github.com/opscode/ohai/blob/d74376a05a3a4f99e0a63d6796fae2d4726e9ba4/lib/ohai/plugins/dmi_common.rb

On Monday, September 23, 2013 at 9:16 AM, Tim Smith wrote:

That seems really odd that it's taking 30 seconds for you.  On my Supermicro systems that command takes less than a second.  Might be worth seeing if there are any IPMI firmware updates to apply from SM.

Tim Smith  - Systems Engineer 
m: +1 707.738.8132

On Sep 23, 2013, at 7:36 AM, Graham Christensen < "> > wrote:

Hello everyone,

I maintain a deployment of SuperMicro servers, and Dell servers, as well as a compliment of cloud machines. What I want to accomplish is as follows:

An ohai plugin to determine if IPMI is available. I can do this with dmidecode, which takes about 30 seconds:

sudo dmidecode --type 38

# dmidecode 2.11

SMBIOS 2.6 present.

Handle 0x2600, DMI type 38, 18 bytes

IPMI Device Information

Interface Type: KCS (Keyboard Control Style)

Specification Version: 2.0

I2C Slave Address: 0x10

NV Storage Device: Not Present

Base Address: 0x0000000000000CA8 (I/O)

Register Spacing: 32-bit Boundaries

And the failure mode, takes less than one second:

sudo dmidecode --type 38

# dmidecode 2.11

# No SMBIOS nor DMI entry point found, sorry.

Unfortunately it takes too long to run on a regular basis. (If anyone has sage wisdom on detecting IPMI in a quicker fashion, it would be appreciated.) Would it be a faux pas to touch a hint file after the first run which indicates it is present? The logic might look like this:

if hint?('has_ipmi')

  ipmi true

else if `dmidecode --type 38`.include? 'IPMI Device Information'

  ipmi true

  save_hint('has_ipmia')

else

  ipmi false

end

Any feedback would be much appreciated!

Thank you very much, 

Graham Christensen