DL380 G6 how to use ilo stonith device

[mho27]

Hello,

trying to use STONITH device "external/riloe".

This is for shutting down a server via a stonith device.

My ilo interface is configured and working. When printing out the help page of the device:

STONITH Device: external/riloe - HP/COMPAQ iLO2 STONITH device

For more information see http://www.hp.com/

List of valid parameter names for external/riloe STONITH device:
        hostlist
        ilo_hostname
        ilo_user
        ilo_password
        ilo_can_reset
        ilo_protocol
        ilo_powerdown_method
        ilo_proxyhost
        ilo_proxyport
For Config info [-p] syntax, give each of the above parameters in order as
the -p value.
Arguments are separated by white space.
Config file [-F] syntax is the same as -p, except # at the start of a line
denotes a comment

But I can't find any explanation for the parameter "hostlist" ?!

When I try:

"stonith -t external/riloe hostlist=visn2 ilo_hostname=10.23.3.103 ilo_user=Administrator ilo_password=somePW ilo_can_reset=1 ilo_protocol=1.2 ilo_powerdown_method=button -S -T off visn2"

resuluts:

info: external/riloe device OK.

But the server isn't shutting down. "visn2" is a VM running on xenserver. Ofcourse my HP Server can't know that so I have no idea what "hostlist" means..

Also I cant find any documentation about "button -S", read it in a book, but no details here, too...

Somebody have ideas to that?

My goal is to force a shutdown of the HP server...

Regards,

Matthias

[waaronb]

What program are you using to send the command? Is "stonith" something supplied by HP, or is it part of some separate package to knock another cluster node off?

The ILO has a command line interface via SSH, and also there's a scripting program where you can use an XML file to send commands such as hitting the power button, but it seems like you're asking for help on some other command and it's syntax.

For instance, you could SSH into the CLI of the ILO and do a simple "power off" command, or "power off hard".

[mho27]

hi,

thank you for your reply.

"stonith" is part of "pacemaker". So you are right, I am asking for the right syntax of this command.

But thanks anyway, now I know there is a CLI of the ILO. Maybe I can use this to get the Node down.

To my mind, the stonith device is used for this, but I have no Idea how to use it exactly. "For more information visit hp.com" didn't help me.

Regards,

Matthias

[Jimmy Vance · Accepted solution]

By definiion iLO is not a STONITH device, it is a management engine. It can be used to remotely power off a server, so it can be used for this function.

I'm not familiar with pacemaker or its utilities. One thing I would suggest is to use external/ipmi instead of external/riloe. The underlying bits in iLO have changed over the years, and this python script may not work properly.

To understand what some of the values are that you're trying to figure out, you might look at the python scripts in /usr/lib64/stonith/plugins/external/riloe and /usr/lib64/stonith/plugins/external/ipmi (locations may differ a bit depending on what distribution your using)

if you decide to try ipmi, iLO3 and iLO4 require the "interface" to be lanplus, and the user priviledge level needs to be set to operator at a minimum.

Using this syntax I was able to shutdown a Gen8 server over iLO4

stonith -t external/ipmi hostname=labunit1 ipaddr=xx.xx.xx.xx userid=admin passwd=pword interface=lanplus -c 4 -T off labunit1

labunit1 doesn't really mean anything, but it wouldn't work without something in those 2 fields.

You need the -c option set to at least 4 or the command will fail

Hope this helps

[mho27]

Hi again,

sorry for late reply.

Your answer helped indeed, thank you.

I was able to Shut down One Server with

ipmitool -H ipipip -U Administrator -P pwpwpw -L Administrator -I lanplus power off

but often I get sth like:

 ERROR: error executing ipmitool: ipmitool: lanplus.c:2153: ipmi_lanplus_send_payload: Assertion `session->v2_data.session_state == LANPLUS_STATE_PRESESSION' failed. Aborted

or

 ERROR: error executing ipmitool: ipmitool: lanplus.c:2167: ipmi_lanplus_send_payload: Assertion `session->v2_data.session_state == LANPLUS_STATE_OPEN_SESSION_RECEIEVED' failed. Aborted

or

Unable to set Chassis Power Control to Down/Offionol: Error in open session response message : insufficient resources for session

Well it doesn't work when somebody is logged in, but even if no one is logged in it only works sometimes, seems the connection is very unstable ?!

Surfing the web gui of ilo is very slow also, maybe this is a problem?

Why you do c 4 ? You have the same problems?

My stonith isn't working anyway...

with:

stonith -t external/ipmi hostname=visn2 ipaddr=ipipip userid=Administrator passwd=pwpwpwpw interface=lanplus -c 4 -T off visn2

external/ipmi[9366]: ERROR: error executing ipmitool: ipmitool: lanplus.c:2153: ipmi_lanplus_send_payload: Assertion `session->v2_data.session_state == LANPLUS_STATE_PRESESSION' failed. Aborted
CRIT: external_reset_req: 'ipmi off' for host visn2 failed with rc 1
 Unable to set Chassis Power Control to Down/Offtool: Error: Unable to establish IPMI v2 / RMCP+ session
CRIT: external_reset_req: 'ipmi off' for host visn2 failed with rc 1
external/ipmi[9691]: ERROR: error executing ipmitool: ipmitool: lanplus.c:2153: ipmi_lanplus_send_payload: Assertion `session->v2_data.session_state == LANPLUS_STATE_PRESESSION' failed. Aborted
CRIT: external_reset_req: 'ipmi off' for host visn2 failed with rc 1
 Unable to set Chassis Power Control to Down/Offtool: Error: Unable to establish IPMI v2 / RMCP+ session
CRIT: external_reset_req: 'ipmi off' for host visn2 failed with rc 1

Even with -c 50 ;-)

Is there anyway to fix this?

**EDIT**

I created another User @Ilo webinterface and now stonith is working! but still need at least 4 tries with (-c) ... thats makes me a little nervous :)

but THANK YOU!!!

[waaronb]

I wonder if sending the command is merely simulating a "momentary press" instead of a "press and hold". Depending on how the OS or server is configured to respond to a momentary press, it may be trying to do a "polite" shutdown and running into the usual issues when users have open documents or whatever.

Just guessing though.

[mho27]

this would be a possible explanation yes.

some more observations of mine:

When I try to ping the ilo interface the connection was very unstable, 2/4 pings were replied.

No idea what the reason was.

After resetting the ethernet port (unplug/plug networkcable) it was stable again.

Now its working "ok" and sometimes it already shuts down after the first try.

so, maybe the user was just a sth like a placebo.. whatever..

greetings