According to what i found, it seems that ipmilan stonith module
is not reliable to say the least.

http://www.tummy.com/journals/entries/jafo_20070809_013453 

It was removed from debian heartbeat package.

http://lists.community.tummy.com/pipermail/linux-ha-dev/2007-August/014864.html


You could try to make it work using information from first url
or write some script (which uses impitool) and try to use it
as external stonith module.

Regards,
bartek

>>> On 2007-09-24 at 19:29, in message <[EMAIL PROTECTED]>,
Dejan
Muhamedagic <[EMAIL PROTECTED]> wrote:
> Hi,
> 
> On Mon, Sep 24, 2007 at 06:14:03PM +0200, Frederik Schueler wrote:
>> Hello,
>> 
>> I am trying to figure out why ipmilan STONITH loops on the
following
>> error:
>> 
>> heartbeat[20237]: 2007/09/19_23:11:40 info: Resetting node castor
with [IPMI 
> Over LAN]
>> heartbeat[20237]: 2007/09/19_23:11:40 ERROR: glib: Unable to setup 
> connection: 16
>> heartbeat[20237]: 2007/09/19_23:11:50 ERROR: glib: IPMI operation
timed 
> out... :(
>> heartbeat[20237]: 2007/09/19_23:11:50 info: glib: Host castor
ipmilan-reset 
> error. Error = 6.
>> 
>> the config looks as follow:
>> 
>> in /etc/ha.d/ha.cf:
>> 
>> stonith ipmilan /etc/ha.d/conf/stonith.ipmilan
>> 
>> 
>> and in /etc/ha.d/conf/stonith.ipmilan:
>> 
>> castor 192.168.7.3 623 md5 admin USER PASS
>> pollux 192.168.7.4 623 md5 admin USER PASS
>> 
>> querying the BMCs using ipmitool works fine:
>> 
>> pollux:~# ipmitool -I lan -H 192.168.7.3 -A MD5 -U USER -P PASS
chassis power 
> status
>> Chassis Power is on
>> pollux:~#
>> 
>> doing the STONITH by hand breaks does not:
>> 
>> pollux:/tmp# stonith -t ipmilan -F /etc/ha.d/conf/stonith.ipmilan
-T
>> reset castor
>> 18:08:11.372069 IP 192.168.7.2.32774 > 192.168.7.3.623: UDP, length
23
>> 18:08:11.372073 IP 192.168.7.2.32774 > 192.168.7.3.623: UDP, length
23
>> 18:08:11.373061 IP 192.168.7.3.623 > 192.168.7.2.32774: UDP, length
22
>> 18:08:11.373061 IP 192.168.7.3.623 > 192.168.7.2.32774: UDP, length
22
>> 
>> ** (process:27593): CRITICAL **: Unable to setup connection: 16
>> 
>> 
>> ** (process:27593): CRITICAL **: IPMI operation timed out... :(
>> 
>> ** INFO: Host castor ipmilan-reset error. Error = 6.
>> 
>> 
>> I captured the chat between stonith and the BMC with tcpdump and 
>> looked into it using wireshark: stonith seems to not use authtype
MD5 but
>>  PLAIN, and gets an AUTH_FAILED from the IPMI card.
>> 
>> Setting both the stonith config and the BMC to authtype NONE does
not do
>> the trick, neither PLAIN/PASSWORD - it keeps throwing the errors
above.
>> 
>> Heartbeat used is 2.1.2, built against openipmi 2.0.7, architecture
is
>> amd64 with 64bit userland.
> 
> Looks like a bug to me. Or a compatibility problem. Whichever,
> can you please file a bug report.
> 
> If this is the IBM BMC, you could also try the external/ibmrsa
> stonith plugin.
> 
> Thanks,
> 
> Dejan
> 
>> 
>> Best regards
>> Frederik Schüler
>> 
>> -- 
>> ENOSIG
> 
> 
> 
>> _______________________________________________
>> Linux-HA mailing list
>> [email protected] 
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha 
>> See also: http://linux-ha.org/ReportingProblems 
> _______________________________________________
> Linux-HA mailing list
> [email protected] 
> http://lists.linux-ha.org/mailman/listinfo/linux-ha 
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to