Description of problem: fence_ilo on an iLO interface with firmware 1.64 (11/11/2004) returns w/o changing power status. Version-Release number of selected component (if applicable): How reproducible: always Steps to Reproduce: 1.fence_ilo -a <ilo-ip> -l <login> -p <password> 2. 3. Actual results: no power status change Expected results: should powercycle the system Additional info: The iLO interface returns: <INFORM>Scripting utility should be updated to the latest version.</INFORM>
Created attachment 117712 [details] log output from fence_ilo -v
What version of fence_ilo are you using? What architecture is the machine? There were some issues with ilo on acpid enabled machines, ia64 machines and those with later versions of the firmware (1.70 and later). These issues have since been addressed in the RHEL3u6 and RHEL4u2 releases of GFS. See bug #161352
$FENCE_RELEASE_NAME="1.32.1", the system is a dual-xeon em64t, and the firmware is 1.64 (e.g. below 1.70). I'm not authorized to view bug #161352, could you summarize? I also don't have access to prereleases/betas of RHEL4U2.
It's possible that em64t is having some of the same issues that ia64 was having. The basic gist is that the acpid was causing the nodes to interpret the power off command as a gracefull poweroff and was causing poor interaction between the ilo and GFS. The issue was resolved with this check-in: http://sources.redhat.com/cgi-bin/cvsweb.cgi/cluster/fence/agents/ilo/fence_ilo.pl?rev=1.3.8.1&content-type=text/x-cvsweb-markup&cvsroot=cluster Let me know if this version of fence_ilo fixes the problem for you.
The CVS version seems to work fine, thank you! Any chance this will make it soon into rhn updates?
Since the CVS version works for you, I'm going to resolve this a duplicate of bug #161352 The fix should be included in the next update to rhn. It's in GFS-6.0.2.25-0 for RHEL3 and fence-1.32.6-0 for RHEL4. *** This bug has been marked as a duplicate of 161352 ***