Hi, On Thu, Dec 18, 2008 at 12:50:58PM +0100, Jo De Troy wrote: > Hello, > > here are the cib, the config file and the logs in the attached tarball. Let > me know if you'd rather have it cut-and-pasted here.
No, it's fine. You can also use hb_report: it will collects everything one needs for the report. > At a certain moment I was able to start them up, so I have no clue what's > wrong. > Afterwards I manually stonith'ed the passive node, to test the stonith > settings and that worked. > Afterwards the stontith resource on that node again failed to startup with > rc=6. stonithd[5961]: 2008/12/17_10:31:21 ERROR: record_new_srsc:3136: empty host list for stonith RA external/riloe stonithd[5961]: 2008/12/17_11:30:02 ERROR: record_new_srsc:3136: empty host list for stonith RA external/riloe Perhaps you were hit by the recently discovered bug which may get excercised on a fresh booted host (fixed on Dec 8). Though so far nobody reported it. As a workaround, you may stop and start the stonith resources and then they should be ok. BTW, why does your clock goes yo-yo? Also: # use_logd yes If you don't, you should use ha_logd. Thanks, Dejan > Thanks again, > Jo > > > 2008/12/18 Dejan Muhamedagic <[email protected]> > > > Hi, > > > > On Wed, Dec 17, 2008 at 04:08:12PM +0100, Jo De Troy wrote: > > > Hello, > > > > > > thanks for the clarification. I seem to be having another configuration > > > issue, do I need to setup another constraint to allow the stonith > > resource > > > to run on a node that's running any other resources? I've setup a > > resource > > > group (ip,fs, apache) and 2 seperate stonith resources, 1 for each iLO. > > The > > > stonith resource that should be running on the node where the resource > > group > > > is running fails to startup. > > > If the resource group is not running it does run. > > > The error I see in the log files is mgmtd[5964]: 2008/12/17_10:02:34 > > ERROR: > > > unpack_rsc_op: Hard error: stonith_lab034_start_0 failed with rc=6. > > > > A configuration error. > > > > > Any idea what I'm doing wrong here? > > > > Please post the CIB and more logs. > > > > Thanks, > > > > Dejan > > > > > Thanks in advance, > > > Jo > > > > > > > > > 2008/12/17 Gary Stansbury <[email protected]> > > > > > > > The only way i've gotten this to work reliably is to NOT use clones at > > all. > > > > You need one primitive stonith resource per node, configured with the > > > > appropriate parameters for that host (including ilo hostname and > > username > > > > and password) and set a constraint to not allow it to run on the host > > with > > > > its own ilo, since it can't suicide itself with the riloe stonith > > plugin. > > > > > > > > > > > > > > > > Gary W. Stansbury II > > > > Lead LAN Engineer > > > > [email protected] > > > > 757-631-3356 > > > > > > > > > > > > > > > > From: "Jo De Troy" <[email protected]> > > > > > > > > To: [email protected] > > > > > > > > Date: 12/17/2008 08:48 AM > > > > > > > > Subject: [Linux-HA] Re: Problem with riloe stonith plugin on SLES > > 10 > > > > SP2 > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hello, > > > > > > > > I'm also trying to use the riloe stonith plugin on SLES10 SP2. > > > > I have a question wrt the cloning on the stonith resource, I'm pretty > > new > > > > to > > > > stonith so forgive me my ignorance. > > > > The riloe plugin has several parameters, 1 is the ilo_hostname. I would > > > > expect that every physical machine has it's own and unique > > ilo_hostname, so > > > > how can a clone work? If I create a stonith resource with ilo_hostname: > > > > nodeA_ilo and I clone this resource to have it running on both nodeA > > and > > > > nodeB. There's only 1 ilo_hostname parameter so this will fail when > > trying > > > > to stonith nodeB via it's iLO hostname nodeB_ilo. I must misunderstand > > this > > > > whole thing, can someone help me out here? > > > > Or can I duplicate the ilo_hostname parameter and will stonith choose > > the > > > > correct one based on some logic? > > > > I guess the same holds for the ilo_user and ilo_password, or not? Do > > these > > > > parameters need the same values on all cluster nodes to have the > > resource > > > > clone work and actually perform the stonith? > > > > > > > > > > > > Thanks in advance, > > > > Jo > > > > _______________________________________________ > > > > Linux-HA mailing list > > > > [email protected] > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha > > > > See also: http://linux-ha.org/ReportingProblems > > > > > > > > > > > > > > > > > > > > > > > > > > ****************************************************************************** > > > > This email and any files transmitted with it are intended solely for > > > > the use of the individual or agency to whom they are addressed. > > > > If you have received this email in error please notify the Navy > > > > Exchange Service Command e-mail administrator. This footnote > > > > also confirms that this email message has been scanned for the > > > > presence of computer viruses. > > > > > > > > Thank You! > > > > > > > > > > ****************************************************************************** > > > > > > > > _______________________________________________ > > > > 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 > > > _______________________________________________ > 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
