[Linux-HA] HA EVMS start at boot

Andrew Beekhof beekhof at gmail.com
Mon Jan 29 02:33:52 MST 2007


On 1/26/07, John Lange <john.lange at open-it.ca> wrote:
> Here is the resource that I am currently loading into heartbeat. I do
> have a notify section but I don't know if the values are proper.
>
> Mostly I deduced these from reading the source code of EvmsSCC.
>
> --------
>
>  <clone id="evmscloneset">
>    <instance_attributes id="evmscloneset_instance_attrs">

as robert pointed out later, you need to also set: notify="true"
globally_unique="false"

>      <attributes>
>        <nvpair id="evmscloneset_clone_max" name="clone_max" value="4"/>
>        <nvpair id="evmscloneset_clone_node_max" name="clone_node_max" value="1"/>
>        <nvpair name="target_role" id="evmscloneset_target_role" value="stopped"/>
>      </attributes>
>    </instance_attributes>
>    <primitive class="ocf" type="EvmsSCC" provider="heartbeat" id="evms">
>      <instance_attributes id="evms_instance_attrs">
>        <attributes>
>          <nvpair name="target_role" id="evms_target_role" value="stopped"/>
>        </attributes>

it would also be best if the following nvpairs were removed

>        <attributes>
>          <nvpair name="CRM_meta_notify_active_uname" id="evms_meta_notify_active_uname" value="vs1 vs2 vs3 vs4"/>
>          <nvpair name="CRM_meta_notify_stop_uname" id="evms_meta_notify_stop_uname" value="vs4 vs3 vs2 vs1"/>
>          <nvpair name="CRM_meta_notify_start_uname" id="evms_meta_notify_start_uname" value="vs4 vs3 vs2 vs1"/>
>          <nvpair name="CRM_meta_timeout" id="evms_meta_timeout" value="300000"/>
>          <nvpair name="CRM_meta_notify_type" id="evms_meta_notify_type" value="pre"/>
>        </attributes>
>      </instance_attributes>
>      <operations>

notify shouldn't be configured like this (see above), you should
remove this next line.

>        <op id="78df7ddc-5a7c-4262-a203-82b5ad7c995d" name="notify" interval="120s" timeout="240s"/>
>      </operations>
>    </primitive>
>  </clone>
>
>
>
>
> On Fri, 2007-01-26 at 13:57 -0700, Robert Wipfel wrote:
> > >>> On Thu, Jan 25, 2007 at 12:21 PM, in message
> > <1169752904.5049.9.camel at ibmlaptop.darkcore.net>, John Lange
> > <john.lange at open-it.ca> wrote:
> > > On Wed, 2007- 01- 24 at 14:32 - 0700, Robert Wipfel wrote:
> > >> >>> On Wed, Jan 24, 2007 at  2:21 PM, in message
> > >> <904050d50701241321n7d41494ag8786077c67a22f92 at mail.gmail.com>, "Andreas Kurz"
> > >> <andreas.kurz at gmail.com> wrote:
> >
> > [....]
> >
> > >> EvmsSCC is needed for making Shared (public) Cluster Containers (SCC)
> > >> visible, as a dependency for hosting an OCFS2 cluster file systems
> > >> over shared cluster segments, since EVMS' Cluster Segment Manager
> > >> (CSM)  won't allow them to be activated until the node has joined the
> > >> cluster. If you have logical volumes on local (or shared) disks,
> > >> partitioned using GUID or DOS segment managers, then iirc boot.evms
> > >> (which also calls evms_activate) should scan them in during boot (if
> > >> chkconfig'd on). Changju will know for sure...
> > >
> > > Ok, I could not get it to work. First challenge was that 2.0.7 is what
> > > ships with SLES 10 so the resource script didn't exist. I installed
> > > 2.0.8 from source by just running "./ConfigureMe install" over top the
> > > existing install. That seems to have worked as the EvmsSCC resource is
> > > now available in the GUI.
> >
> > > Are there some parameters that need to be added to the resource or what
> > > might the problem be?
> >
> > I'm told you need to set (notify=true). Maybe you could post your CIB
> > and we can take a look.
> >
> > > Does anyone know where a more recent release of the Novell "Conquering
> > > the High Availability Storage Foundation" document (found here:
> > > http://wiki.novell.com/index.php/SUSE_Linux_Enterprise_Server ) might be
> > > obtained? It describes the exact situation we are trying to configure
> > > but unfortunately ends before the parts we are missing.
> >
> > There will be updates coming soon. Sorry there's not more there now...
> >
> > Hth,
> > Robert
> >
> > _______________________________________________
> > Linux-HA mailing list
> > Linux-HA at lists.linux-ha.org
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > See also: http://linux-ha.org/ReportingProblems
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>


More information about the Linux-HA mailing list