[Linux-HA] Question regarding primitives and groups inside a clone

Raoul Bhatia [IPAX] r.bhatia at ipax.at
Mon Sep 17 08:58:33 MDT 2007


as far as i understand, a clone ressource can be used to tell the crm
to start a resource "clone_node_max" on one host and "clone_max" times
within the whole cluster.

if only one node is started, the clone should therefore run only
"clone_node_max" times within the whole cluster.

i have a ocfs2 clone (one primitive inside the clone):

>        <clone id="clone_ocfs2_www">
>          <instance_attributes>
>            <attributes>
>              <nvpair name="clone_max"      value="2"/>
>              <nvpair name="clone_node_max" value="1"/>
>            </attributes>
>          </instance_attributes>
>          <primitive class="ocf" id="ocfs2_www" provider="heartbeat" type="Filesystem">
>            <operations>
>              <op id="ocfs2_www_mon" interval="120s" name="monitor" timeout="60s"/>
>            </operations>
>            <instance_attributes>
>              <attributes>
>                <nvpair id="ocfs2_www_attr_0" name="device"    value="/dev/drbd0"/>
>                <nvpair id="ocfs2_www_attr_1" name="directory" value="/data/www"/>
>                <nvpair id="ocfs2_www_attr_2" name="fstype"    value="ocfs2"/>
>              </attributes>
>            </instance_attributes>
>          </primitive>
>        </clone>

which is working as expected:

> Clone Set: clone_ocfs2_www
>     ocfs2_www:0 (heartbeat::ocf:Filesystem):    Stopped
>     ocfs2_www:1 (heartbeat::ocf:Filesystem):    Started webcluster01

when using a group instead of the primitive, things do not work as
i think they should. the result is, that the group is started clone_max

> Clone Set: apache2_clone
>     Resource Group: grrr:0
>         apache2:0       (heartbeat::ocf:apache):        Started webcluster01
>         mysql-proxy:0   (lsb:mysql-proxy):      Started webcluster01
>     Resource Group: grrr:1
>         apache2:1       (heartbeat::ocf:apache):        Started webcluster01
>         mysql-proxy:1   (lsb:mysql-proxy):      Started webcluster01

the only way to avoid this is to use a constraint like:

>        <rsc_colocation id="no_dual_clone" to="grrr:0" from="grrr:1" score="-INFINITY" />

my related configuration is:

>        <clone id="apache2_clone">
>          <instance_attributes>
>            <attributes>
>              <nvpair name="clone_max"      value="2"/>
>              <nvpair name="clone_node_max" value="1"/>
>            </attributes>
>          </instance_attributes>
>          <group id="grrr">
>            <primitive id="apache2" class="ocf" type="apache" provider="heartbeat">
>              <operations>
>                <op id="apache2_mon" name="monitor" interval="10s" timeout="20s"/>
>              </operations>
>              <instance_attributes id="34d8027e-02f2-4b02-ad13-4509de2dadc3">
>                <attributes>
>                  <nvpair name="httpd"       value="/usr/sbin/apache2" />
>                  <nvpair name="options"     value="-k start" />
>                  <nvpair name="configfile"  value="/etc/apache2/apache2.conf" />
>                  <nvpair name="target_role" value="started"/>
>                </attributes>
>              </instance_attributes>
>            </primitive>
>            <primitive id="mysql-proxy" class="lsb" provider="heartbeat" type="mysql-proxy">
>              <operations>
>                <op id="mysql-proxy_mon" name="monitor" interval="10s" timeout="20s" />
>              </operations>
>            </primitive>
>          </group>
>        </clone>

is this an intended behavior?

kind regards,
raoul bhatia
ps: please note that i stripped of some id="xxxxxxxxxxxxxxx" to make the
     xml more readable.
DI (FH) Raoul Bhatia M.Sc.          email.          r.bhatia at ipax.at
Technischer Leiter

IPAX - Aloy Bhatia Hava OEG         web.          http://www.ipax.at
Barawitzkagasse 10/2/2/11           email.            office at ipax.at
1190 Wien                           tel.               +43 1 3670030
FN 277995t HG Wien                  fax.            +43 1 3670030 15

More information about the Linux-HA mailing list