On Thu, Mar 10, 2011 at 3:54 PM, Artur <linu...@netdirect.fr> wrote: > Hello, > > I'm currently switching to Heartbeat (3.0.3) and Pacemaker (1.0.9.1) on > Debian Squeeze with CRM/CIB setup. > This is the first time i try to configure it so please be kind with a > newbie. :) > > I would like to setup an active/active cluster with 2 nodes, with sticky > resources and a connectivity check. > At this time i have a basic working solution without connectivity check. > The configuration follows at the bottom of the email. > But i am unable to configure the connectivity check with the pingd resource. > > The (basic) configuration explained : > - on node p01 there are 2 sticky resources called WWW1 (apache web > server) and VIP1 (virtual IP) > - on node p17 there is 1 sticky resource called VIP2 (virtual IP) > > In my mind this is how it should work : > - if a node is down its sticky resources go to the other node (this works) > - if a node has no connectivity the resources go to the other node > (unable to make it work) > - if a down node goes up, the sticky resources are migrated back on it > (this works in the current setup with no connectivity check) > > I created a pingd primitive and cloned it as explained in tutorials and > tried some rules but with no success. > > I tried to add the following rules with no success : > > location www1-on-p01-connected WWW1 \ > rule $id="www1-on-p01-connected-rule" pingd: defined pingd \ > rule $id="www1-on-p01-connected-rule-0" -1000: not_defined pingd or > pingd lte 0 \
change -1000 to -INFINITY > rule $id="www1-on-p01-connected-rule-1" 1000: #uname eq p01 > > location vip2-on-p17-connected VIP2 \ > rule $id="vip2-on-p17-connected-rule" pingd: defined pingd \ > rule $id="vip2-on-p17-connected-rule-0" -1000: not_defined pingd or > pingd lte 0 \ here too > rule $id="vip2-on-p17-connected-rule-1" 1000: #uname eq p17 > > This is the current setup without active connectivity check but with > cloned pingd primitive : > > node $id="52dadc12-ada6-46a0-8474-639a62dfa3ad" p17 > node $id="6d96beed-abd9-4ad1-9a92-b6560abc0475" p01 > primitive VIP1 ocf:heartbeat:IPaddr2 \ > params ip="192.168.1.201" cidr_netmask="32" iflabel="vip1" \ > op monitor interval="30s" > primitive VIP2 ocf:heartbeat:IPaddr2 \ > params ip="192.168.1.202" cidr_netmask="32" iflabel="vip2" \ > op monitor interval="30s" > primitive WWW1 ocf:heartbeat:apache \ > params configfile="/etc/apache2/apache2.conf" > primitive pingd ocf:pacemaker:pingd \ > params host_list="192.168.1.254" multiplier="1000" \ > op monitor interval="15s" timeout="5s" > clone pingdclone pingd \ > meta globally-unique="false" > location vip2-on-p17 VIP2 250: p17 > location www1-on-p01 WWW1 250: p01 > colocation www1-with-vip1 inf: WWW1 VIP1 > order www1-after-vip1 inf: VIP1 WWW1 > property $id="cib-bootstrap-options" \ > stonith-enabled="false" \ > no-quorum-policy="ignore" \ > dc-version="1.0.9-74392a28b7f31d7ddc86689598bd23114f58978b" \ > cluster-infrastructure="Heartbeat" > > Any ideas about how to make it work ? > > -- > > Best regards, > > Artur. > > _______________________________________________ > Linux-HA mailing list > Linux-HA@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@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems