On Sun, Mar 7, 2010 at 6:14 PM, Marc-Christian Petersen <[email protected]> wrote:
> Hello all,
>
> first: I'm completely new to pacemaker.
>
> I want to setup a scenario like this:
>
>
> Machine A)
>        - has IP address 10.0.0.251
>
>        - DRBD resources
>                drbd1 (for apache2)
>                drbd2 (for mysql)
>                drbd3 (for cacti)
>
>        - LVM ontop of every drbd resource (vg11, vg12, vg13)
>                vg11 (for apache)
>                vg12 (for mysql)
>                vg13 (for cacti)
>
>        - apache2
>          - virtual IP 10.0.0.250
>          - /cluster/www
>          - iptables FWMARK to route SNMP requests via the node where Cacti is
>            as ONLY the IP of Cacti is allowed to do SNMP (the monitored 
> devices
>            only have 10.0.0.253 as allowed IP)
>
>        - NFS client mount from Machine B for the Cacti RRD files
>
>
> Machine B)
>        - has IP address 10.0.0.252
>
>        - DRBD resources
>                drbd1 (for apache2)
>                drbd2 (for mysql)
>                drbd3 (for cacti)
>
>        - LVM ontop of every drbd resource
>                vg11 (for apache)
>                vg12 (for mysql)
>                vg13 (for cacti)
>
>        - NFS Server
>
>        - mysql
>          - virtual IP 10.0.0.254
>          - /cluster/mysql
>
>        - cacti
>          - virtual IP 10.0.0.253
>          - /cluster/cacti
>          - iptables SNAT so 10.0.0.253 is the outgoing IP to talk SNMP
>
>
>
> my current configuration is:
> ----------------------------
>
> node machineA
> node machineB
>
> primitive drbd1 ocf:linbit:drbd params drbd_resource="drbd1" op monitor 
> interval="15s"
> primitive drbd2 ocf:linbit:drbd params drbd_resource="drbd2" op monitor 
> interval="15s"
> primitive drbd3 ocf:linbit:drbd params drbd_resource="drbd3" op monitor 
> interval="15s"
>
> primitive fs_apache2 ocf:heartbeat:Filesystem params 
> device="/dev/vg11/apache2" directory="/cluster/www"   fstype="ext4" 
> options="noatime,nodiratime,barrier=0"
> primitive fs_cacti   ocf:heartbeat:Filesystem params device="/dev/vg13/cacti" 
>   directory="/cluster/cacti" fstype="ext4" 
> options="noatime,nodiratime,barrier=0"
> primitive fs_mysql   ocf:heartbeat:Filesystem params device="/dev/vg12/mysql" 
>   directory="/cluster/mysql" fstype="ext4" 
> options="noatime,nodiratime,barrier=0"
>
> primitive ip_apache2 ocf:heartbeat:IPaddr2 params ip="10.0.0.250"
> primitive ip_cacti   ocf:heartbeat:IPaddr2 params ip="10.0.0.253"
> primitive ip_mysql   ocf:heartbeat:IPaddr2 params ip="10.0.0.254"
>
> primitive lvm_vg11 ocf:heartbeat:LVM params volgrpname="vg11" exclusive="yes" 
> op monitor interval="10" timeout="30" depth="0"
> primitive lvm_vg12 ocf:heartbeat:LVM params volgrpname="vg12" exclusive="yes" 
> op monitor interval="10" timeout="30" depth="0"
> primitive lvm_vg13 ocf:heartbeat:LVM params volgrpname="vg13" exclusive="yes" 
> op monitor interval="10" timeout="30" depth="0"
>
> primitive svc_apache2 lsb:apache2
> primitive svc_cacti   lsb:cacti
> primitive svc_mysql   lsb:mysql
>
> group apache2 lvm_vg11 ip_apache2 fs_apache2 svc_apache2 meta 
> target-role="Started"
> group cacti   lvm_vg13 ip_cacti   fs_cacti   svc_cacti   meta 
> target-role="Started"
> group mysql   lvm_vg12 ip_mysql   fs_mysql   svc_mysql   meta 
> target-role="Started"
>
> ms ms_drbd1 drbd1 meta master-max="1" master-node-max="1" clone-max="2" 
> clone-node-max="1" notify="true" target-role="Started"
> ms ms_drbd2 drbd2 meta master-max="1" master-node-max="1" clone-max="2" 
> clone-node-max="1" notify="true" target-role="Started"
> ms ms_drbd3 drbd3 meta master-max="1" master-node-max="1" clone-max="2" 
> clone-node-max="1" notify="true" target-role="Started"
>
> location prefer-fs_apache fs_apache2 100: machineA
> location prefer-fs_cacti  fs_cacti   100: machineB
> location prefer-fs_mysql  fs_mysql   100: machineB
>
> colocation apache_on_drbd1 inf: apache2 ms_drbd1:Master
> colocation cacti_on_drbd3  inf: cacti   ms_drbd3:Master
> colocation mysql_on_drbd2  inf: mysql   ms_drbd2:Master
>
> order apache2_after_drbd1 inf: ms_drbd1:promote apache2:start
> order cacti_after_drbd3   inf: ms_drbd3:promote cacti:start
> order mysql_after_drbd2   inf: ms_drbd2:promote mysql:start
>
> property $id="cib-bootstrap-options" \
>        dc-version="1.0.7-6e1815972fc236825bf3658d7f8451d33227d420" \
>        cluster-infrastructure="openais" \
>        expected-quorum-votes="2" \
>        stonith-enabled="false" \
>        no-quorum-policy="ignore" \
>        last-lrm-refresh="1267978201"
>
>
> everything seems to work correctly, but as you may see, the missing part is
> the NFS stuff (the iptables stuff also but that's easy ;)
>
> How do I configure pacemaker to:
>
> - start the NFS server on the node where Cacti is

well you'd need to add an NFS resource and use a colocaiton constraint
to tie it to cacti

> - mount the NFS share from the NFS server after the NFS server is up, but 
> only:
>  - if the current node does not have /cluster/cacti active

anti-colocate (score of -inf) the mount resource with nfs resource and
use and ordering constraint to tell it to start afterwards

>    i.e. when machineB goes dead, umount /cluster/cacti on machineA and move
>    every resoure from machineB to A and when B comes back, switch all 
> resources
>    who belong to B back to B and mount the NFS share /cluster/cacti on A 
> again.
>
>
> Thanks in advance.
>
>
>
> --
> Kind regards
>        Marc-Christian Petersen
> _______________________________________________
> 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

Reply via email to