Hi

I have a 4 node cluster and am trying to get SCSI Persistent Reservations working correctly.
I was hoping to get a get a more detailed how to on getting it working.

What is throwing me off at the moment is the above line in the /etc/init.d/scsi_reserve script.

if [ -f /etc/sysconfig/scsi_reserve ] ; then
. /etc/sysconfig/scsi_reserve
fi

I am not sure how to generate this file or what it should look like.

I have tried the tool fence_scsi -n 1,2,3 and 4 etc but get the following error:

Unable to determine nodeid for 1 etc.

According to documentation on URL http://kbase.redhat.com/faq/docs/DOC-17809 LVM2 cluster volumes are required for persistent reservation fencing.

I configured /dev/sdc1 which is a 1G SAN disk with HP qlogic drivers and added to the rivenvg (LVM Volume group).

My question is how to get /dev/sdc1 configured for persistent reservation fencing and if anyone had any advise on how to get this working correctly.

Packages
cman-2.0.98-1.el5_3.4
sg3_utils-1.25-1.el5
sg3_utils-libs-1.25-1.el5
lvm2-2.02.40-6.el5
lvm2-cluster-2.02.26-1.el5

Cluster Configuration

<?xml version="1.0"?>
<cluster alias="rivencluster" config_version="29" name="rivencluster">
<fence_daemon clean_start="0" post_fail_delay="0" post_join_delay="3"/>
<clusternodes>
<clusternode name="node3prv.rmb.co.za" nodeid="1" votes="1">
<fence>
<method name="1">
<device name="SCSIRES" node="node3prv.rmb.co.za"/>
</method>
</fence>
</clusternode>
<clusternode name="node2prv.rmb.co.za" nodeid="2" votes="1">
<fence>
<method name="1">
<device name="SCSIRES" node="node2prv.rmb.co.za"/>
</method>
</fence>
</clusternode>
<clusternode name="node1prv.rmb.co.za" nodeid="3" votes="1">
<fence>
<method name="1">
<device name="SCSIRES" node="mode1prv.rmb.co.za"/>
</method>
</fence>
</clusternode>
<clusternode name="node4prv.rmb.co.za" nodeid="4" votes="1">
<fence>
<method name="1">
<device name="SCSIRES" node="node4prv.rmb.co.za"/>
</method>
</fence>
</clusternode>
</clusternodes>
<cman/>
<fencedevices>
<fencedevice agent="fence_scsi" name="SCSIRES"/>
</fencedevices>
<rm>
<failoverdomains>
<failoverdomain name="RivenDomain" ordered="1" restricted="1">
<failoverdomainnode name="node1prv.rmb.co.za" priority="1"/>
<failoverdomainnode name="node2prv.rmb.co.za" priority="1"/>
</failoverdomain>
</failoverdomains>
<resources>
<clusterfs device="/dev/mapper/rivenvg-rivenLV1" force_unmount="1" fsid="51133" fstype="gfs" mountpoint="/data" name="RivenGFS" options=""/>
<script file="/etc/init.d/siena" name="Software Environment Script"/>
<ip address="IP" monitor_link="1"/>
</resources>
<service autostart="1" domain="RivenDomain" name="Siena-Environment" recovery="relocate">
<script ref="Software Environment Script"/>
</service>
<service autostart="1" domain="RivenDomain" name="IP" recovery="relocate">
<ip ref="IPADDRESS"/>
</service>
</rm>
<quorumd device="/dev/sdb1" interval="3" min_score="1" tko="23" votes="3"/>
<totem consensus="4800" join="60" token="10000" token_retransmits_before_loss_const="20"/>
<gfs_controld plock_rate_limit="0"/>
<dlm plock_ownership="1"/>
</cluster>