cancel
Showing results forΒ 
Search instead forΒ 
Did you mean:Β 

Auto switch of a LVMVolumeGroup resource

lallo
Level 2

Hi all,

I have installed a VCS version 5.0 MP3 on a two RedHat EL 5.3 64bit nodes, this servers share four SAN disks that are connected by two fibre channel on each server, the multiple links with the SAN are managed with linux multipath daemon.
I have configured two types of resources in a one service group: Volume group and File system. The cluster, at the boot time, should import the Volume Group (and activate the logical volume that it contain) and mount the filesystems that is associated with the logical volume. This steps works fine without problems and the Volume group and File system start correctly and works fine, but, after some minutes, these resources do an auto-switch on the second node without reason.
In the engine.log I can see this at the switch time :

The first message speak about a resource "oradbvg" (is one of my shared volume group) that became online on the second node, but I don't start the resource on the second node.


2009/11/13 14:38:49 VCS INFO V-16-1-10299 Resource oradbvg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (Not initiated by VCS)
2009/11/13 14:38:49 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group oracle
2009/11/13 14:38:49 VCS NOTICE V-16-1-10233 Clearing Restart attribute for group oracle on all nodes
2009/11/13 14:38:49 VCS INFO V-16-1-10299 Resource sysvg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (Not initiated by VCS)
2009/11/13 14:38:49 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group oracle
2009/11/13 14:38:49 VCS NOTICE V-16-1-10233 Clearing Restart attribute for group oracle on all nodes
2009/11/13 14:38:49 VCS INFO V-16-1-10299 Resource redovg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (Not initiated by VCS)
2009/11/13 14:38:49 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group oracle
2009/11/13 14:38:49 VCS NOTICE V-16-1-10233 Clearing Restart attribute for group oracle on all nodes
2009/11/13 14:38:49 VCS WARNING V-16-6-15034 (rh5masiemdb2) violation:Offlining group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS WARNING V-16-6-15034 (rh5masiemdb2) violation:Offlining group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10167 Initiating manual offline of group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10300 Initiating Offline of Resource oradbvg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10300 Initiating Offline of Resource redovg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10300 Initiating Offline of Resource sysvg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10167 Initiating manual offline of group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS WARNING V-16-6-15034 (rh5masiemdb2) violation:Offlining group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS INFO V-16-6-15002 (rh5masiemdb2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/violation rh5masiemdb2 oracle   successfully
2009/11/13 14:38:49 VCS INFO V-16-6-15002 (rh5masiemdb2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/violation rh5masiemdb2 oracle   successfully
2009/11/13 14:38:49 VCS NOTICE V-16-1-10167 Initiating manual offline of group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS INFO V-16-6-15002 (rh5masiemdb2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/violation rh5masiemdb2 oracle   successfully
2009/11/13 14:38:49 VCS INFO V-16-1-10299 Resource archivevg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (Not initiated by VCS)
2009/11/13 14:38:49 VCS ERROR V-16-1-10214 Concurrency Violation:CurrentCount increased above 1 for failover group oracle
2009/11/13 14:38:49 VCS NOTICE V-16-1-10233 Clearing Restart attribute for group oracle on all nodes
2009/11/13 14:38:49 VCS WARNING V-16-6-15034 (rh5masiemdb2) violation:Offlining group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10167 Initiating manual offline of group oracle on system rh5masiemdb2
2009/11/13 14:38:49 VCS NOTICE V-16-1-10300 Initiating Offline of Resource archivevg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:49 VCS INFO V-16-6-15002 (rh5masiemdb2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/violation rh5masiemdb2 oracle   successfully
2009/11/13 14:38:51 VCS ERROR V-16-2-13067 (rh5masiemdb1) Agent is calling clean for resource(archivevg) because the resource became OFFLINE unexpectedly, on its own.
2009/11/13 14:38:51 VCS ERROR V-16-2-13067 (rh5masiemdb1) Agent is calling clean for resource(oradbvg) because the resource became OFFLINE unexpectedly, on its own.
2009/11/13 14:38:51 VCS ERROR V-16-2-13067 (rh5masiemdb1) Agent is calling clean for resource(redovg) because the resource became OFFLINE unexpectedly, on its own.
2009/11/13 14:38:51 VCS ERROR V-16-2-13067 (rh5masiemdb1) Agent is calling clean for resource(sysvg) because the resource became OFFLINE unexpectedly, on its own.
2009/11/13 14:38:51 VCS INFO V-16-1-10305 Resource redovg (Owner: unknown, Group: oracle) is offline on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:52 VCS INFO V-16-1-10305 Resource oradbvg (Owner: unknown, Group: oracle) is offline on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:52 VCS INFO V-16-1-10305 Resource sysvg (Owner: unknown, Group: oracle) is offline on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:52 VCS INFO V-16-1-10305 Resource archivevg (Owner: unknown, Group: oracle) is offline on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:52 VCS NOTICE V-16-1-10446 Group oracle is offline on system rh5masiemdb2
2009/11/13 14:38:52 VCS INFO V-16-10031-15005 (rh5masiemdb2) triggers:???:nfs_postoffline:(postoffline) Invoked with arguments rh5masiemdb2, oracle
2009/11/13 14:38:52 VCS INFO V-16-6-15002 (rh5masiemdb2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/nfs_postoffline rh5masiemdb2 oracle   successfully
2009/11/13 14:38:52 VCS INFO V-16-6-15004 (rh5masiemdb2) hatrigger:Failed to send trigger for postoffline; script doesn't exist
2009/11/13 14:38:52 VCS INFO V-16-2-13068 (rh5masiemdb1) Resource(redovg) - clean completed successfully.
2009/11/13 14:38:52 VCS INFO V-16-2-13068 (rh5masiemdb1) Resource(oradbvg) - clean completed successfully.
2009/11/13 14:38:52 VCS INFO V-16-2-13068 (rh5masiemdb1) Resource(sysvg) - clean completed successfully.
2009/11/13 14:38:52 VCS INFO V-16-2-13068 (rh5masiemdb1) Resource(archivevg) - clean completed successfully.
2009/11/13 14:38:53 VCS INFO V-16-1-10307 Resource oradbvg (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (Not initiated by VCS)
2009/11/13 14:38:53 VCS NOTICE V-16-1-10300 Initiating Offline of Resource vip (Owner: unknown, Group: oracle) on System rh5masiemdb1
2009/11/13 14:38:53 VCS INFO V-16-1-10307 Resource sysvg (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (Not initiated by VCS)
2009/11/13 14:38:53 VCS INFO V-16-1-10307 Resource redovg (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (Not initiated by VCS)
2009/11/13 14:38:53 VCS INFO V-16-1-10307 Resource archivevg (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (Not initiated by VCS)
2009/11/13 14:38:53 VCS INFO V-16-6-15004 (rh5masiemdb1) hatrigger:Failed to send trigger for resfault; script doesn't exist
2009/11/13 14:38:53 VCS INFO V-16-6-15004 (rh5masiemdb1) hatrigger:Failed to send trigger for resfault; script doesn't exist
2009/11/13 14:38:53 VCS INFO V-16-6-15004 (rh5masiemdb1) hatrigger:Failed to send trigger for resfault; script doesn't exist
2009/11/13 14:38:53 VCS INFO V-16-6-15004 (rh5masiemdb1) hatrigger:Failed to send trigger for resfault; script doesn't exist
2009/11/13 14:38:54 VCS INFO V-16-1-10305 Resource vip (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:38:54 VCS NOTICE V-16-1-10300 Initiating Offline of Resource archivefs (Owner: unknown, Group: oracle) on System rh5masiemdb1
2009/11/13 14:38:54 VCS NOTICE V-16-1-10300 Initiating Offline of Resource databasefs (Owner: unknown, Group: oracle) on System rh5masiemdb1
2009/11/13 14:38:54 VCS NOTICE V-16-1-10300 Initiating Offline of Resource redofs (Owner: unknown, Group: oracle) on System rh5masiemdb1
2009/11/13 14:38:54 VCS NOTICE V-16-1-10300 Initiating Offline of Resource sysfs (Owner: unknown, Group: oracle) on System rh5masiemdb1
2009/11/13 14:38:56 VCS INFO V-16-1-10305 Resource archivefs (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:38:56 VCS INFO V-16-1-10305 Resource databasefs (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:38:56 VCS INFO V-16-1-10305 Resource redofs (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:38:56 VCS INFO V-16-1-10305 Resource sysfs (Owner: unknown, Group: oracle) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:38:56 VCS ERROR V-16-1-10205 Group oracle is faulted on system rh5masiemdb1
2009/11/13 14:38:56 VCS NOTICE V-16-1-10446 Group oracle is offline on system rh5masiemdb1
2009/11/13 14:38:56 VCS INFO V-16-1-10493 Evaluating rh5masiemdb1 as potential target node for group oracle
2009/11/13 14:38:56 VCS INFO V-16-1-50010 Group oracle is online or faulted on system rh5masiemdb1
2009/11/13 14:38:56 VCS INFO V-16-1-10493 Evaluating rh5masiemdb2 as potential target node for group oracle
2009/11/13 14:38:56 VCS NOTICE V-16-1-10301 Initiating Online of Resource archivevg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:56 VCS NOTICE V-16-1-10301 Initiating Online of Resource oradbvg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:56 VCS NOTICE V-16-1-10301 Initiating Online of Resource redovg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:56 VCS NOTICE V-16-1-10301 Initiating Online of Resource sysvg (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:56 VCS INFO V-16-10031-15005 (rh5masiemdb1) triggers:???:nfs_postoffline:(postoffline) Invoked with arguments rh5masiemdb1, oracle
2009/11/13 14:38:56 VCS INFO V-16-6-15002 (rh5masiemdb1) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/nfs_postoffline rh5masiemdb1 oracle   successfully
2009/11/13 14:38:56 VCS INFO V-16-6-15004 (rh5masiemdb1) hatrigger:Failed to send trigger for postoffline; script doesn't exist
2009/11/13 14:38:58 VCS INFO V-16-1-10298 Resource sysvg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:58 VCS NOTICE V-16-1-10301 Initiating Online of Resource sysfs (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:58 VCS INFO V-16-1-10298 Resource archivevg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:58 VCS NOTICE V-16-1-10301 Initiating Online of Resource archivefs (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:58 VCS INFO V-16-1-10298 Resource oradbvg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:58 VCS NOTICE V-16-1-10301 Initiating Online of Resource databasefs (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:58 VCS INFO V-16-1-10298 Resource redovg (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:58 VCS NOTICE V-16-1-10301 Initiating Online of Resource redofs (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:38:59 VCS INFO V-16-1-10298 Resource sysfs (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:59 VCS INFO V-16-1-10298 Resource databasefs (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:59 VCS INFO V-16-1-10298 Resource redofs (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:59 VCS INFO V-16-1-10298 Resource archivefs (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:38:59 VCS NOTICE V-16-1-10301 Initiating Online of Resource vip (Owner: unknown, Group: oracle) on System rh5masiemdb2
2009/11/13 14:39:07 VCS INFO V-16-1-10298 Resource vip (Owner: unknown, Group: oracle) is online on rh5masiemdb2 (VCS initiated)
2009/11/13 14:39:07 VCS NOTICE V-16-1-10447 Group oracle is online on system rh5masiemdb2
2009/11/13 14:39:07 VCS NOTICE V-16-1-10448 Group oracle failed over to system rh5masiemdb2
2009/11/13 14:39:07 VCS INFO V-16-6-15004 (rh5masiemdb2) hatrigger:Failed to send trigger for postonline; script doesn't exist
2009/11/13 14:39:21 VCS INFO V-16-1-50135 User root fired command: MSG_CLUSTER_STOP_SYS from localhost
2009/11/13 14:39:21 VCS NOTICE V-16-1-10322 System rh5masiemdb1 (Node '0') changed state from RUNNING to LEAVING
2009/11/13 14:39:21 VCS NOTICE V-16-1-10300 Initiating Offline of Resource ntfr (Owner: unknown, Group: ClusterService) on System rh5masiemdb1
2009/11/13 14:39:22 VCS INFO V-16-1-10305 Resource ntfr (Owner: unknown, Group: ClusterService) is offline on rh5masiemdb1 (VCS initiated)
2009/11/13 14:39:22 VCS NOTICE V-16-1-10446 Group ClusterService is offline on system rh5masiemdb1
2009/11/13 14:39:22 VCS NOTICE V-16-1-10010 Stopping all agents
2009/11/13 14:39:22 VCS INFO V-16-1-10493 Evaluating rh5masiemdb1 as potential target node for group ClusterService
2009/11/13 14:39:22 VCS INFO V-16-1-10494 System rh5masiemdb1 not in RUNNING state
2009/11/13 14:39:22 VCS INFO V-16-1-10493 Evaluating rh5masiemdb2 as potential target node for group ClusterService
2009/11/13 14:39:22 VCS NOTICE V-16-1-10490 Resetting Evacuating (lastval=0) for group ClusterService for node rh5masiemdb1
2009/11/13 14:39:22 VCS NOTICE V-16-1-10301 Initiating Online of Resource ntfr (Owner: unknown, Group: ClusterService) on System rh5masiemdb2
2009/11/13 14:39:22 VCS NOTICE V-16-1-10322 System rh5masiemdb1 (Node '0') changed state from LEAVING to EXITING
2009/11/13 14:39:22 VCS NOTICE V-16-1-10322 System rh5masiemdb1 (Node '0') changed state from EXITING to EXITED


Thanks in advantage for the support.
Bye

1 ACCEPTED SOLUTION

Accepted Solutions

Ratheeshk
Level 2
 Got the issue.

DiskReservation resource has to be used with LVMVolumeGroup.
But in this case multiple path to the disk isnot supported.

So if you want to use DMMP+LVM2+VCS , Use DiskReservation with Single path 

If you need Multipath use DMP + VxVM + VCS  :)

View solution in original post

9 REPLIES 9

lallo
Level 2
I have forget to write that if I manually import  the volume groups and mount the file systems all works fine, the problem coming out only when I use the VCS for manage the file systems/volume groups on the nodes.

Gaurav_S
Moderator
Moderator
   VIP    Certified
Hello,

Looking in errors it indicates a concurrency violation... so for sure once the VG is online on any node.... somehow VCS is understanding that VG has got imported on second node as well...

As you are saying that when you import manually all works fine... we need to review the configuration once...

can you please paste the main.cf here ?

Also, one more thought, it could be an issue with your LLT, quite possible that other node is not knowing that first node has group imported & hence trying to import the group.... some temporary failures with your LLT links might cause this issue... please paste your LLT config also here:

# lltstat -vvn | head
# lltstat
# cat /etc/llttab
# cat /etc/llthosts

Also, by any chance you know if LLT is connected on VLANS or cross cables ? can you verify the links ?

Gaurav

lallo
Level 2

Hi Gaurav, thanks a lot for you support. I have put following the informations that you have required.
For the hearthbeats links the situation is this : Now, we have two heartbeats connected by two cross cables with the second node, for the future the heartbeats cards will be connected through a switch (maybe two for redundancy reasons) with a VLAN dedicated for the haertbeat traffic.
The heartbeats are, on each nodes, two internal Nics of type "Broadcom NetXtreme II BCM5709 1000Base-T (C0)" alias "eth0" and "eth1"

[root@rh5masiemdb1 ~]# more /etc/VRTSvcs/conf/config/main.cf
include "types.cf"

cluster rh5masiemdb (
        UserNames = { admin = hKLdKFkHLgLLjTLfKI }
        Administrators = { admin }
        )

system rh5masiemdb1 (
        )

system rh5masiemdb2 (
        )

group ClusterService (
        SystemList = { rh5masiemdb1 = 0, rh5masiemdb2 = 1 }
        AutoStartList = { rh5masiemdb1, rh5masiemdb2 }
        OnlineRetryLimit = 3
        OnlineRetryInterval = 120
        )

        NIC csgnic (
                Device = bond0
                )

        NotifierMngr ntfr (
                SnmpConsoles = { "10.155.147.18" = Information }
                SmtpServer = "10.155.147.50"
                SmtpRecipients = { "arc_dbvcs_ma_soc@terna.it" = Information }
                )

        ntfr requires csgnic


        // resource dependency tree
        //
        //      group ClusterService
        //      {
        //      NotifierMngr ntfr
        //          {
        //          NIC csgnic
        //          }
        //      }


group oracle (
        SystemList = { rh5masiemdb1 = 0, rh5masiemdb2 = 1 }
        )

        LVMVolumeGroup redovg (
                VolumeGroup = oraredovg
                StartVolumes = 1
                )

        Mount redofs (
                MountPoint = "/oracle/redo"
                BlockDevice = "/dev/mapper/oraredovg-oraredolv"
                FSType = ext3
                FsckOpt = "-n"
                CkptUmount = 0
                )

        redofs requires redovg


        // resource dependency tree
        //
        //      group oracle
        //      {
        //      Mount redofs
        //          {
        //          LVMVolumeGroup redovg
        //          }
        //      }

[root@rh5masiemdb1 ~]# lltstat -vvn | head
LLT node information:
    Node                 State    Link  Status  Address
   * 0 rh5masiemdb1      OPEN
                                  eth0   UP      00:21:5E:36:44:24
                                  eth1   UP      00:21:5E:36:44:26
     1 rh5masiemdb2      OPEN
                                  eth0   UP      00:21:5E:36:33:D8
                                  eth1   UP      00:21:5E:36:33:DA
     2                   CONNWAIT
                                  eth0   DOWN

[root@rh5masiemdb1 ~]# lltstat
LLT statistics:
    61637      Snd data packets
    0          Snd retransmit data
    5983644    Snd connect packets
    1601709    Snd independent ACKs
    19992      Snd piggyback ACKs
    0          Snd independent NACKs
    0          Snd piggyback NACKs
    91117      Snd loopback packets
    31607      Rcv data packets
    0          Rcv out of window
    0          Rcv duplicates
    0          Rcv datagrams dropped
    0          Rcv multiblock data
    0          Rcv misaligned data
    0          Snd chained header
LLT errors:
    2          Rcv not connected
    0          Rcv unconfigured
    0          Rcv bad dest address
    0          Rcv bad source address
    0          Rcv bad generation
    0          Rcv no buffer
    0          Rcv malformed packet
    0          Rcv bad SAP
    0          Rcv bad STREAM primitive
    0          Rcv bad DLPI primitive
    0          Rcv DLPI error
    90         Snd not connected
    0          Snd no buffer
    0          Snd stream flow drops
    0          Snd no links up
    0          Rcv bad checksum
    0          Rcv bad udp/ether source address
    0          Snd udp queue full drops

[root@rh5masiemdb1 ~]# cat /etc/llttab
set-node rh5masiemdb1
set-cluster 0
link eth0 eth-00:21:5e:36:44:24 - ether - -
link eth1 eth-00:21:5e:36:44:26 - ether - -

[root@rh5masiemdb1 ~]# cat /etc/llthosts
0 rh5masiemdb1
1 rh5masiemdb2

Gaurav_S
Moderator
Moderator
   VIP    Certified
Hello ,

I hope the configuration for other volume groups is same.. nothing unusual looks here.... Autostartlist can be populated however I can't think of Autostartlist making any difference here.

 LLT also looks ok, ( I hope llttab & llthosts files are identical on both the nodes -except hostnames)

Have you configured any triggers ? for sure something is triggering the online of the VG resources, as per VCS it is happening outside of VCS but still worth to check any triggers.

Also anything external do you think of  (like some database script or some other script ) ? because what I see from above error messages is 3 of the VGs were triggered online at same time...

Also it would be worth to check the system logs to figure out if a manual vgimport command was triggerd....(this should be logged even when script triggers)...

Gaurav

Ratheeshk
Level 2
 I think the issue is with vgdisplay command output.
VCS uses vdsiaplay command for monitoring the status of LVMVolumeGroup. Since the output of vgdisplay command on both the cluster nodes are same in case of RHEL5.3, VCS thinks the VG is online on the other node also.

If you shutdown one node of your 2 node cluster, VG will not fail.
could you please confirm ?.

I am also looking for a solution for this issue.

Ratheeshk
Level 2
 Got the issue.

DiskReservation resource has to be used with LVMVolumeGroup.
But in this case multiple path to the disk isnot supported.

So if you want to use DMMP+LVM2+VCS , Use DiskReservation with Single path 

If you need Multipath use DMP + VxVM + VCS  :)

Taoufik_Grira
Level 2

Hi lallo,

I'm having the same issue with my VCS: VCS 5.0 MP4 on two Red Hat Linux 5.4 x86_64 servers connected to an EVA4400 disk array via FC. I'm using embedded Linux multipath daemon.
Could you please tell us what was the fix, or whether you have got an official response from Symantec.
Ratheeshk: is what you have written above the official uniq procedure provided by Symantec or it's a solution among others.

Thanks for your help,

Taoufik_Grira
Level 2
Hi,

according to symanec official docs Linux LVM is not supported with VCS disk groups. Veritas Volume Manager is necesary in this case.

Regards
Taoufik Grira
System & storage engineer

g_lee
Level 6
Taoufik,

To clarify:

You can't use the VCS DiskGroup bundled agent to control a Linux LVM group; however there is an LVMVolumeGroup agent if you want to use LVM

As Ratheeshk pointed out, you need to use DiskReservation agent with LVMVolumeGroup resources to ensure the disks are only accessible to one system; however, per Bundled Agents reference guide p36 ( http://sfdoccentral.symantec.com/sf/5.0MP3/linux/pdf/vcs_bundled_agents.pdf :(
"The DiskReservation agent cannot be used to reserve disks that have multiple paths."

So if multipathing is required, use DMP+VxVM rather than linux native multipathing+LVM

regards,
Grace