sundar mahadevan
2009-May-23 21:05 UTC
[Ocfs2-users] cluster manager does not start/run ERROR: OemInit2: Attempting to open the CMDiskFile for a multi-node RAC
Hi Members,
I 'm trying to install 9i RAC (initial installation) on a 2 node
cluster (ocfs2 on opensuse 11.1) on my home test boxes both 900MHz
with 512 MB RAM. I can either go for 9i or 10gR1 because of the RAM
limitation. I'm using OCFS2 and the disk is visible/accessible to both
the boxes. The permissions are set right. I'm executing the
ocmstart.sh script as root and i have root added to dba group which
has rwx on all required directories starting from /u01. I have
cmquorumfile and sharedsrvctlconfig file on /u01/oradata/orcl/orcl.
There are no error messages on /var/log/messages. I know that Oracle
does not support opensuse and supports only Enterprise Linux, Redhat,
SLES and Asianux. Newbie. Help please!!!!!!!
sunny1:~ # cat /u01/oradata/orcl/product/9.2/db_1/oracm/log/cm.log
oracm, version[ 9.2.0.2.0.47 ] started {Sat May 23 15:53:16 2009 }
KernelModuleName is hangcheck-timer {Sat May 23 15:53:16 2009 }
OemNodeConfig(): Network Address of node0: 10.1.1.1 (port 9998)
{Sat May 23 15:53:16 2009 }
OemNodeConfig(): Network Address of node1: 10.1.1.2 (port 9998)
{Sat May 23 15:53:16 2009 }>ERROR: OemInit2: Attempting to open the CMDiskFile for a multi-node RAC
on a non-NFS, non-OCFS, or non-raw device cluster, tid = main:-1210403136 file =
oem.c, line = 482 {Sat May 23 15:53:16 2009 }
>ERROR: OemInit2: If the CMDiskFile is supposed to be an NFS or OCFS file,
please make sure that the relevant shared file system is mounted properly, tid =
main:-1210403136 file = oem.c, line = 483 {Sat May 23 15:53:16 2009 }
>ERROR: OemInit2: If the CMDiskFile is supposed to be a raw device, please
make sure that it has been created properly, tid = main:-1210403136 file =
oem.c, line = 484 {Sat May 23 15:53:16 2009 }
--- DUMP GROUP STATE DB ---
--- END OF GROUP STATE DUMP ---
--- Begin Dump --->TRACE: LogListener: Spawned with tid 0xb70c6b90., tid = -1223922800 file
= logging.c, line = 116 {Sat May 23 15:53:16 2009 }
oracm, version[ 9.2.0.2.0.47 ] started {Sat May 23 15:53:16 2009
}>TRACE: Can't read registry value for HeartBeat, tid =
main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 }
>TRACE: Can't read registry value for PollInterval, tid =
main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 }
>TRACE: Can't read registry value for WatchdogTimerMargin, tid =
main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 }
>TRACE: Can't read registry value for WatchdogSafetyMargin, tid =
main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 }
KernelModuleName is hangcheck-timer {Sat May 23 15:53:16 2009
}>TRACE: Can't read registry value for ClientTimeout, tid =
main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 }
>TRACE: InitNMInfo: setting clientTimeout to 620s based on MissCount 620
and PollInterval 1000ms, tid = main:-1210403136 file = nmconfig.c, line = 137
{Sat May 23 15:53:16 2009 }
>TRACE: InitClusterDb(): getservbyname on CMSrvr failed - 0 : assigning
9998, tid = main:-1210403136 file = nmconfig.c, line = 212 {Sat May 23 15:53:16
2009 }
OemNodeConfig(): Network Address of node0: 10.1.1.1 (port 9998)
{Sat May 23 15:53:16 2009 }>TRACE: OemCreateListenPort: bound at 9998, tid = main:-1210403136 file =
oem.c, line = 858 {Sat May 23 15:53:16 2009 }
>TRACE: InitClusterDb(): found my node info at 0 name sunny1, priv
sunny1, port 3623, tid = main:-1210403136 file = nmconfig.c, line = 265 {Sat May
23 15:53:16 2009 }
OemNodeConfig(): Network Address of node1: 10.1.1.2 (port 9998)
{Sat May 23 15:53:16 2009 }>TRACE: InitClusterDb(): Local Node(0) NodeName[sunny1], tid =
main:-1210403136 file = nmconfig.c, line = 283 {Sat May 23 15:53:16 2009 }
>TRACE: InitClusterDb(): Cluster(Oracle) with (2) Defined Nodes, tid =
main:-1210403136 file = nmconfig.c, line = 286 {Sat May 23 15:53:16 2009 }
>TRACE: OEMInits(): CM Disk File (/u01/oradata/orcl/orcl/cmquorumfile),
tid = main:-1210403136 file = oem.c, line = 243 {Sat May 23 15:53:16 2009 }
>ERROR: OemInit2: Attempting to open the CMDiskFile for a multi-node RAC
on a non-NFS, non-OCFS, or non-raw device cluster, tid = main:-1210403136 file =
oem.c, line = 482 {Sat May 23 15:53:16 2009 }
>ERROR: OemInit2: If the CMDiskFile is supposed to be an NFS or OCFS file,
please make sure that the relevant shared file system is mounted properly, tid =
main:-1210403136 file = oem.c, line = 483 {Sat May 23 15:53:16 2009 }
>ERROR: OemInit2: If the CMDiskFile is supposed to be a raw device, please
make sure that it has been created properly, tid = main:-1210403136 file =
oem.c, line = 484 {Sat May 23 15:53:16 2009 }
>TRACE: IncrementEventValue: *(80f2920) = (1, 1), tid = main:-1210403136
file = unixinc.c, line = 253 {Sat May 23 15:53:17 2009 }
--- End Dump ---
sundar mahadevan
2009-May-24 22:43 UTC
[Ocfs2-users] cluster manager does not start/run ERROR: OemInit2: Attempting to open the CMDiskFile for a multi-node RAC
To add some more details to this issue: /u01/oradata/orcl/orcl/cmquorumfile mentioned in the error message is accessible from the shell prompt. I even checked the file $ORACLE_HOME/oracm/admin/cmcfg.ora and the entry in it states CmDiskFile=/u01/oradata/orcl/orcl/cmquorumfile which is accessible. So this makes me doubt, if OCFS2 does not recognise this file. Newbie. Help Plsssssssssss. On Sat, May 23, 2009 at 5:05 PM, sundar mahadevan <sundarmahadevan82 at gmail.com> wrote:> Hi Members, > > I 'm trying to install 9i RAC (initial installation) on a 2 node > cluster (ocfs2 on opensuse 11.1) on my home test boxes both 900MHz > with 512 MB RAM. I can either go for 9i or 10gR1 because of the RAM > limitation. I'm using OCFS2 and the disk is visible/accessible to both > the boxes. The permissions are set right. I'm executing the > ocmstart.sh script as root and i have root added to dba group which > has rwx on all required directories starting from /u01. I have > cmquorumfile and sharedsrvctlconfig file on /u01/oradata/orcl/orcl. > There are no error messages on /var/log/messages. I know that Oracle > does not support opensuse and supports only Enterprise Linux, Redhat, > SLES and Asianux. Newbie. Help please!!!!!!! > > > sunny1:~ # cat /u01/oradata/orcl/product/9.2/db_1/oracm/log/cm.log > oracm, version[ 9.2.0.2.0.47 ] started {Sat May 23 15:53:16 2009 } > KernelModuleName is hangcheck-timer {Sat May 23 15:53:16 2009 } > OemNodeConfig(): Network Address of node0: 10.1.1.1 (port 9998) > ?{Sat May 23 15:53:16 2009 } > OemNodeConfig(): Network Address of node1: 10.1.1.2 (port 9998) > ?{Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: Attempting to open the CMDiskFile for a multi-node RAC on a non-NFS, non-OCFS, or non-raw device cluster, tid = main:-1210403136 file = oem.c, line = 482 {Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: If the CMDiskFile is supposed to be an NFS or OCFS file, please make sure that the relevant shared file system is mounted properly, tid = main:-1210403136 file = oem.c, line = 483 {Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: If the CMDiskFile is supposed to be a raw device, please make sure that it has been created properly, tid = main:-1210403136 file = oem.c, line = 484 {Sat May 23 15:53:16 2009 } > --- DUMP GROUP STATE DB --- > --- END OF GROUP STATE DUMP --- > --- Begin Dump --- >>TRACE: ? ?LogListener: Spawned with tid 0xb70c6b90., tid = -1223922800 file = logging.c, line = 116 {Sat May 23 15:53:16 2009 } > oracm, version[ 9.2.0.2.0.47 ] started {Sat May 23 15:53:16 2009 } >>TRACE: ? ?Can't read registry value for HeartBeat, tid = main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?Can't read registry value for PollInterval, tid = main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?Can't read registry value for WatchdogTimerMargin, tid = main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?Can't read registry value for WatchdogSafetyMargin, tid = main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 } > KernelModuleName is hangcheck-timer {Sat May 23 15:53:16 2009 } >>TRACE: ? ?Can't read registry value for ClientTimeout, tid = main:-1210403136 file = unixinc.c, line = 1011 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?InitNMInfo: ?setting clientTimeout to 620s based on MissCount 620 and PollInterval 1000ms, tid = main:-1210403136 file = nmconfig.c, line = 137 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?InitClusterDb(): getservbyname on CMSrvr failed - 0 : assigning 9998, tid = main:-1210403136 file = nmconfig.c, line = 212 {Sat May 23 15:53:16 2009 } > OemNodeConfig(): Network Address of node0: 10.1.1.1 (port 9998) > ?{Sat May 23 15:53:16 2009 } >>TRACE: ? ?OemCreateListenPort: ?bound at 9998, tid = main:-1210403136 file = oem.c, line = 858 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?InitClusterDb(): ?found my node info at 0 name sunny1, priv sunny1, port 3623, tid = main:-1210403136 file = nmconfig.c, line = 265 {Sat May 23 15:53:16 2009 } > OemNodeConfig(): Network Address of node1: 10.1.1.2 (port 9998) > ?{Sat May 23 15:53:16 2009 } >>TRACE: ? ?InitClusterDb(): Local Node(0) NodeName[sunny1], tid = main:-1210403136 file = nmconfig.c, line = 283 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?InitClusterDb(): Cluster(Oracle) with (2) Defined Nodes, tid = main:-1210403136 file = nmconfig.c, line = 286 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?OEMInits(): CM Disk File (/u01/oradata/orcl/orcl/cmquorumfile), tid = main:-1210403136 file = oem.c, line = 243 {Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: Attempting to open the CMDiskFile for a multi-node RAC on a non-NFS, non-OCFS, or non-raw device cluster, tid = main:-1210403136 file = oem.c, line = 482 {Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: If the CMDiskFile is supposed to be an NFS or OCFS file, please make sure that the relevant shared file system is mounted properly, tid = main:-1210403136 file = oem.c, line = 483 {Sat May 23 15:53:16 2009 } >>ERROR: ? ?OemInit2: If the CMDiskFile is supposed to be a raw device, please make sure that it has been created properly, tid = main:-1210403136 file = oem.c, line = 484 {Sat May 23 15:53:16 2009 } >>TRACE: ? ?IncrementEventValue: *(80f2920) = (1, 1), tid = main:-1210403136 file = unixinc.c, line = 253 {Sat May 23 15:53:17 2009 } > --- End Dump --- >