09-16-2010 09:20 PM
Hi All,
I am using the CFSTypes.cf and CVMTypes.cf file types and am not able to mount the diskgroup named "oracledg" on VCS nodes. I am enclosing the main.cf file used in cluster configuration. Can anyone please help?
I get the following error message:
010/09/17 07:46:02 VCS ERROR V-16-20007-1010 (node1) CVMVolDg:oracle_cvmvoldg:online:online_change_activation: can not change activation of dg oracledg to shared-write
2010/09/17 07:46:02 VCS WARNING V-16-20007-1025 (node1) CVMVolDg:oracle_cvmvoldg:online:Can not set diskgroup oracledg activation to sw
2010/09/17 07:46:02 VCS ERROR V-16-20007-1045 (node1) CVMVolDg:oracle_cvmvoldg:online:Initial check failed
2010/09/17 07:46:03 VCS INFO V-16-2-13716 (node1) Resource(oracle_cvmvoldg): Output of the completed operation (online)
==============================================
VxVM vxdg ERROR V-5-1-3268 activation failed: Disk group oracledg: shared-write: Invalid mode for non-shared disk group
RROR V-16-2-13066 (node2) Agent is calling clean for resource(oracle_cvmvoldg) because the resource is not up even after online completed.
2010/09/17 07:48:01 VCS INFO V-16-2-13716 (node2) Resource(oracle_cvmvoldg): Output of the completed operation (clean)
==============================================
/var/VRTSvcs/lock/oracle_cvmvoldg_oracledg_stat: No such file or directory
==============================================
2010/09/17 07:48:01 VCS INFO V-16-2-13068 (node2) Resource(oracle_cvmvoldg) - clean completed successfully.
2010/09/17 07:48:01 VCS INFO V-16-2-13071 (node2) Resource(oracle_cvmvoldg): reached OnlineRetryLimit(2).
2010/09/17 07:48:02 VCS ERROR V-16-1-10303 Resource oracle_cvmvoldg (Owner: unknown, Group: cfs) is FAULTED (timed out) on sys node2
2010/09/17 07:48:02 VCS NOTICE V-16-1-10300 Initiating Offline of Resource cfs_online_phantom (Owner: unknown, Group: cfs) on System node2
2010/09/17 07:48:04 VCS INFO V-16-1-10305 Resource cfs_online_phantom (Owner: unknown, Group: cfs) is offline on node2 (VCS initiated)
2010/09/17 07:48:04 VCS ERROR V-16-1-10205 Group cfs is faulted on system node2
2010/09/17 07:48:04 VCS NOTICE V-16-1-10446 Group cfs is offline on system node2
2010/09/17 07:48:04 VCS INFO V-16-1-10493 Evaluating node1 as potential target node for group cfs
2010/09/17 07:48:04 VCS INFO V-16-1-50010 Group cfs is online or faulted on system node1
2010/09/17 07:48:04 VCS INFO V-16-1-10493 Evaluating node2 as potential target node for group cfs
2010/09/17 07:48:04 VCS INFO V-16-1-50010 Group cfs is online or faulted on system node2
2010/09/17 07:48:04 VCS ERROR V-16-2-13066 (node1) Agent is calling clean for resource(oracle_cvmvoldg) because the resource is not up even after online completed.
2010/09/17 07:48:05 VCS INFO V-16-6-0 (node2) postoffline:Invoked with arg0=node2, arg1=cfs
2010/09/17 07:48:05 VCS INFO V-16-6-15002 (node2) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/postoffline node2 cfs successfully
2010/09/17 07:48:05 VCS INFO V-16-2-13716 (node1) Resource(oracle_cvmvoldg): Output of the completed operation (clean)
==============================================
/var/VRTSvcs/lock/oracle_cvmvoldg_oracledg_stat: No such file or directory
==============================================
2010/09/17 07:48:05 VCS INFO V-16-2-13068 (node1) Resource(oracle_cvmvoldg) - clean completed successfully.
2010/09/17 07:48:05 VCS INFO V-16-2-13071 (node1) Resource(oracle_cvmvoldg): reached OnlineRetryLimit(2).
2010/09/17 07:48:07 VCS ERROR V-16-1-10303 Resource oracle_cvmvoldg (Owner: unknown, Group: cfs) is FAULTED (timed out) on sys node1
2010/09/17 07:48:07 VCS NOTICE V-16-1-10300 Initiating Offline of Resource cfs_online_phantom (Owner: unknown, Group: cfs) on System node1
2010/09/17 07:48:08 VCS INFO V-16-1-10305 Resource cfs_online_phantom (Owner: unknown, Group: cfs) is offline on node1 (VCS initiated)
2010/09/17 07:48:08 VCS ERROR V-16-1-10205 Group cfs is faulted on system node1
2010/09/17 07:48:08 VCS NOTICE V-16-1-10446 Group cfs is offline on system node1
2010/09/17 07:48:08 VCS INFO V-16-1-10493 Evaluating node1 as potential target node for group cfs
2010/09/17 07:48:08 VCS INFO V-16-1-50010 Group cfs is online or faulted on system node1
2010/09/17 07:48:08 VCS INFO V-16-1-10493 Evaluating node2 as potential target node for group cfs
2010/09/17 07:48:08 VCS INFO V-16-1-50010 Group cfs is online or faulted on system node2
2010/09/17 07:48:10 VCS INFO V-16-6-0 (node1) postoffline:Invoked with arg0=node1, arg1=cfs
2010/09/17 07:48:10 VCS INFO V-16-6-15002 (node1) hatrigger:hatrigger executed /opt/VRTSvcs/bin/triggers/postoffline node1 cfs successfully
Best Regards,
Bruce
Solved! Go to Solution.
09-19-2010 04:44 AM
Ok..
So,
-- whether disk_0 is a shared disk & visible across the cluster nodes ? You can verify this with serial number of the disk:
# /etc/vx/diag.d/vxdmpinq /dev/vx/rdmp/disk_0
OR
If you run vxdisk -o alldgs list on node 2, you should see disk ora1 imported on second node as well....
-- Secondly, error appearing in engine-A.log says, no such device or address.... that means volume or FS is not ready..
Since you imported the diskgroup manually, did you started the volumes mannually ? can u see if volumes inside diskgroup are enabled active ?
# vxprint -qthg oracledg (all the volumes should be ENABLED ACTIVE), any other state won't work...
To start volumes
# vxvol -g oracledg startall
Once volumes are ENABLED ACTIVE, can you see a filesystem on it ?
# fstyp -v /dev/vx/rdsk/oracledg/oravol
If you see the filesystem, it should be able to mount...
Since you are using cfs commands, I believe you have already added mounts using cfsmntadm... If not then have a look at man page of cfsmntadm..
Gaurav
09-18-2010 02:42 AM
what is the SF version & OS version ?
Is the diskgroup oracledg imported ? (check with # vxdg list)
Point to note us, CVMVolDG agent doesn't import the diskgroup, as part of "online" the resource, it just changes the activation mode of the diskgroup... see here
http://sfdoccentral.symantec.com/sf/5.1/solaris/html/sfcfs_admin/ch05s09.htm
In case diskgroup is not imported... I would suggest to manually import the diskgroup from CVM master node..
--Freeze the cluster group
# hagrp -freeze cfs -persistent
-- identify master node
# vxdctl -c mode
-- from master node
# vxdg -s import oracledg
-- unfreeze group
# hagrp -unfreeze cfs -persistent
-- probe resource
# hares -probe oracle_cvmvoldg
# hagrp -online cfs -sys <node1/node2> (select proper node)
Gaurav
09-18-2010 06:54 AM
I also checked the online script for CVMVolDg agent in /opt/VRTSvcs/bin/CVMVolDg directory.... the error you are getting
2010/09/17 07:46:02 VCS ERROR V-16-20007-1045 (node1) CVMVolDg:oracle_cvmvoldg:online:Initial check failed
See the error code above is 1045, from online script:
# check if everything is imported and enabled etc.
cvmvoldg_check_all "ONLINE" _cvm_res
if [ $_cvm_res -ne 0 ] ; then
VCSAG_LOG_MSG "E" "Initial check failed" 1045
exit $CVMVOLDG_FAILURE
fi
seeing above, I think my first comment looks on the right track..... see if diskgroup is imported... if yes, then whether it is imported with shared flag or no..
# vxdg list
Gaurav
09-19-2010 01:43 AM
Hi,
I made the changes incluster and have created the shared diskgroup. I just want to mount a cluster file system on a 2-node cluster but am unable to get the desired result.
node1 * / # vxdisk -o alldgs list
DEVICE TYPE DISK GROUP STATUS
c0t0d0s2 auto:SVM - - SVM
disk_0 auto:cdsdisk ora1 oracledg online shared
disk_1 auto:none - - online invalid
disk_2 auto:none - - online invalid
disk_3 auto:none - - online invalid
disk_4 auto:none - - online invalid
node1 * / # cfscluster status
Node : node1
Cluster Manager : running
CVM state : running
MOUNT POINT SHARED VOLUME DISK GROUP STATUS
/oracle oravol oracledg NOT MOUNTED
I am getting the following error:
node1 * / # cfsmount /oracle
Mounting...
Error: V-35-50: Could not mount [/dev/vx/dsk/oracledg/oravol] at /oracle on node1
UX:vxfs mount: ERROR: V-3-24706: /dev/vx/dsk/oracledg/oravol no such device or filesystem on it missing one or more devices
09-19-2010 04:44 AM
Ok..
So,
-- whether disk_0 is a shared disk & visible across the cluster nodes ? You can verify this with serial number of the disk:
# /etc/vx/diag.d/vxdmpinq /dev/vx/rdmp/disk_0
OR
If you run vxdisk -o alldgs list on node 2, you should see disk ora1 imported on second node as well....
-- Secondly, error appearing in engine-A.log says, no such device or address.... that means volume or FS is not ready..
Since you imported the diskgroup manually, did you started the volumes mannually ? can u see if volumes inside diskgroup are enabled active ?
# vxprint -qthg oracledg (all the volumes should be ENABLED ACTIVE), any other state won't work...
To start volumes
# vxvol -g oracledg startall
Once volumes are ENABLED ACTIVE, can you see a filesystem on it ?
# fstyp -v /dev/vx/rdsk/oracledg/oravol
If you see the filesystem, it should be able to mount...
Since you are using cfs commands, I believe you have already added mounts using cfsmntadm... If not then have a look at man page of cfsmntadm..
Gaurav
09-19-2010 11:52 PM
Hi Gaurav,
I recreated the shared disk group, followed by cluster file system and it is working fine now. I guess I must have created the shared disk group from the "slave" cluster node earlier. Thank you for your help.
Regards,
Bruce
09-20-2010 01:05 AM
good to know its working..
just FYI, all config changes can happen from master node only... might have missed something else :)
Gaurav