如果CTSSD服务无法正常启动,则在11.2中将无法正常启动HAS=>CRS导致RAC无法正常使用。
具体的诊断需要参考$GRID_HOME/log/$HOSTNAME/ctssd/octssd.log 的日志,例如:
2010-12-23 16:45:31.287: [ GIPC][1257901344] gipcCheckInitialization: possible incompatible non-threaded init from [prom.c : 687], original from [clsss.c : 5325] [ default][1257901344]Failure 4 in trying to open SV key SYSTEM.version.localhost [ default][1257901344]procr_open_key error 4 errorbuf : PROCL-4: The local registry key to be operated on does not exist. or PROCL-5 2010-12-23 16:45:31.289: [ CTSS][1257901344]clsctss_r_av2: Error [3] retrieving Active Version from OLR. Returns [19]. 2010-12-23 16:45:31.292: [ CTSS][1257901344](:ctss_init16:): Error [19] retrieving active version. Returns [19]. 2010-12-23 16:45:31.292: [ CTSS][1257901344]ctss_main: CTSS init failed [19] 2010-12-23 16:45:31.292: [ CTSS][1257901344]ctss_main: CTSS daemon aborting [19].
触发该错误,主要有2种可能:
- ocr所在ASM DISK的用户权限存在问题
- ocr所在ASM DISK在节点之间设备名不一致
- HAS的状态不正常
- bug 10420872
对于问题2 建议通过kfed read /dev/oracleasm/disks/OCRVOTE01 | grep kfbh.type 来确认节点之间的ASM DISK一致。
对于问题3 可以尝试 重启 HAS 服务:
su - root crsctl stop has -f crsctl start has
Comment