11204 in exadata x6
Hi,
MRP has been hung since 5 hours. The lag grows. MRP says waiting for log. RFS is shipping all logs.
The primary crashed and was re-booted. Dataguard broker has the standby database with mode 'APPLY-ON', but MRP was hung.
We only come to know when our script runs.
Alert logs MRP hangs at this point
Fri Sep 02 12:35:41 2016
Media Recovery Waiting for thread 1 sequence 9510
Fri Sep 02 12:36:30 2016
Archived Log entry 4749 added for thread 2 sequence 10172 ID 0x52ae51d3 dest 1:
Fri Sep 02 12:36:37 2016
RFS[39]: Selected log 11 for thread 2 sequence 10173 dbid 1380968317 branch 915104253
Fri Sep 02 12:38:10 2016
RFS[36]: Possible network disconnect with primary database
Fri Sep 02 12:38:10 2016
RFS[37]: Possible network disconnect with primary database
Fri Sep 02 12:47:21 2016
RFS[40]: Assigned to RFS process 240738
RFS[40]: Selected log 7 for thread 1 sequence 9512 dbid 1380968317 branch 915104253
Fri Sep 02 12:47:21 2016
Archived Log entry 4752 added for thread 1 sequence 9512 ID 0x52ae51d3 dest 1:
I stop the database and start it, everything comes back to normal and the database catches up. Doesnt even copy the archivelogs (which by the way are deleted at the primary already). Why does MRP get hung like that? There is no corruption!
Also in a real-time query on environment (Read only with apply) the checksum of the archivelogs on the standby side are different. Is that normal?
I understand that the primary writes directly into the standby log file, which in turn creates the archivelogs.
Please do let me know.
.
Thanks!