Description:
It maybe nice when the ndb_mgmd is rotating the cluster log automatically to note this on top of the next log together with some more information.
Reasoning:
ndb_mgmd will keep rotating, and since Cluster is supposed to never go down (hehe) the log files could fill up with 'noise'. So, if you check cluster logs, it's pretty dull. If you put some generic information at least on top of each file, you have something to go with.
How to repeat:
(feature request)
Suggested fix:
For example, when a cluster log file rotates, on top, we might find:
.. [MgmSrvr] INFO -- Cluster log rotated.
.. [MgmSrvr] INFO -- NDB Cluster Management Server. mysql-5.1.35 ndb-6.3.26-GA
.. [MgmSrvr] INFO -- Configuration 1 used
.. [MgmSrvr] INFO -- Node 3: Data usage is 0%(22 32K pages of total 67200)
.. [MgmSrvr] INFO -- Node 3: Index usage is 0%(16 8K pages of total 64032)
.. [MgmSrvr] INFO -- Node 3: Resource 0 min: 67242 max: 67967 curr: 86
.. [MgmSrvr] INFO -- Node 3: Resource 3 min: 67200 max: 67200 curr: 22
.. [MgmSrvr] INFO -- Node 3: Resource 5 min: 128 max: 128 curr: 64
.. [MgmSrvr] INFO -- Node 4: Data usage is 0%(22 32K pages of total 67200)
.. [MgmSrvr] INFO -- Node 4: Index usage is 0%(16 8K pages of total 64032)
.. [MgmSrvr] INFO -- Node 4: Resource 0 min: 67242 max: 67967 curr: 86
.. [MgmSrvr] INFO -- Node 4: Resource 3 min: 67200 max: 67200 curr: 22
.. [MgmSrvr] INFO -- Node 4: Resource 5 min: 128 max: 128 curr: 64
Thus forcing to how the:
* version
* the status of the data nodes (through forced ALL DUMP 1000)
* and show what configuration was active
© 2024, Oracle Corporation and/or its affiliates