During "db2cluster -cm -enter -maintenance -all" command
processing, a check is performed to ensure that no mounts are in
maintenance on any host in the cluster. Sometimes, this check
incorrectly detects some mounts to be in maintenance mode on
remote hosts even though they are not.
To diagnose this issue, run "db2cluster -cm -enter -maintenance
-all" with "-debug" option. In the generated ibm.db2.cluster.*
file, check for the following snippets:
2018-12-14-10.07.51.059031+540 I460379A1197 LEVEL: Info
PID : 34603410 TID : 1 PROC :
db2cluster
INSTANCE: NODE : 000
HOSTNAME: db2host1
EDUID : 1
FUNCTION: DB2 UDB, high avail services,
sqlhaMountMaintenanceDoesFileExistOnHost, probe:0
DATA #1 : Codepath, 8 bytes
4
DATA #2 : String, 31 bytes
/var/db2/.db2_file
DATA #3 : String, 10 bytes
db2host2
DATA #4 : signed integer, 4 bytes
0
DATA #5 : String, 10 bytes
db2host1
DATA #6 : Boolean, 1 bytes
true
DATA #7 : String, 10 bytes
db2host1
DATA #8 : unsigned integer, 2 bytes
5
DATA #9 : String, 120 bytes
/var/db2/.db2_file not found
< - - - - DATA#9 should contain "not found" string
failure - examine the system log on the remotehost for
additional information
CALLSTCK: (Static functions may not be resolved correctly, as
they are resolved to the nearest symbol)
[0] 0x000000010007449C
sqlhaMountMaintenanceDoesFileExistOnHost + 0x7FC
[1] 0x00000001002310D0
IPRA.$sqlhaDisplayMountsWithConflictingCMandMountMaintenanceMode
__FCbCPb + 0xA30
[2] 0x0000000100230084
sqlhaUIEnterMaintenanceModeAll__FP9sqlhaUICB + 0x1204
[3] 0x0000000100005E7C main + 0x4B9C
[4] 0x00000001000002F8 __start + 0x70
..
2018-12-14-10.07.51.060700+540 I463849A337 LEVEL: Error
PID : 34603410 TID : 1 PROC :
db2cluster
INSTANCE: NODE : 000
HOSTNAME: db2host1
EDUID : 1
FUNCTION: DB2 UDB, high avail services,
sqlhaUIEnterMaintenanceModeAll, probe:1624
MESSAGE : Failed to enter maintenance mode. |
