When db2 is shutting down database, it attempts to flush all
the dirty pages in bufferpool. But it finds some dirty pages of
temp tables has no owner. This stale direty pages cause the
db2agent stuck in the following place:
##### Object: /usr/lib/sparcv9/libc.so.1
__systemcall(0xffffffff7ffee3d0,0x35,0x2,0x1cf019f,0xffffffff7ff
ee490,0x1) + 0x34
semop(0x1cf019f,0xffffffff7ffee490,0x1,0x3750,0x1ffff,0x20000) +
0x18
##### Object: /SAN03/data/db2/db2dmp01/sqllib/adm/db2sysc
__1cRSQdDLO_SLATCH_CAS64NstateToString6FklkpckL_v_(0x20001cd58,0
x0,0x10000,0x280000,0x0,0x18780000) + 0x130
__1cRSQdDLO_SLATCH_CAS64SgetConflictComplex6MkL_v_(0x221628960,0
x0,0x20001cd58,0x1ffff,0x2a0a28,0x28) + 0x470
__1cRSQdDLO_SLATCH_CAS64LgetConflict6MkL_i_(0x221628960,0x10000,
0x1fc00,0x1fc00,0x10000,0x10000) + 0xf0
__1cTsqlbVerifyNoDirties6FpnMSQdDLB_GLOBALS_pnQSQdDLB_OBJECT_DES
C_I_v_(0x221628960,0x10000,0x102c1d140,0x0,0x1,0x1) + 0x106c
__1cUsqlbFlushForDLSubset6FpnJSQdDLP_LSN8_IHiipnMSQdDLB_GLOBALS_
_i_(0x221a33e58,0x10000,0x0,0x102c09d80,0x102dd02c0,0xffffffffff
fc0000) + 0xd48
__1cIsqlpthrd6FpnUsqle_agent_privatecb_L_i_(0x102c08180,0x100000
,0x7bc8,0x220010080,0x220017080,0x8a00) + 0x2f4
__1cSsqlbTerminateAgent6FpnMSQdDLB_GLOBALS_i_i_(0x102c09d80,0x34
3,0x1,0x0,0x0,0x1) + 0xd54
__1cTsqleAgentDissociate6FpnNsqle_agent_cb_i_i_(0x200eb8080,0xff
ffffff7ffef41c,0x0,0x338,0x102b80630,0x100) + 0x3a0
__1cRsqleTermDbConnect6FpnNsqle_agent_cb_pnIsqledbcb_pnFsqlca_pn
VsqleTermConnectStatus__i_(0x200eb8080,0x0,0x102c08368,0xfffffff
f7ffef8f4 ,0x200eb0080,0x17e4) + 0x20fc
__1cIsqlesrsp6FpnUsqle_agent_privatecb_CpnFsqlca_i_i_(0x102c0818
0,0x1000000,0x102c08368,0x18280000,0x0,0x200eb8080) + 0xc84
db2pd -tcbstats may show some temp table exists whose
Application Handle does not exist. For example, in db2pd
-tcbstats, you may find "TEMP (00030,00006) <1537><S" where 1537
is the Application Handle created this TEMP table ID 6 in
tempspace ID 30. But Application Handle 1537 does not exist in
db2pd -applications nor in db2 snapshot for applications. |
