In a huge DPF environment, snapshot table functions like
SYSIBMADM.SNAPTBSP_PART may fail with the following error:
SQL0443N Routine "SNAP_GET_TBSP_PART_V91" (specific name
"*P_GET_TBSP_PART_V91_AP") has returned an error SQLSTATE with
diagnostic text "Status error for file: ". SQLSTATE=38553
While calling snapshot table functions, db2 fetches data from
all logical nodes of all the physical nodes. During this
operation, temporary files which get created to serve data
retrieval may get unlinked incorrectly and cause the function to
fail with SQL0443N.
This condition can occur if sqllib/tmp directory exists on NFS
mounted filesystem.
db2diag.log from the co-ordinating node shows:
XXXX-XX-XX-XX.XX.XX.XXXXXX+XXX I696374A430 LEVEL: Info
PID : 3273166 TID : 1 PROC :
db2agent (TESTDB)74
INSTANCE: db2inst1 NODE : 074 DB : TESTDB
APPHDL : 74-54 APPID: *N74.db2inst1.090616160251
AUTHID : DB2INST1
FUNCTION: DB2 UDB, base sys utilities, sqleMergeSqlca, probe:20
MESSAGE : Received sqlcode -443 for request 80000000 from node
number 212
The corresponding entries in the db2diag.log file from node 212
show
XXXX-XX-XX-XX.XX.XX.XXXXXX+XXX I146013A355 LEVEL: Error
(OS)
PID : 4051408 TID : 772
FUNCTION: DB2 Common, OSSe, OSSHFile::getSize, probe:10
MESSAGE : ECF=0x90000034 File is on a remote system and the link
to that system is not active
CALLED : OS, -, fstat OSERR: ESTALE
(52)
ARG #1 : Hex, 4 bytes
0x6
When fstat fails, the attempt to close the temporary file ends
in an error like this --
XXXX-XX-XX-XX.XX.XX.XXXXXX+XXX I147081A335 LEVEL: Error
(OS)
PID : 2942268 TID : 772
FUNCTION: DB2 Common, OSSe, OSSHFile::close, probe:10
MESSAGE : ECF=0x90000038 The path contains an unmounted virtual
file system
CALLED : OS, -, close OSERR: ESTALE
(52)
ARG #1 : Hex, 4 bytes
0x6 |
