# Problem description
When running in DB2 pureScale environment, the capture program
of Q replication terminates with error SQL2038N:
# Diagnostic information
In the capture log from Q replication, the following error
messages are seen:
2017-08-09-11.22.04.294755 ASN8041D
"Q Capture" : "PRD2DRP" : "LogrdThread" : db2LogRead API is
sendi
ng us backwards in theDB2 Log: First LSN is
"0000:0007:05e9:5cff:0000:0159:b38c:6f4a" while Next Start LSN
is "0000:0000:0000:00
00:0000:0000:0000:0000"
2017-08-09-11.22.04.294914 ASN0552E
"Q Capture" : "PRD2DRP" : "logrd" : The program encountered an
SQL error. The server name is "PROD". The SQL request is
"db2ReadLog". The table name is "". The SQLCODE is "-2038". The
SQLSTATE is " ". The SQLERRMC is "". The SQLERRP is
"SQLPRLOG".
SQL2038N A database system error "" occurred during processing.
In db2diag.log, A "checksum error" is reported for a particular
transaction log. In this case, it's S0000091.LOG.
2017-08-09-11.21.48.603790+540 I17348223A570 LEVEL: Info
PID : 22545070 TID : 2521457 PROC :
db2sysc 0
INSTANCE: db2prod NODE : 000 DB : PROD
APPHDL : 0-4084 APPID: *N0.db2prod.170809012214
AUTHID : db2prod HOSTNAME: prodsrv01
EDUID : 2521457 EDUNAME: db2shred.1 (PROD) 0
FUNCTION: DB2 UDB, recovery manager, sqlpshrEdu, probe:14000
DATA #1 :
Merging LFR scan flags after processing 16 pages.
Ping extent flags: 842
Pong extent flags: 3
Merged flags: 843
2017-08-09-11.22.03.947425+540 I17348794A505 LEVEL:
Warning
PID : 22545070 TID : 2754963 PROC :
db2sysc 0
INSTANCE: db2prod NODE : 000 DB : PROD
HOSTNAME: prodsrv01
EDUID : 2754963 EDUNAME: db2lfr.3 (PROD) 0
FUNCTION: DB2 UDB, recovery manager, sqlplfrVerifyLogPages,
probe:500
DATA #1 :
LFR Scan Num = 749
LFR Scan Caller's EDUID = 2753935
Checksum error when reading page 32773 from log file
S0000091.LOG
Alternatively, in the db2diag.log, you may see a message like:
2018-08-28-00.40.17.163119+540- I35517094A570 LEVEL:
Warning
PID : 20054628 TID : 228098 PROC :
db2sysc 4
INSTANCE: db2inst1 NODE : 004 DB :
DBNAME
APPHDL : 4-247 APPID:
+ACo-N4.db2inst1.180824224024
AUTHID : QCAPTURE HOSTNAME: HOST1
EDUID : 228098 EDUNAME: db2shred.3 (DBNAME) 4
FUNCTION: DB2 UDB, recovery manager, sqlpshrEdu, probe:6985
DATA +ACM-1 : +ADw-preformatted+AD4-
Standby did not receive close information for 108513. Moving LSO
from 28307396094303 to 28307396096001
The capture program of Q replication has to be restarted.
Normally it does not encounters the error on this log file any
more. After the incident, running db2cklog against the log
S0000091.LOG returns success and no checksum error is found.
This issue is a timing issue which could occur when Db2
reads from remote log stream. Last page can be read while
written to, causing checksum error. |
