DB2 - Problem description
Problem IC68726 | Status: Closed |
WHEN NO MORE FCM BUFFER AVAILABLE, IT MAY CAUSE DB2 OPERATIONS (E.G. SNAPSHOT MONITORING) TO HANG. | |
product: | |
DB2 FOR LUW / DB2FORLUW / 980 - DB2 | |
Problem description: | |
This problem effects v91, v95, v97. In partition database environment (where multiple nodes needs to communicate with each other via FCM), if one of the nodes run out of the FCM buffer to communicate back (send reply) to other nodes, it could cause the receiver nodes to wait forever for the reply, which causes hang. The following db2diag.log entries will be written when snapshot monitoring runs out of FCM buffers: 2010-03-25-02.04.24.272731+000 I14931064A484 LEVEL: Error PID : 648562 TID : 1 PROC : db2agntp 1 INSTANCE: xxxxxx NODE : 001 APPHDL : 0-413 FUNCTION: DB2 UDB, fast comm manager, sqkfBufferManager::AllocBuffer, probe:55 MESSAGE : Ran out low priority buffers. Total=246125; Num in freelist=49226; med thres=49225; hi thres=24612 DATA #1 : Hexdump, 4 bytes 0x0FFFFFFFFFFF6184 : 0000 000A .... 2010-03-25-02.04.24.272901+000 I14931549A332 LEVEL: Warning PID : 648562 TID : 1 PROC : db2agntp 1 INSTANCE: xxxxxx NODE : 001 APPHDL : 0-413 FUNCTION: DB2 UDB, buffer dist serv, sqlkdSendReply, probe:80 RETCODE : ZRC=0x85590027=-2057764825=SQLKF_NO_BUFFER "No FCM Buffer available" 2010-03-25-02.04.24.273315+000 I14931882A376 LEVEL: Warning PID : 648562 TID : 1 PROC : db2agntp 1 INSTANCE: xxxxxx NODE : 001 APPHDL : 0-413 FUNCTION: DB2 UDB, buffer dist serv, sqlkdSendReply, probe:90 MESSAGE : Replying to node = DATA #1 : Hexdump, 2 bytes 0x0FFFFFFFFFFF6570 : 0000 .. 2010-03-25-02.04.24.273460+000 I14935899A303 LEVEL: Warning PID : 648562 TID : 1 PROC : db2agntp 1 INSTANCE: db2ob8 NODE : 001 APPHDL : 0-413 FUNCTION: DB2 UDB, base sys utilities, sqlePdbRemap, probe:100 MESSAGE : ZRC 85590027 remapped to SQLCODE -6040 2010-03-25-02.04.24.625916+000 I14936203A293 LEVEL: Severe PID : 648562 TID : 1 PROC : db2agntp 1 INSTANCE: xxxxxx NODE : 001 APPHDL : 0-413 FUNCTION: DB2 UDB, database monitor, sqm_sqlmPdbRequestRouter, probe:90 RETCODE : ZRC=0xFFFFE868=-6040 Stacktrace on the receiver node will be similar to the following: <StackTrace> -------Frame------ WaitRecvReady ReceiveBuffer getNextReply getNextBuffer sqlkd_rcv_buffer sqlkd_rcv_get_next_buffer sqlkd_rcv_init sqlkdReceiveReply sqlmReceiveAndProcessReplies sqlmSendMonitorRequest sqlmonssbackend sqlesrvr sqleMappingFnServer sqlerKnownProcedure sqlerCallDL sqljs_ddm_excsqlstt sqljsParseRdbAccessed sqljsParse | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * ALL * **************************************************************** * PROBLEM DESCRIPTION: * * See Problem Description above. * **************************************************************** * RECOMMENDATION: * * Upgrade to DB2 Version 9.8 Fix Pack 4. * **************************************************************** | |
Local Fix: | |
Ensure FCM buffers is set to an appropriate value. If needed increase the database manager configuration parameter FCM_NUM_BUFFERS. | |
available fix packs: | |
DB2 Version 9.8 Fix Pack 4 for AIX and Linux | |
Solution | |
First fixed in DB2 Version 9.8 Fix Pack 4. | |
Workaround | |
not known / see Local fix | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 19.05.2010 05.08.2011 05.08.2011 |
Problem solved at the following versions (IBM BugInfos) | |
9.8.FP4 | |
Problem solved according to the fixlist(s) of the following version(s) | |
9.8.0.4 |