Article ID: 315405
Article Last Modified on 2/22/2007
APPLIES TO
- Microsoft Host Integration Server 2000 Standard Edition
This article was previously published under Q315405
SYMPTOMS
COM+ components that use ActiveX Data Objects (ADO) and connection pooling to do two-phase commit (2PC) transactions with either the Open Database Connectivity(ODBC) Driver for DB2 or with OLE DB provider for DB2 (DB2OLEDB) may receive the following SEND state error under load:
CAUSE
The underlying code that implements the 2PC transactions has a timing problem, which can become evident under load. During the timing problem, a transaction is marked as complete and is returned to the connection pool. However, all of the Syncpoint commands have not yet finalized the transaction on the host system. If another client uses this connection again before the Syncpoint commands have completed, the error occurs.
RESOLUTION
To resolve this problem, obtain the latest service pack for Host Integration Server 2000. For additional information, click the following article number to view the article in the Microsoft Knowledge Base:
328152 How to Obtain the Latest Host Integration Server 2000 Service Pack
The English version of this fix has the file attributes (or later) that are listed in the following table. The dates and times for these files are listed in coordinated universal time (UTC). When you view the file information, it is converted to local time. To find the difference between UTC and local time, use the Time Zone tab in the Date and Time tool in Control Panel:
Date Time Version Size File name
--------------------------------------------------------------
03-Jun-2002 12:33 5.0.0.836 221,456 Wappc32.dll
NOTE: Because of file dependencies, the most recent fix that contains the preceding files may also contain additional files.
WORKAROUND
To work around this problem, disable either ODBC Connection pooling or OLE DB Resource pooling, depending on which technology is being used by the COM+ component.
STATUS
Microsoft has confirmed that this is a problem in Microsoft Host Integration Server 2000. This problem was first corrected in Host Integration Server 2000 Service Pack 1.
MORE INFORMATION
If you examine the DB2APIx.atf trace file to find the symptoms, you see that the DB2 provider that is in use tries to issue a distributed data management (DDM) architecture command, which results in a STATE_CHECK error response:
APPC ----------------------------------------------- 12:48:35.0234
APPC SEND_DATA request
APPC
APPC Type:Sync VerbID:013D4002
APPC
APPC ---- Verb Parameter Block at address 0148EDC0 ----
APPC 0F000000 00000000 00000000 02000000 <................>
APPC 10AF1D01 70501601 00004E00 5A00EB0E <....pP....N.Z...>
APPC 00000000 <.... >
APPC ---- Data at address 0EEB005A ----
APPC 004ED051 00010048 200A0044 2113C4C2 <.N.Q...H ..D!.DB>
APPC F2C5C4D7 40404040 40404040 40404040 <2EDP@@@@@@@@@@@@>
APPC C4C2F2E2 C1D2C9E2 40404040 40404040 <DB2AARON@@@@@@@@>
APPC 4040D4E2 E4D9F0F0 F1404040 40404040 <@@MSAMG01@@@@@@@>
APPC 40404040 F5F0F0F0 F0F0F0F2 0001 <@@@@50000002.. >
APPC ----------------------------------------------- 12:48:35.0234
APPC SEND_DATA response, result = STATE_CHECK
APPC
APPC Type:Sync VerbID:013D4002
APPC
APPC ---- Verb Parameter Block at address 0148EDC0 ----
APPC 0F000000 00020000 000000F2 02000000 <...........2....>
APPC 10AF1D01 70501601 00004E00 5A00EB0E <....pP....N.Z...>
APPC 00000000 <.... >
You can also identify this problem by looking up the return codes in the Host Integration Server (HIS) 2000 SDK. Note the following:
- A primary return code of 0x0002 indicates AP_STATE_CHECK.
- A secondary return code of 0x000000F2 indicates AP_SEND_DATA_NOT_SEND_STATE.
Although the last request that was issued by the DB2 provider was a RECEIVE_AND_POST request, the final asynchronous response has not yet been received. You can see this by following the APPC Conversation ID in the trace:
APPC ----------------------------------------------- 12:48:35.0218
APPC RECEIVE_AND_POST request
APPC
APPC Type:Sync VerbID:013B4002
APPC
APPC ---- Verb Parameter Block at address 01122AE8 ----
APPC 0D000000 00000000 00000000 02000000 <................>
APPC 10AF1D01 70501601 00000101 00000002 <....pP..........>
APPC 00000000 84FDA70E FEFFFFFF 00000000 <....d.x.........>
APPC ----------------------------------------------- 12:48:35.0218
APPC RECEIVE_AND_POST response, result = OK
APPC
APPC Type:Sync VerbID:013B4002
APPC
APPC ---- Verb Parameter Block at address 01122AE8 ----
APPC 0D000000 00000000 00000000 02000000 <................>
APPC 10AF1D01 70501601 00040101 00000002 <....pP..........>
APPC 00000000 84FDA70E FEFFFFFF 00000000 <....d.x.........>
The correct final response, which is the Syncpoint Forget command from the host, is received almost immediately after the SEND_STATE error occurs, therefore demonstrating the timing problem:
APPC ----------------------------------------------- 12:48:35.0250
APPC RECEIVE_AND_POST response, result = OK
APPC
APPC Type:Async Method:Post Hwnd:FFFFFFFE VerbID:013B4002
APPC
APPC ---- Verb Parameter Block at address 01122AE8 ----
APPC 0D000000 00000000 00000000 02000000 <................>
APPC 10AF1D01 70501601 01020101 00000002 <....pP..........>
APPC 06000000 84FDA70E FEFFFFFF 00000000 <....d.x.........>
APPC ---- Data at address 0EA7FD84 ----
APPC 0001040A 0008 <...... >
For more information about DDM commands, see the documentation on the following The Open Group Web site:
Database Interoperability Consortium
http://www.opengroup.org/dbiop/
For more information about Syncpoint and 2PC with IBM host systems, see the following IBM documentation:
Sync Point Services Architecture Reference
IBM document number SC31-8134-00
Keywords: kbbug kbfix kbhostintegserv2000presp1fix kbhostintegserv2000sp1fix KB315405
