|
View previous topic :: View next topic
|
| Author |
Message |
John Poulakos
Active User
Joined: 13 Jun 2012
Posts: 178
Location: United States
|
|
|
|
I am working with an old proprietary software product and the damn thing intercepts DB2 and puts out its own useless user abend. From the message it puts out, I can tell it's a DB -310 error, but I can't tell which host variable is causing the error. I'm hoping its in the SQLCA, but I am having trouble interpreting it. This is all the information I have:
********************************* TOP OF DATA **********************************
*** SYSTEM FATAL ERROR:
DATE: 12/26/18 TIME: 22:44:15
***** LAST DBIO CALL:
DBFUNC: INS DBNUM: 0061 DBSTATUS: *DB27 STATUS-2: 0036
BUFFER: |201900011021021 22991021220 OTHER ST GRANTS & REIMBU|
|R 0000000000000000000000000000000000000000000000000000000000|
|0000000000000000000130|
***** LAST DBMS CALL IS DB2:
FUNC : INSERT CURSOR: STATUS: -310
TABLE : RRPT
BUFFER: 201900011021021 22991021220 OTHER ST GRANTS & REIMB
000
0000
SQLCA : *E2D8D3C3C140404000000088FFFFFECA0002F1F3* *SQLCA h 13*
: *4040404040404040404040404040404040404040* * *
: *4040404040404040404040404040404040404040* * *
: *4040404040404040404040404040404040404040* * *
: *4040404040404040C4E2D5E7D9C9C8C2FFFFFF12* * DSNXRIHB *
: *4040404040404040404040F2F2F0F2F3A0286F1C* * 22023 ? *
******************************** BOTTOM OF DATA ******************************** |
|
| Back to top |
|
 |
John Poulakos
Active User
Joined: 13 Jun 2012
Posts: 178
Location: United States
|
|
|
|
| I believe the host variable number is the 13 at the end of the row. It is the beginning of errmsg and nothing else would hold this content. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3051
Location: NYC,USA
|
|
|
|
| Quote: |
| I believe the host variable number is the 13 at the end of the row. It is the beginning of errmsg and nothing else would hold this content. |
If that's so then that must be a Decimal field having a nondecimal data and if not then identify all the Decimal fields present and look for the values before INSERT. |
|
| Back to top |
|
|
John Poulakos
Active User
Joined: 13 Jun 2012
Posts: 178
Location: United States
|
|
|
|
Without knowing which host variable is giving the error, I don't even know if it's a field I think is numeric. All the fields that are numeric according to the file definition that feeds this program are zeros (as shown in the buffer). But, I don't have the actual source code and I can only see the SQL that is shown in the DBRM, which doesn't show host variables. Knowing which host variable is causing the error would at least offer some clue, which is more than I have now.
This is a 20+ year old legacy system that is going away... just trying to get it to work one last time. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3051
Location: NYC,USA
|
|
|
|
| Quote: |
| Without knowing which host variable is giving the error, I don't even know if it's a field I think is numeric. All the fields that are numeric according to the file definition that feeds this program are zeros (as shown in the buffer). |
Do you not know what you are passing to this called module? did you get to know what is the bad record from the input data set? As far as I can tell its a bad data for this record , how about comparing this record with good ones? |
|
| Back to top |
|
|
John Poulakos
Active User
Joined: 13 Jun 2012
Posts: 178
Location: United States
|
|
|
|
| The program fails on the first record. If I remove that, it fails on the next record, etc. etc. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3051
Location: NYC,USA
|
|
|
|
| Quote: |
| The program fails on the first record. If I remove that, it fails on the next record, etc. etc. |
That means the Data Set has all bad records. If this program executes in production without this issue then best way is to compare this Data Set to Production one and ask for corrections accordingly to the source system. If its a production issue then compare this Data Set to last Successful run in Production and see the differences and fix it. |
|
| Back to top |
|
|
John Poulakos
Active User
Joined: 13 Jun 2012
Posts: 178
Location: United States
|
|
|
|
If it were that simple, they wouldn't have given it to me. I only get the really ugly problems. This is an annual program that was overlooked when this legacy system was migrated to a new mainframe in January. It has never been run on this mainframe. This software was originally written for some database other than DB2. Rather than rewrite programs for DB2 they created an interface that translates database operations into DB2 SQL and performs data mapping functions. It's very complex, ugly and a good part of it is a "black box".
I only posted this topic to see if somebody could say for sure that the host variable number is in the first 2 bytes of the SQLCA message area when a DB2 -310 is issued. That would help me find things in a rather large dump, because fortunately the "black box" died on a S0C4 after completing the DB2 rollback. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3051
Location: NYC,USA
|
|
|
|
| Quote: |
| I only posted this topic to see if somebody could say for sure that the host variable number is in the first 2 bytes of the SQLCA message area when a DB2 -310 is issued. That would help me find things in a rather large dump, because fortunately the "black box" died on a S0C4 after completing the DB2 rollback. |
No. It is not as per me. You can see SQLCA copybook and when sqlcode -310 then in the Error it self it gives the position # than in SQLCA.
However, I still wonder why you can't display the host variables (used for this black box module call )before the call in the calling program to determine what is going wrong? Without this I don't think of any other way to troubleshoot the issue, since error displays that you posed are so customized and needs extra efforts to translate it and one could surely misinterpret them.
| Quote: |
| DBFUNC: INS DBNUM: 0061 DBSTATUS: *DB27 STATUS-2: 0036 |
Worth paying attention to this 61 if not then 36 position# if that's the variable having non decimal value. Did you check that already? |
|
| Back to top |
|
|
|
|
