InaPlex Forums

Using inaport 6.0 build 7280 and Goldmine 8.5 Premium.

2 Inaport profiles are being used - one to delete data from contsupp (a detail tab) and the other to update with new data. Profile is run from a batch file scheduled overnight.

This has been working fine, but recently has developed an issue where the delete profile fails with various exception handles. I've enabled logging and have verified that that the recid exists (it fails on different recids). Have also reindexed Goldmine and the problem still occurs. When I run the profile but reduce the number to process (to for example 10 records), then the profile runs. I'm all out of ideas. I've posted the log where it fails below, any help would be greatly appreciated!

687415    indexes.TableIndex:  match(rec):  table = _Details  using sql: select recid from contsupp
where recid='EC5NWT0!^SUL92S'
687415    goldmine.GMPhysDB:  sql query: sql = select recid from contsupp
where recid='EC5NWT0!^SUL92S'
687415    goldmine.GMPhysDB:  query: number of rows returned = 1
687415    db.DBLogTable:  match: table = _Details  ParentKey = NULL  Num matches found = 1  matchVal found = EC5NWT0!^SUL92S
687415    goldmine.GM:  SeekExact: handle = 86581456  seekVal = EC5NWT0!^SUL92S  result = true  seekCode = 1
687431    exceptions.MVException: 
========
MVException: ======
 msg = Fail deleting row from GoldMine table with handle 86581456

 stack = com.inaplex.inaport.exceptions.MVException 

Hi Dan

Almost invariably this type of problem is a database or GoldMine index issue of some kind. Unfortunately, the GoldMine API does not provide any information about why an operation fails.

Things to check:

1. As you did, reindex the database - perhaps try again just before a run

2. Are there any triggers on any of the tables?

3. Any chance there are multiple inaport jobs running at the same time?

4. Anything else hitting the GoldMine database?

If all else fails, can you run a SQL trace and send it to support?

Regards

Admin

Thanks for getting back to me. In reply to your questions:

 1) Tried reindexing again, but this didn't help.

2) There is a trigger on the conthist table but not on anything inaport is using.

3) I have a batch file that runs a profile every 2 hours (loading a detail tab to display current back orders). Could this be causing the issue?

4) We have GISME running on the Goldmine database and also it will be interrogated by wMobile.

 When I run a profile it spends a long time with the status "Building indexes" - is this normal? There are now over a million records in the contsupp table.

Re a SQL trace - using SQL Server Profiler would it be just the SQL statements you need to see?

Hi Dan

When I wrote it, GISME had a trigger on CONTSUPP (to detect changes to email and web addresses) , and it would not surprise me if wMobile does as well.

If you batch file is running at the same time as the Inaport delete job, I could imagine this causing an issue.

With over a million records in the contsupp, I would *strongly* recommend moving to using SQL matching. This will dramtically improve performance, and may cause the issue to go away. See here for a tutorial on SQL matching:
http://www.inaplex.com/Resources/Tutorials.aspx

Regards

Admin