Abort Matchning

[Former Member]

Hi All,

We are experiencing performance issues in our MDM database where we have ~500k records. If we match 1 record vs all, this operation takes ~40 minutes to even get the pop-up saying that the matching has started. That means that we cannot even press the "Cancel" button. Since this operation runs on the MDM server it doesn't help with closing MDM Data Manager. If we would run a "all vs all" operation I figure it would take several days to even get the cancel button.

My question is: How do I abort this operation without having to restart the repository?

Regards,

Christian

[Former Member]

Hi Christian,

Looking into your problem, I would advice you some workaround.

Since you are Matching 1 vs All, the one record is getting matched with all the records present in your repository i.e 500k records, and that is huge enough to drag down the performance of the system.

Hence, I would advice you to use the other matching types present there:

1. Selected vs Selected - Selected records will be matched against each other.

2. Select vs result - Matches the selected records with the record set derived from the search criteria.

3. Results vs Results - Matches all records that result from the search criteria.

Avoid Selected Vs all and Results Vs All as All in your case implies to 500K records, which will lower the performance.

You can run a search query and get the records as a result or else, select some 100 recrds ( just for testing) and use Selected vs Selected.

Try to use, above adviced 3 ways.

I hope this helps you.

Thanks and Regards

Nitin Jain

[Former Member]

Our plan is to avoid using those "vs All"-commands, but most likely someone will mistakenly use one of these commands resulting in a 2h delay in the work.

I still would like to know if there is any way of aborting one of these command or disabling the use of those commands (i.e. removing access to the buttons)?

[Former Member]

Hi,

We are still facing this problem when users by mistake make an all vs all matching. The systems locks up for a few hours, resulting in us having to restart the server.

Has anyone faced the same problems I would be grateful if you could give me an suggestions on how to avoid similar mistakes in the future.

Kind Regards,

Christian

[Former Member]

Hi Christian,

maybe a workaround would be to guide users using workflows. You can than specify in the matching step to match only workflow records against each other. The users will than only see the result but do not have to launch matching, which would avoid accidental execution against all records.

Best regards

Christian