Saving rule-set takes too long

  • 0
  • 1
  • Question
  • Updated 2 years ago
  • Answered

Hello,

    I have problem with saving some rulesets.

The first one is made up of 2910 tables and the second one is made up of 220 tables.

The refresh/save icon shows hourglass saying that the activity is still in progress after some days have spent since it was started.

Luigi

Photo of luigidep

luigidep

  • 622 Points 500 badge 2x thumb

Posted 2 years ago

  • 0
  • 1
Photo of Mouhssine SAIDI

Mouhssine SAIDI

  • 4,782 Points 4k badge 2x thumb
Hi,

Can you look at the inventory tab if you have any table(s) listed out for each rule set. (may be it's a navigator just a refresh).

Regards,

Mouhssine 
Photo of luigidep

luigidep

  • 622 Points 500 badge 2x thumb

Hi,

   the inventory actually store the objects related to the rulesets and i can list them and view the detail structure since the ruleset view still show hourglass at the refresh/save column but now only for one of the two


Could be best practice to obtain different disjoint rulesets from a single large set (oracle schema) of tables? In this way we could lighten the workload involve in provisioning and masking.

Luigi

Photo of Mouhssine SAIDI

Mouhssine SAIDI

  • 4,782 Points 4k badge 2x thumb
Hi,

The last paragraph in your replay isn't very clear for me, would you to please re explain it.

Regards,
Mouhssine
Photo of luigidep

luigidep

  • 622 Points 500 badge 2x thumb

I'll try to re-explain.

If we have an oracle schema with 2000 tables we can create two new oracle schemas. Then

we grant select privileges on the first 1000 tables to one schema and we grant select privileges on the last 1000 tables to the second new-created  schema.

Then we create two separate rule-set one the first new schema  and the other on the second schema.  And so we split the original oracle schema in two disjoint rulesets. The profiling and masking task should result be lightened.

Photo of Mouhssine SAIDI

Mouhssine SAIDI

  • 4,782 Points 4k badge 2x thumb
Hi,

If you are just splitting things just to lighten profiling and masking jobs, let me tell you that you haven't to.

You can play with streams and parallel parameters and also the jvm xms and xmx params to tune the java heap for masking job.

So you can just grant all tables to the same schema and create one ruleset and later in play with kind of tuning profiling and masking jobs.

Regards,

Mouhssine
Photo of luigidep

luigidep

  • 622 Points 500 badge 2x thumb

Thank you

    very much.


Luigi

Photo of luigidep

luigidep

  • 622 Points 500 badge 2x thumb

Excuse me.

I have another question.

I have a rule set that i have profiled with column level expressions.

Then I want to apply data level expressions profiling to the same set of tables that we know should be greedy of computing resources.

I would not use the same rule set because it is already associated with the inventory that hold column level profiling results. So I create a new ruleset with the same set of tables as the preceeding one and then i'll apply the datalevel profiling to it.

Given these premises i ask you how can i merge the profiling results that now are split between two inventories? Is there a way to export the inventories merge them and then import the resulting inventory?

I'll keep separate profiling job for data level versus column level expressions because the latter always ends successfully while the former often fails or remains hanging.

Luigi