Relevance DB Occupying too Much Space

Experience Manager (PaaS/OnPrem)
1

Hi All,

Currently, we are using Bloomreach 12.6.9 version.
Our Relevance DB is occupying too much space. It has stored almost 22M rows within the span of 6-8 months.

Number of records in request log has reached 12.5M rows and occupied 26GB space and Visitors reached 9M rows and occupied 15GB.

Is there a way to delete/purge the data which is older than 6 months ?

Thanks,
Yeshwanth.

2

Hi Yeshwanth,

Wow that’s a lot of data. The pages you are looking for are

The stores can be configured to truncate stale data. Set the property maxAgeDays to 180 so that all data older than approximately 6 months is deleted.

You might want to upgrade to Bloomreach 13.4 so that you can configure the property cleanupJobCronTrigger to specify when the truncation jobs are executed.

3

Thank you @michiel_eggermont

4

Hi,

We’re facing the same space issue and we are trying to configure the requestlog data store as specified in documentation.

However, the cleanup cronjob doesn’t seem to be working. This is how our /targeting:targeting/targeting:datastores/targeting:requestlog is configured (cron expression set to run every 2 minutes for testing):

/targeting:requestlog:
  jcr:primaryType: targeting:datastore
  cleanupJobCronTrigger: '*/2 * * * *'
  dataSource: jdbc/targetingDS
  maxAgeDays: 35
  targeting:storefactoryclass: com.onehippo.cms.targeting.storage.sql.DelegatingSqlStoreFactory

When we check our targetting DB, the table requestlog still has same number of records.

Are we missing anything? We’re using Bloomreach 14.5.0.

Thanks,
Jaime

5

Hello jamie,

Post configuring the requestlog datastore in bloomreach. we have followed below steps to optimize the tables.

  1. Login to DB
  2. Capture results
  3. Optimize table requestlog;
  4. capture the results again and compare the results with step2. (DB size and row count should be reduced)

Thanks,
Yeshwanth.

6

Hi @Yeshwanth_Lagala ,

Thanks for your quick reply on an old thread :slight_smile:

But does it mean we need additional steps to cleanup the table? I would expect this to be managed by Bloomreach itself via mentioned configuration.

Does configuring requestlog datastore have any effect then? It doesn’t even seem like the job is triggered as specified by our Cron expression.

Regards,
Jaime