Data Retention Period and Database Purge

[saharhanna]

                   Hello,

I have a question about HDS data retention period.

When we installed the system more than a year ago, we changed the data retention settings in the registry (Registry Key: days) for some tables to 400 because we expected that the database would be near full after this duration.

However, the HDS has only filled 30% of its size and a purge occurred in the past days that purged Agent tables and many other tables.

I am curious about the purge process: Does it happen only when the database reaches 80% regardless of the data retention period, or does it check both conditions (data retention period expiry and 80% database limit)?

Furthermore, the purge job happened in a confusing manner as it purged data between certain past dates (dates between August 2012 and Jan 2013 are lost, whereas data before August can still be found)

Isn't the purge job supposed to dump older data?

Can someone please explain?

[Gergely Szabo]

Hi,

it's by default 80%. I wonder, why it got activated at 30% - where did you see that percentage?

A usable technote explaining how it works and how it can be changed:

http://www.cisco.com/en/US/products/sw/custcosw/ps1001/products_tech_note09186a008014f8b9.shtml

G.

[saharhanna]

Hello Gergely,

Thanks for replying

I did not see that percentage anywhere. I just used to check the HDS used space and it was 30 GB, and knowing that we have a 100 GB database, I calculated the 30%

Thank you for the document, but the problem is it is not clear on the reason of the purge (80% full or data retention values in registry). According to the document, "the daily purge adjustment to percent full" which usually defines the 80% limit is only found in loggers.

Is there any way I can change or check this value for HDS?

Could there be any other reasons that could result in a purge?

[Gergely Szabo]

Hi,

you can use the "icmdba" tool to check (and expand, if necessary) the Logger, awdb and hds databases. A detailed table pops up after selecting the database, and then "Space Used Summary" from the "Data" menu.

To modify the purge settings, you may want to use ICM setup (you will have to stop the Logger instance you are working on); actually, these settings are in the registry, even at the HDS server, but not sure about how and when they are picked up, so ICM setup would be a safer option.

The reason of the purge is the percent full value, so if it goes beyond that, the system should begin deleting oldest data up to the point it reaches the purge adjustment value, regardless of the retention day settings.

G.