cancel
Showing results for 
Search instead for 
Did you mean: 

PI DB growing faster

former_member601407
Participant
0 Kudos

Hi,

Our team BASIS consultant complained about rapid increase in production PI DB in the last 2-3 weeks. I checked in Message monitoring section, there wasn't any unwanted interface running and all interfaces are functioning as usual. I did a change in persistence of audit logs like, the audit log was not showing for messages that are older than 1 day, where i changed the parameter "messaging.auditLog.memoryCache = false". But this change was did 2 months before.

Could you please guide me on this. What are the PI tables that are to be monitored when this kind of scenario occurs. And what precautionary steps has to be taken to avoid rapid DB growth.

PI version: PI 7.5 Single Stack SP 10

jens.schwendemann ryan.crosby

JaySchwendemann
Active Contributor
0 Kudos

I don't think it is wise to particularly "mention" people within your OP just to gather additional attention of those people. You may easily pick the wrong people (see Ryan's answer that he is not a basis guy) which may (and will ultimately) put people off

Accepted Solutions (0)

Answers (2)

Answers (2)

Ryan-Crosby
Active Contributor
0 Kudos

Hi Subin,

I am a developer so I cannot speak to items that fall within the BASIS realm. System and application configuration related to the database and persistence within the database should be understood well and handled by that team.

Regards,

Ryan Crosby

JaySchwendemann
Active Contributor

I second that. Basis should know which tables have increased and based on that know (ideally) or google (realistically) where the root cause may be. If Basis is not willing to do the latter, they should at least tell you the increasing table names (or tell you the file system directories if that is where the increasing happens)

former_member601407
Participant
0 Kudos

i consulted with BASIS consultant and these are the table names which was increasing rapidly.

BC_MSG_LOG - synchronous message logs and retain time configured in xpi: messaging system is 7days (messaging.log.retainTime)

BC_MSG_VERSION - async messages staging payloads. xiadapter.stage.conf : "MS=3"

BC_MSG - asynchronous messages. “xiadapter.outbound.persistDuration.default” and “xiadapter.inbound.persistDuration.default” set to 3 days.

My doubt is, in above table BC_MSG config, persistDuration is set to 3 days. But i can see the audit log and message payload getting retained for more than 20 days (async & sync). Then what is the use of having BC_MSG table if we have 20 days retention period set.

Now if i change from 3 days to default 1 day, then the audit logs and message payloads will get affected in PI ? or just the table data will be cleared after 1 day?

Ryan-Crosby
Active Contributor
0 Kudos

Hi Subin,

We have our xiadapter.outbound.persistDuration.default and (inbound too) set for 14 days and we have no issues. One important distinction here would be to confirm that the configuration hasn't been altered to also persist synchronous messages which are not persisted by default. For example, in our system we handle a little over a million messages a week, but synchronous messages account for over a million of those messages.

Regards,

Ryan Crosby

former_member601407
Participant
0 Kudos

Hi Ryan,

In 2 days, another 4GB has been used mostly by above 3 tables. So u r saying that to check on sync messgaes config.. which is this one... messaging.log.retainTime = 7 days (this is default)..

I think this might be because of audit log persistence which I changed 2 months back...messaging.auditLog.memoryCache = false. these audit log persistence gets stored under which table ?

If one of the three tables mentioned then I have to revert the changes.

Ryan-Crosby
Active Contributor
0 Kudos

Hi Subin,

I am not familiar with the parameter but it sounds like you may have disabled the use of memory cache for storage (which is what we have) and possibly diverted the information into the database on account of what you are experiencing. Your BASIS team should be able to confirm or find out what the purpose of that parameter is for and determine if that is what is causing the DB growth.

Regards,

Ryan Crosby

Sriram2009
Active Contributor
0 Kudos

Hi Subin.

You may have to define the policy to keep the database of 3 to 4 weeks time remaining all the data can be archived or delete from the DB, so that you can manage the disk space. refer the sap note 2592310 and follow the SAP wiki link https://wiki.scn.sap.com/wiki/display/XI/PI+database+size+is+growing+large+SXMSCLUP%2C+SXMSCLUR

Regards

SS