mdb database size is growing a lot every day
Article ID: 383980
Updated On:
Products
Issue/Introduction
Size of mdb database of a Domain Manager is growing a lot every day (Some GB on average per day).
Why is it growing continuously ?
Environment
Client Automation - All Versions
Cause
A possible cause is if Domain Manager is linked to an Enterprise Server and there are some problems with Upload Replication Job executed by the Engine.
Execute following SQL Query to check the number of rows in table ca_replication_history :
sp_spaceused ca_replication_history
Example :
In below example table ca_replication_history contains 1.9 billion of rows
name |rows |reserved |data |index_size |unused
-----------------------|--------------------|------------------|------------------|------------------|------------------
ca_replication_history |1925401077 |747986656 KB |240778512 KB |507052784 KB |155360 KB
This table is used by Upload Replication job to replicate the deletion actions to Enterprise database.
When a row is replicated it is deleted in table ca_replication_history
Due to huge number of rows, Upload Replication job could take several hours or days to process all the rows and replication does not work.
Resolution
- Check the Engine logs to see if Upload replication job is working fine (ex: C:\Program Files (x86)\CA\DSM\logs\engine*.log)
- If Upload Replication has problems to process the deletions operation, check the number of rows in table ca_replication_history
If this table contains several millions or rows, it could be emptied with following SQL Query :
TRUNCATE TABLE ca_replication_history - If the table ca_replication_history is filled with huge number of rows every day, check if amagent is executed on the Agents with option /COLLECT.
Ex:
caf start amagent args /COLLECT
Avoid to execute amagent with option /COLLECT everyday as it could generate huge number of rows in ca_replication_history.
/COLLECT option could be used on the agents once a week for example.
Feedback
