Troubleshooting
Problem
Using IBM InfoSphere Guardium product, you notice the database disk space is growing at a fast rate. Upon investigation, you find out GDM_ERROR table is increasing rapidly mostly with PARSER_ERROR errors.
Symptom
- Database disk space growing fast
- GDM_ERROR table space increasing rapidly with errors of type PARSER_ERROR
Cause
Sniffer not able to handle some types of SQLs (usually a sniffer bug). This failure to handle the SQLs occurs frequently enough that GDM_ERROR table grows at a fast rate which then translates to database disk space usage growth.
Diagnosing The Problem
The problem referenced in this article is occurring if both of the following are true:
- Database space growing fast. Use command support show db-status used % to show database used space.
- GDM_ERROR used space increasing at a fast rate. Use command support show db-top-tables all to show the biggest tables. GDM_ERROR will be the biggest, or one of the biggest tables.
Resolving The Problem
To resolve the problem:
1.- Install the latest sniffer patch from FixCentral
Note: Latest sniffer patch contains the latest fixes to identified defects related to sniffer inability to handle certain SQLs which leads to this problem.
2.- Run below cli commands to enable logging of parser errors and restart the inspection core to apply the change:
store log sql parser_errors
restart inspection-core
Note: You can wait few days after installing the latest sniffer patch before deciding to run this step and only perform this step if GDM_ERROR continues growing rapidly with the last sniffer patch. Alternatively, you can perform this step on any case.
3.- Optional: cleanup GDM_ERROR from current data with one of below options:
a) immediate cleanup. Use command support clean DAM_data parser_errors to perform the cleanup
b) wait for normal purge to take care of the purging (you'll start to see decrease in disk space usage when the GDM_ERROR purge period has passed).
Note that default purge period for GDM_ERROR is 30 days, but you can reduce it if you want the purge to occur sooner . Use commands show purge objects age/store purge objects age to list/change purge age for GDM_ERROR.
4.- Run an Optimize to optimize the TURBINE database. This will reorganize the data in all tables, including GDM_ERROR, which will result in reflecting the current actual reduced size. To run the Optimize, use the diag cli command and select options
- 4 Perform Maintenance Actions
2 TURBINE optimize
Was this topic helpful?
Document Information
Modified date:
16 June 2018
UID
swg21700128