Category: Databases
-
Unified Audit Trail: SQL Text Not Recorded in Syslog
The Oracle Database 12.2 or later uses Unified Auditing to track activity. However, certain fields are not captured in the syslog of Unified Audit Trail, impacting the completeness of log data. To address this, a simple method of sharing audit records with SIEM tools like WAZUH involves using SQL/JSON functions to stream data in NDJSON format. This can be done using a shell script scheduled with cron.
-
Fuzzy Matching in PostgreSQL: A Guide to pg_trgm extension
PostgreSQL’s pg_trgm extension facilitates efficient similarity searches using trigram matching, breaking text into trigrams for pattern matching and enabling the use of distance operators. It supports fuzzy matching and indexing for speedier searches, allowing for functions like similarity() and operator usage.
-
How to Remove Orphan Large Objects in PostgreSQL
To manage Binary Large Objects (BLOBS) in PostgreSQL, use the lo_unlink() function to remove them properly. Failing to do so can leave orphaned records, but vacuumlo can help clean them up.
-
Comparing Oracle Users’ Decrypted Passwords with RockYou2024
The “RockYou2024” password leak contains 9,948,575,739 unique plaintext passwords, making it valuable for cybercriminals aiming to perform brute-force attacks. Ensuring the security of Oracle databases is crucial, as weak passwords pose a significant vulnerability. To mitigate this risk, it is recommended to compare database passwords against the 1 million most common ones. A provided SQL script guides this process and demonstrates how to perform the comparison.
-
pg_amcheck: PostgreSQL corruption detection tool
PostgreSQL’s pg_amcheck is an essential tool for ensuring data integrity. Like Sherlock Holmes, it diligently detects anomalies, offering early detection of data corruption from hardware failures, software bugs, or system crashes. This tool, though not glamorous, is vital for proactive maintenance. Administrators can use pg_amcheck to identify issues.
-
Unusable and Parallel indexes on Flashback Data Archive Tables in 19.22
After updating to release 19.22, issues arose with Flashback Data Archive (FDA). Despite bug fixes(35957640), internal indexes were created with the DEFAULT PARALLEL degree. Additionally, split operations caused related indexes to become unusable. A debug of the FBDA process revealed that the split operations lack an “UPDATE INDEXES” clause and no index rebuild operation is done internally. The test case provided highlights the issues. These concerns have been raised in a Service Request for resolution.
Osman’s DBlog 