Common Reasons for INSERT Failures in GBase Databases

When an application suddenly can't insert rows, the root cause usually falls into one of a few well-known categories: space, configuration, data format, or transactions. This post lists the most frequent culprits and the commands you can use to verify them.

Main Causes and How to Check

• Dbspace Full – The most obvious blocker. Run onstat -d to see space usage. If a dbspace is exhausted, add storage or purge old data.
• Non‑fragmented Table Hit Page Limit – A single fragment can hold at most about 16,775,134 pages. When you hit that ceiling, the insert fails. Use oncheck -pt dbname:tablename to check page usage and fragment the table if needed.
• Encoding Incompatibility – Garbled characters or a mismatch between the client and server character sets will cause insert failures. Verify the locale settings on both sides.
• Lock Conflicts – If the table lock mode is set to PAGE, concurrent writers can block each other. onstat -k helps identify lock waits.
• Rolling Back Long Transactions – If a long transaction is rolling back, its inserted rows are undone, making it appear as if data was never written. Check onstat -x for transactions with an R flag and a non‑zero rb_time.

Quick Diagnostic Commands

onstat -d                     # dbspace status
oncheck -pt dbname:tablename  # page count
onstat -k                     # lock waits
onstat -x                     # transaction / rollback status

In most cases, one of these checks will expose the bottleneck. Knowing them helps you restore insert operations quickly in your gbase database environment.