feat(proxy): Oracle session TNS dump capture#78
Merged
Conversation
- Multi-row: rows now captured correctly (0x07 separator between rows, 0x08 footer detection) - Multi-column: limit column name scanning to area before 0x06 0x22 marker, eliminating false positive column names from row data - DATE type: decode Oracle DATE (7 bytes) to YYYY-MM-DD HH:MM:SS - Single-column multi-row: fix premature pendingQuery completion by not completing queries from legacy Response handler (0x08) which fires during auth negotiation - NULL values: properly shown as empty string (0x00 length byte) - Column names: require minimum 2 chars to avoid false positives Tested with 7 query patterns against Oracle 19c on abynonprod: all produce correct column names, multiple rows, and decoded values. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The connect data offset in TNS Connect packets is from the start of the full packet (including 8-byte header), not from the payload start. Subtract tnsHeaderSize when indexing into the payload. This fixes DBeaver/JDBC thick driver connections returning 'database not found' because SERVICE_NAME was read from the wrong offset. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Try both interpretations of the connect data offset: 1. Raw offset (works for v315+ thin client with extended data) 2. Offset minus header size (works for JDBC thick with inline data) Validate by checking if the extracted string contains DESCRIPTION or SERVICE_NAME keywords. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace fixed-offset SQL extraction with scanning approach:
1. Scan offsets 40-70 for length-prefixed SQL strings
2. Fallback: find SQL keywords (SELECT, INSERT, etc.) directly in payload
3. Extract SQL text until end of printable ASCII
Fixes JDBC thin driver SQL appearing as 'ELECT' (offset 54 vs 50),
and supports go-ora which uses yet another offset.
Tested with all 3 client types through local proxy:
- Python oracledb thin: SELECT 42 → {ANSWER: 42}
- JDBC thin (ojdbc11): SELECT 1 → logged correctly
- Go go-ora: SELECT 42 → logged correctly
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Stream rows directly to database as they arrive (no memory buffering) - Create query record eagerly when SQL is intercepted - Capture rows from continuation packets (func=0x06) in multi-packet results - Detect end-of-data (ORA-01403) across any response packet - Robust SQL extraction scanning offsets 40-70 + keyword fallback - Update query record with duration on completion Tested: 117-row result set captures 49 rows (first+continuation packets). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Binary packet capture per session for offline replay and TTC parser development. Includes dump writer/reader, REST API for download, Python replay script, configurable retention, and max size cap. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
CLAUDE.md: update description, add Oracle proxy to stack/structure, add DBB_LISTEN_ORA env var, update connection flow diagram. docs/oracle.md: comprehensive TNS/TTC protocol documentation covering v315+ packet format (4-byte length), TTC function codes, piggyback sub-operations, SQL extraction offsets per client, result row format, Oracle NUMBER/DATE encoding, multi-packet results, and known limitations. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Binary dump format captures raw TNS packet traffic for offline replay and analysis. Includes DumpWriter, DumpReader, and CleanupOldDumps. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
New env vars: DBB_ORACLE_DUMP_DIR, DBB_ORACLE_DUMP_MAX_SIZE, DBB_ORACLE_DUMP_RETENTION. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Captures raw TNS packets in both directions during relay. Dump writer is initialized after upstream connection, closed on session cleanup. Periodic cleanup goroutine removes old dumps based on retention config. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
GET /connections/:uid/dump - download raw TNS dump (admin/viewer) DELETE /connections/:uid/dump - delete dump file (admin only) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Python script for analyzing .dbbat-dump files. Updated CLAUDE.md with DBB_ORACLE_DUMP_* environment variable documentation. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Check error returns on f.Close() and os.Remove()
- Add t.Parallel() to all test functions
- Use assert.GreaterOrEqual instead of assert.True for comparisons
- Use defer func() { _ = r.Close() }() pattern
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
DumpWriter/DumpReaderininternal/proxy/oracle/dump.gowith full round-trip testsDBB_ORACLE_DUMP_DIR,DBB_ORACLE_DUMP_MAX_SIZE,DBB_ORACLE_DUMP_RETENTIONGET /connections/:uid/dump(download) andDELETE /connections/:uid/dump(admin delete)scripts/replay_dump.py) for offline analysisTest plan
go vetcleanDBB_ORACLE_DUMP_DIRset, connect Oracle client, verify dump files created🤖 Generated with Claude Code