Load testing of IBProvider v5.21.1 and Firebird v3.0.7

Hello All.

We have good news for everyone who uses IBProvider in critical and complex systems.

A «large», multithreaded, load testing of the latest release of IBProvider v5.21.1 and Firebird v3.0.7 (SS, Win, x64) has been carried out.

This is more than 7 million tests that were performed on a 10-core processor.

The main point of this testing was that the global cache of the SQL query parser was used (sql_cache__scope=db).

As usual, the built-in client for Firebird (dbclient_type=fb.direct) was used.

Testing completed with no visible problems.

See additional info at our web site: https://www.ibprovider.com/eng/news/n_201225.html

Best regards, IBProvider Team.

Firebird high-level native client for Node.js / TypeScript updated to v2.1.1

Firebird high-level native client for Node.js / TypeScript updated to v2.1.1 with a few changes .

driver-native
  • Feature #62 – Change value of getDefaultLibraryFilename in MacOS.
  • Feature #63 – Make getMaster return OS error when failing to load the client library.
driver
  • Fix #69 – Subsequent use of Statement::execute* methods may send incorrect value of input parameters.

Firebird automated testing site

Alexey Kovyazin recently shared some interesting news:

Hi All,

FYI – http://firebirdtest.com
This website is a frontend of FBT + OLTP+automated scripts, supported by Pavel Zotov and some IBSurgeon employees.

It is running FBT tests on Windows server (8Gb, 4cores) for SuperServer and Classic for 3.0 and 4.0, and OLTP-EMUL for SuperServer 3.0/4.0.

Tests are triggered automatically, when new build appear at
http://web.firebirdsql.org/download/snapshot_builds/win/3.0/
http://web.firebirdsql.org/download/snapshot_builds/win/4.0/

As a result of test run, the report is generated.
Description
P – passed (ok)
Failed test due to failed assert is marked as F (failed), pink
Failed due to runtime marked as E (error), yellow.
Failed due to Firebird crash – F red/violet

For each error there is a hint (mouseover) with details of the error.

After click on the cell, the new page will be opened with more details and history of runs.

Detailed links:

  1. Total results for each kind of outcome and server mode
  2. ERRORS MAP: distribution of Firebird errors which occured during runs
  3. FRESH FAILS: tests which failed for LAST execution only
  4. FRESH ERRORS: tests which raised run-time error for LAST execution only
  5. STILL FAILS: tests which fail for at least two last runs
  6. Has fails: tests with at least one fail for previous runs
  7. Has runtime errors: tests with at least one runtime error for previous runs
  8. FRESH PASSED: successful tests with previous outcome = FAIL/ERROR/SKIPPED or absent
  9. All tests: outcomes – all results for last 30 runs
  10. All tests: A. Skipped: tests which was intentionally excluded from execution

Firebird OLTP Emul results
OLTP Emul simulates high load OLTP aplication, and uses many
sophisticated SQLs in multi-thread environment.

Our plan for the future:
1) Publish sources of automated scripts to GitHub (need to clean them a bit)
2) Perform Linux tests and include report results to the frontend
3) improve web representation – https certificate, design, etc
4) Migrate to the new FBT test framework when its adoption will be completed

We invite all interested to join the effort.
Pavel Zotov is a coordinator of the project – pz@ibase.ru.

Regards,
Alexey Kovyazin
IBSurgeon