fix(dursto): open databases read-only when checking out

[victor-dumitrescu]

Relates to RV-984

What

Opens RocksDB databases as read-only when checking out.

Why

Stumbled upon this while trying to diagnose an issue with the long tests (#1084) where I noticed that the size of a commit is (surprisingly) correlated to the the number of proptest cases executed. This happens because every test case starts with a checkout, which, in read-write mode, will result in a new LOG file, with the previous log file stored asLOG.old.XXXX. This balloons the size of a commit from ~1.5 MiB for 10 test cases to ~22 MiB for 256 test cases. With this fix, the initial commit drops to 0.24 MiB.

How

By using open_cf_descriptors_read_only instead of open_cf_descriptors for checking out.

An alternative fix would have been to limit the number of log files that are kept using set_keep_log_file_num. We could fall back on it if there's a good reason to keep opening DBs in read-write mode.

Manually Testing

make all

Tasks for the Author

• Link all Linear issues related to this MR using magic words (e.g. part of, relates to, closes).
• Eliminate dead code and other spurious artefacts introduced in your changes.
• Document new public functions, methods and types.
• Make sure the documentation for updated functions, methods, and types is correct.
• Add tests for bugs that have been fixed.
• Explain changes to regression test captures when applicable.
• Write commit messages in agreement with our guidelines.
• Self-review your changes to ensure they are high-quality.
• Complete all of the above before assigning this MR to reviewers.

[github-actions]

Benchmark results for revision 50c4db4:

Metric	Duration	TPS
Mean	1.469153385s	27.227
Worst	1.495357675s	26.749
Best	1.461093996s	27.377
Standard Deviation	±7.13852ms	±0.131

Full results

Run	Transfers	Duration	TPS
1	40	1.495357675s	26.749
2	40	1.469156373s	27.227
3	40	1.471735669s	27.179
4	40	1.469755949s	27.215
5	40	1.474529408s	27.127
6	40	1.470110972s	27.209
7	40	1.472418872s	27.166
8	40	1.461682638s	27.366
9	40	1.462207625s	27.356
10	40	1.469265796s	27.224
11	40	1.468918769s	27.231
12	40	1.465177772s	27.300
13	40	1.461093996s	27.377
14	40	1.463393889s	27.334
15	40	1.468697464s	27.235
16	40	1.464140267s	27.320
17	40	1.473601017s	27.144
18	40	1.468084585s	27.246
19	40	1.464606573s	27.311
20	40	1.46913239s	27.227

Compare the results above with those for the default branch.

[emturner]

I came across this yesterday too! But this time in the context of the file-handles investigation.

The diff is identical to the one I hadn't pushed yet 😅 so LGTM

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.19%. Comparing base (e8d4cce) to head (0b2db40).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1089      +/-   ##
==========================================
+ Coverage   90.16%   90.19%   +0.02%     
==========================================
  Files         130      130              
  Lines       28391    28394       +3     
  Branches    28391    28394       +3     
==========================================
+ Hits        25599    25609      +10     
+ Misses       2027     2020       -7     
  Partials      765      765              

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.