Transaction-log version refusal skips a newest segment it cannot read
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 5/5
- Estimated time
- Over a week
- Newbie friendliness
- 35/100
Research direction
Start by reading TransactionLogStore::load() and the header probe described in the issue, then trace how activation handles unreadable newest segments. Decide whether an unreadable segment should stop the version check or remain tolerated, accounting for transient I/O errors. Done means the version refusal behavior is explicitly resolved so this build cannot append version-1 segments after an unreadable newer-format segment, or the limitation is documented.
Written by the indexing model from the issue text.
Description
Problem
From release N on (rocksdb-js#967), TransactionLogStore::load() refuses to open a store whose newest segments include one written by a newer release. The check is a header probe that walks down from the newest segment. A segment it cannot read (permissions, an I/O error) is treated as unversioned, like a malformed one, and the walk moves on.
So suppose the newest segment is a newer-format segment this build cannot read, for example one created 0600 by a newer release running as a different user. The probe skips it, activation's open fails and is tolerated as an unreadable newest segment, and this build starts appending version-1 segments after it. That is the outcome the refusal exists to prevent.
Options
- Treat an unreadable newest segment as fatal when a probe cannot determine its version. That changes today's tolerance for unreadable newest segments, which also covers transient errors.
- Leave it, documented: the version cannot be known without reading the file, and the trigger needs a newer release writing under another identity.
Review-Deferral: 9827cc9d497ceb78 severity=minor category=data-integrity
— Claude Opus 5.5 (dev-agent)
- Dominant language
- C++
- Stars
- 24
- Forks
- 3
- Avg merge
- 1d 5h
- Merged PRs (30d)
- 53
Getting set up
- No Dockerfile or Docker Compose file
- No pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from HarperFast/rocksdb-js
-
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
HarperFast/rocksdb-js#964 ·
Maintainers usually reply within 1 day
-
Transaction committedPosition is the watermark key, not the batch's first-entry addressPossibly taken @kriszyp claimed this 2 days ago. Open
Difficulty 1/5 Under an hour Newbie friendliness 85/100
HarperFast/rocksdb-js#961 · 1 assignee ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
HarperFast/rocksdb-js#945 ·
Maintainers usually reply within 1 day
-
Difficulty 4/5 3-5 days Newbie friendliness 45/100
HarperFast/rocksdb-js#972 ·
Maintainers usually reply within 1 day
-
Difficulty 4/5 3-5 days Newbie friendliness 55/100
HarperFast/rocksdb-js#963 ·
Maintainers usually reply within 1 day
All issues in HarperFast/rocksdb-js
Similar issues
-
bug intrinsic medium priority rt_lib
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
OpenPrinting/splix#33 · 3 comments ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
sumatrapdfreader/sumatrapdf#6417 · 3 comments ·
-
[BUG] Cython cache evicts most recent instead of oldest entry on cleanupPossibly taken @scoder claimed this today. OpenBuild System defect
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
Maintainers usually reply within 2 days
-
Difficulty 2/5 1-3 hours Newbie friendliness 75/100