fix: handle nested low-level files in compaction by davidby-influx · Pull Request #26911 · influxdata/influxdb

davidby-influx · 2025-10-15T23:27:55Z

Previously, when a lower-level file (level 1, 2, or 3) was in between level 4 or higher files, it would not be compacted, and it would prevent full compactions from running.

This fix allows lower level files to be compacted even if sandwiched in between higher level files or as the first/earliest file in a shard.

Exampe problem from a customer system:

-rw-r--r--.  1 root root 2.1G Aug  5 20:59 000016684-000000007.tsm
-rw-r--r--.  1 root root 2.1G Aug  5 21:02 000016684-000000008.tsm
-rw-r--r--.  1 root root 2.1G Aug  5 21:04 000016684-000000009.tsm
-rw-r--r--.  1 root root 376M Aug  5 21:05 000016684-000000010.tsm
-rw-r--r--.  1 root root 2.1G Aug  5 18:00 000016812-000000004.tsm
-rw-r--r--.  1 root root 1.4G Aug  5 18:00 000016812-000000005.tsm
-rw-r--r--.  1 root root 1.3G Aug  5 21:21 000016844-000000002.tsm
-rw-r--r--.  1 root root 2.1G Aug  5 18:00 000016948-000000004.tsm
-rw-r--r--.  1 root root 1.4G Aug  5 18:00 000016948-000000005.tsm
-rw-r--r--.  1 root root 2.1G Aug  5 18:00 000017076-000000004.tsm

There is a rogue level 2 file packed within fully compacted files

Fixes #26794

Add `FileStore.SupportsCompactionPlanning()` which allows compaction planners to check if a `FileStore` supports compaction planning. Currently this means the `FileStore` must have a TSM filename parsing function available.

…rations

davidby-influx · 2025-10-15T23:34:02Z

@devanbenz - please walk through compact_test.go and check inputs and outputs for sanity before reading the changes in compact.go. Let me know of any concerns or additional tests you recommend.

The new tests start at the comment // Start of nested test cases???? on line 3576. Prior changes in the file are simply reformatting for readability.

gwossum

Couple of questions

tsdb/engine/tsm1/compact.go

gwossum

It looks like it will solve the nested file issue and I can't find any new issues it introduces.

devanbenz

Just the one comment about tests. Overall this looks good to me. I ran the code locally and wrote out a few TSM files to disk using inch + influxd, afterwards when I had several files I went ahead and manually renamed the levels on them to have a nested and a leading level 2 file. It was captured by the compaction without issue.

Here is the files on disk as you can see I have nested lower level files:

-rw-r--r--@  1 devan  wheel   1492013 Oct 28 15:18 000000008-000000002.tsm
-rw-r--r--@  1 devan  wheel   1484013 Oct 28 15:18 000000016-000000002.tsm
-rw-r--r--@  1 devan  wheel   1327005 Oct 28 15:18 000000024-000000002.tsm
-rw-r--r--@  1 devan  wheel    284013 Oct 28 15:18 000000025-000000001.tsm
-rw-r--r--@  1 devan  wheel    284013 Oct 28 15:18 000000026-000000001.tsm
-rw-r--r--@  1 devan  wheel    284005 Oct 28 15:18 000000027-000000004.tsm
-rw-r--r--@  1 devan  wheel    284013 Oct 28 15:20 000000028-000000005.tsm
-rw-r--r--@  1 devan  wheel  21818009 Oct 28 15:32 000000159-000000004.tsm
-rw-r--r--@  1 devan  wheel  22375789 Oct 28 15:34 000000287-000000004.tsm
-rw-r--r--@  1 devan  wheel  22450341 Oct 28 15:36 000000415-000000004.tsm
-rw-r--r--@  1 devan  wheel  22396817 Oct 28 15:41 000000543-000000004.tsm
-rw-r--r--@  1 devan  wheel   5620965 Oct 28 15:42 000000575-000000003.tsm
-rw-r--r--@  1 devan  wheel   1431685 Oct 28 15:42 000000583-000000002.tsm
-rw-r--r--@  1 devan  wheel   1424133 Oct 28 15:42 000000591-000000002.tsm
-rw-r--r--@  1 devan  wheel   1424321 Oct 28 15:42 000000599-000000002.tsm
-rw-r--r--@  1 devan  wheel     90985 Oct 28 15:42 000000600-000000001.tsm

After letting influxd sit for a bit after modifying the cold write duration (~5 minutes)

-rw-r--r--@  1 devan  wheel  93835106 Oct 28 16:12 000000543-000000005.tsm
-rw-r--r--@  1 devan  wheel  22415417 Oct 28 16:02 000000672-000000004.tsm
-rw-r--r--@  1 devan  wheel  22290681 Oct 28 16:04 000000800-000000004.tsm
-rw-r--r--@  1 devan  wheel   1424521 Oct 28 16:04 000000808-000000002.tsm
-rw-r--r--@  1 devan  wheel   1417977 Oct 28 16:04 000000816-000000002.tsm
-rw-r--r--@  1 devan  wheel   1414785 Oct 28 16:05 000000824-000000002.tsm
-rw-r--r--@  1 devan  wheel    180693 Oct 28 16:05 000000825-000000001.tsm
-rw-r--r--@  1 devan  wheel    182217 Oct 28 16:05 000000826-000000001.tsm
-rw-r--r--@  1 devan  wheel    182793 Oct 28 16:05 000000827-000000001.tsm
-rw-r--r--@  1 devan  wheel    181793 Oct 28 16:05 000000828-000000001.tsm
-rw-r--r--@  1 devan  wheel    135033 Oct 28 16:05 000000829-000000001.tsm

The lower level nested files are picked up and everything seems good.

devanbenz · 2025-10-28T18:39:05Z

tsdb/engine/tsm1/compact_test.go

+			},
+		},
+		{
+			name: "Mixed generations with 3 level 2 files",


Can you add an additional test similar to this one with 3 level 3 files?

devanbenz

LGTM

Simplify and refactor compaction planner to handle low-level files occurring before higher-level (more compacted) files. (cherry picked from commit a099cb7) Fixes #26947

Simplify and refactor compaction planner to handle low-level files occurring before higher-level (more compacted) files. (cherry picked from commit a099cb7)

Simplify and refactor compaction planner to handle low-level files occurring before higher-level (more compacted) files. (cherry picked from commit a099cb7) Co-authored-by: davidby-influx <72418212+davidby-influx@users.noreply.github.com>

davidby-influx and others added 17 commits October 10, 2025 13:55

feat: cache generation and sequence to reduce TSM filename parsing

6511b59

Fixes #26794

chore: format to reduce nugatory diffs

a9cb5cc

chore: add TSM generation and sequence parsing to tests

4f82862

fix: better errors and checks

60a7ad7

fix: use nil check version of FileStore.ParseFileName

08c9374

fix: option function ordering for FileStore

6b8629f

fix: refactor for cleanliness

d2c450d

fix: remove incorrect generation caching from DefaultPlanner.findGene…

1b2d353

…rations

fix: handle orphaned files in compaction

13c7a52

fix: test results modified for new compaction algorithm

67de4ab

fix: allow level 3 generations in Plan - WIP

b10ef5d

fix: working tests for nested low-level files

6ff5606

fix: make handling nested low-level files mandatory

e54de7c

chore: refactor for clarity

19da027

fix: RLock for findGenerations, errCompactionInProgress.Is refactor

750b7cc

chore: remove unused functions and outdated comments

21c7ba0

davidby-influx self-assigned this Oct 15, 2025

davidby-influx added area/tsm kind/bug team/edge labels Oct 15, 2025

davidby-influx requested a review from devanbenz October 15, 2025 23:31

davidby-influx changed the title ~~fix: handle nsted low-level files in compaction~~ fix: handle nested low-level files in compaction Oct 15, 2025

davidby-influx and others added 4 commits October 16, 2025 12:09

Merge branch 'master-1.x' into DSB_compaction_fixes

d9fce1a

chore: added customer shard file list to compaction planner tests

b1640e0

chore: update test case to match customer scenario

520f012

chore: fix typo

c6b4ea0

davidby-influx marked this pull request as ready for review October 21, 2025 23:33

davidby-influx requested a review from gwossum October 21, 2025 23:33

fix: added tests for edge cases

fcb0313

gwossum reviewed Oct 22, 2025

View reviewed changes

tsdb/engine/tsm1/compact.go Show resolved Hide resolved

tsdb/engine/tsm1/compact.go Outdated Show resolved Hide resolved

chore: update comment

4052c76

gwossum previously approved these changes Oct 24, 2025

View reviewed changes

devanbenz previously approved these changes Oct 28, 2025

View reviewed changes

chore: add tests per review comments

502ed65

davidby-influx dismissed stale reviews from devanbenz and gwossum via 502ed65 October 28, 2025 21:24

devanbenz approved these changes Oct 28, 2025

View reviewed changes

davidby-influx merged commit a099cb7 into master-1.x Oct 28, 2025
9 checks passed

davidby-influx deleted the DSB_compaction_fixes branch October 28, 2025 21:39

This was referenced Oct 28, 2025

handle nested lower level files in compaction [port to 1.12] #26947

Closed

handle nested lower level files in compaction [port to main-2.x] #26948

Closed

davidby-influx linked an issue Nov 26, 2025 that may be closed by this pull request

Issue with findGenerations caching depending on input #26680

Closed

This was referenced Dec 23, 2025

handle nested low-level files in compaction [port to main-2.x] #27063

Closed

OSS 2.9.0 cherry-picks #27048

Closed

davidby-influx linked an issue Mar 17, 2026 that may be closed by this pull request

TSM file state leads to inescapable loop where compactFull is not running #26681

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: handle nested low-level files in compaction#26911

fix: handle nested low-level files in compaction#26911
davidby-influx merged 24 commits intomaster-1.xfrom
DSB_compaction_fixes

davidby-influx commented Oct 15, 2025

Uh oh!

davidby-influx commented Oct 15, 2025

Uh oh!

gwossum left a comment

Uh oh!

Uh oh!

Uh oh!

gwossum left a comment

Uh oh!

devanbenz left a comment

Uh oh!

devanbenz Oct 28, 2025

Uh oh!

devanbenz left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

davidby-influx commented Oct 15, 2025

Uh oh!

davidby-influx commented Oct 15, 2025

Uh oh!

gwossum left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

gwossum left a comment

Choose a reason for hiding this comment

Uh oh!

devanbenz left a comment

Choose a reason for hiding this comment

Uh oh!

devanbenz Oct 28, 2025

Choose a reason for hiding this comment

Uh oh!

devanbenz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants