MOD-8097, MOD-8114 Fix memory counting by meiravgri · Pull Request #5204 · RediSearch/RediSearch

meiravgri · 2024-11-12T16:07:46Z

This PR fixes bugs in statistics updates during gc.

wrong update of `range->invertedIndexSize` when returns empty from the fork process

In the previous implementation, when the inverted index of a numeric range returned with no docs from the fork process, range->invertedIndexSize stat was set to 0.
This is wrong since an empty range consumes memory of inverted index metadata struct size + block struct size + initial block capacity size.
In this PR range->invertedIndexSize is decreased by nbytesCollected and increased by nbytesAdded.

In addition, once the gc parent finishes iterating all the nodes received from the fork, it starts cleaning empty nodes.
NumericRangeTree_TrimEmptyLeaves returns the amount of memory released due to empty nodes removal.
This is where we update the stats of a completely removed range (inverted index).

fix overflow in spec->stats.invertedSize during `FGC_updateStats`

before the fix:

sctx->spec->stats.invertedSize += bytesAdded - bytesCollected;

Since all the variables are of type size_t, and bytesAdded might be smaller than bytesCollected, sctx->spec->stats.invertedSize might get overflowed.

the fix is to split this into two lines.

sctx->spec->stats.invertedSize += bytesAdded;
sctx->spec->stats.invertedSize -= bytesCollected;

Fix bytes collected by the gc statistics

Currently, the number of bytes collected by the gc inclused bytesAdded, but it doesn’t include bytesAdded during the cleaning process.

This might result in counting bytes added by the parent in bytesCollected, when releasing an entire inverted index.

After the fix, we will include also bytesAdded in the gc stats

  gc->stats.totalCollected += bytesCollected;
  gc->stats.totalCollected -= bytesAdded;

Additional changes in this PR

Add total invertedIndexSize to the numeric tree struct
Since opening an index in WRITE mode, might result in creating it, debug commands open inverted index for READ. If the invertedindex was not initialized yet, the result will be in the format of an empty index.
add numeric tree struct overhead to the memory count in IndexSpec_TotalMemUsage

Additional bug fixes

Fix unsafe read of a numeric tree without holding a valid reference to its spec in the GC thread - MOD-8114

replaces all OpenNumericIndex calls.

rename createIndex to createPopulateTermsInvIndex add function to tests/index_utils generalize FGCTest suit add inheriting suite for tags some cosmetics in cpp_range ExtTest.testDynamicLoading: log when assert fails (the test halts upon assertion and the log will be nerver pronted) in IndexTest and QueryTest use IndexSpec_RemoveFromGlobals instead of StrongRef_Release to ensure the spec is properly cleaned otherwise some structs like the global prefix trie nodes, potentially accessed by following tests, might hold a copy of the spec refrence that was already released and cuase a crash.

update in gc and in addition add cpp_range test and cpp_forkgc test

decrease gc->stats.totalCollected by bytesAdded to not count bytes added by us remove reseting currNode->range->invertedIndexSize in applyNumIdx called upon each node recieved from the child process instead, we will count the memory released when trimming the empty leaves call IndexSpec_TotalMemUsage from RediSearch_MemUsage instead of code duplication add IndexSpec_collect_numeric_overhead that counts the memory of NumericRangeTree in a given spec call this function to add it to IndexSpec_TotalMemUsage update LLApiTest testInfo* accordingly

codecov · 2024-11-12T17:27:32Z

Codecov Report

Attention: Patch coverage is 86.20690% with 8 lines in your changes missing coverage. Please review.

Project coverage is 86.59%. Comparing base (613bc67) to head (5a3e205).
Report is 4 commits behind head on master.

Files with missing lines	Patch %	Lines
src/fork_gc.c	72.72%	6 Missing ⚠️
src/debug_commands.c	88.88%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5204      +/-   ##
==========================================
- Coverage   86.59%   86.59%   -0.01%     
==========================================
  Files         193      193              
  Lines       34769    34767       -2     
==========================================
- Hits        30109    30107       -2     
  Misses       4660     4660

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

🚨 Try these New Features:

Flaky Tests Detection - Detect and resolve failed and flaky tests

…king_fixes

… if it was an empty tree

it will be counted when the range is deleted

don't intialize the numeric inverted index in the child if it's empty

src/debug_commands.c

src/fork_gc.c

src/spec.c

tests/cpptests/index_utils.cpp

fix IndexSpec_collect_numeric_overhead and add test fix sending NULL to FGC_sendFixed

GuyAv46 · 2024-11-21T05:44:35Z

tests/pytests/common.py

    env.assertEqual(db_res, res, message=message)
+
+# fields should be in capital letters
+def getInvertedIndexInitialSize_MB(fields: list[str]) -> float:


No need to fix it here as well?

redisearch-backport-pull-request · 2024-11-21T07:39:20Z

Backport failed for 2.8, because it was unable to cherry-pick the commit(s).

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin 2.8
git worktree add -d .worktree/backport-5204-to-2.8 origin/2.8
cd .worktree/backport-5204-to-2.8
git switch --create backport-5204-to-2.8
git cherry-pick -x 357d356db01fac23d522344bfb0855cb8a38f4d2

redisearch-backport-pull-request · 2024-11-21T07:39:21Z

Backport failed for 2.6, because it was unable to cherry-pick the commit(s).

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin 2.6
git worktree add -d .worktree/backport-5204-to-2.6 origin/2.6
cd .worktree/backport-5204-to-2.6
git switch --create backport-5204-to-2.6
git cherry-pick -x 357d356db01fac23d522344bfb0855cb8a38f4d2

redisearch-backport-pull-request · 2024-11-21T07:39:22Z

Backport failed for 2.10, because it was unable to cherry-pick the commit(s).

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin 2.10
git worktree add -d .worktree/backport-5204-to-2.10 origin/2.10
cd .worktree/backport-5204-to-2.10
git switch --create backport-5204-to-2.10
git cherry-pick -x 357d356db01fac23d522344bfb0855cb8a38f4d2

redisearch-backport-pull-request · 2024-11-21T07:39:23Z

Backport failed for 8.0, because it was unable to cherry-pick the commit(s).

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin 8.0
git worktree add -d .worktree/backport-5204-to-8.0 origin/8.0
cd .worktree/backport-5204-to-8.0
git switch --create backport-5204-to-8.0
git cherry-pick -x 357d356db01fac23d522344bfb0855cb8a38f4d2

* expose openNumericKeysDict, requires a write/read argument. replaces all OpenNumericIndex calls. * fork gc -initialiize index pointer only once rename createIndex to createPopulateTermsInvIndex add function to tests/index_utils generalize FGCTest suit add inheriting suite for tags some cosmetics in cpp_range ExtTest.testDynamicLoading: log when assert fails (the test halts upon assertion and the log will be nerver pronted) in IndexTest and QueryTest use IndexSpec_RemoveFromGlobals instead of StrongRef_Release to ensure the spec is properly cleaned otherwise some structs like the global prefix trie nodes, potentially accessed by following tests, might hold a copy of the spec refrence that was already released and cuase a crash. * add inverted index total size to the numeric tree root update in gc and in addition add cpp_range test and cpp_forkgc test * FIX: decrease gc->stats.totalCollected by bytesAdded to not count bytes added by us remove reseting currNode->range->invertedIndexSize in applyNumIdx called upon each node recieved from the child process instead, we will count the memory released when trimming the empty leaves call IndexSpec_TotalMemUsage from RediSearch_MemUsage instead of code duplication add IndexSpec_collect_numeric_overhead that counts the memory of NumericRangeTree in a given spec call this function to add it to IndexSpec_TotalMemUsage update LLApiTest testInfo* accordingly * fixes * fix new cpp forkgc test * fix unsafe read * fix status check * numeric tree debug commands - if tree was not intizlied yet, reply as if it was an empty tree * fix reset range->invertedIndexSize in applyNumIdx when index is empty. it will be counted when the range is deleted * add initial inverted index size if numeric to test don't intialize the numeric inverted index in the child if it's empty * fix gc warning * little fix * changes rt->invertedIndexSize to rt->invertedIndexesSize fix IndexSpec_collect_numeric_overhead and add test fix sending NULL to FGC_sendFixed * CalculateNumericInvertedIndexMemory fix identation and break when failing * rename openNumericKeysDict variable write to initialize and change type to bool * change OPEN_INDEX_READ to DONT_CREATE_INDEX and OPEN_INDEX_WRITE to CREATE_INDEX add coments to tests * use Likst instead of list to support older python versopns * remove type from function args * fix --------- Co-authored-by: GuyAv46 <[email protected]> (cherry picked from commit 357d356)

* MOD-8097, MOD-8114 Fix memory counting (#5204) * expose openNumericKeysDict, requires a write/read argument. replaces all OpenNumericIndex calls. * fork gc -initialiize index pointer only once rename createIndex to createPopulateTermsInvIndex add function to tests/index_utils generalize FGCTest suit add inheriting suite for tags some cosmetics in cpp_range ExtTest.testDynamicLoading: log when assert fails (the test halts upon assertion and the log will be nerver pronted) in IndexTest and QueryTest use IndexSpec_RemoveFromGlobals instead of StrongRef_Release to ensure the spec is properly cleaned otherwise some structs like the global prefix trie nodes, potentially accessed by following tests, might hold a copy of the spec refrence that was already released and cuase a crash. * add inverted index total size to the numeric tree root update in gc and in addition add cpp_range test and cpp_forkgc test * FIX: decrease gc->stats.totalCollected by bytesAdded to not count bytes added by us remove reseting currNode->range->invertedIndexSize in applyNumIdx called upon each node recieved from the child process instead, we will count the memory released when trimming the empty leaves call IndexSpec_TotalMemUsage from RediSearch_MemUsage instead of code duplication add IndexSpec_collect_numeric_overhead that counts the memory of NumericRangeTree in a given spec call this function to add it to IndexSpec_TotalMemUsage update LLApiTest testInfo* accordingly * fixes * fix new cpp forkgc test * fix unsafe read * fix status check * numeric tree debug commands - if tree was not intizlied yet, reply as if it was an empty tree * fix reset range->invertedIndexSize in applyNumIdx when index is empty. it will be counted when the range is deleted * add initial inverted index size if numeric to test don't intialize the numeric inverted index in the child if it's empty * fix gc warning * little fix * changes rt->invertedIndexSize to rt->invertedIndexesSize fix IndexSpec_collect_numeric_overhead and add test fix sending NULL to FGC_sendFixed * CalculateNumericInvertedIndexMemory fix identation and break when failing * rename openNumericKeysDict variable write to initialize and change type to bool * change OPEN_INDEX_READ to DONT_CREATE_INDEX and OPEN_INDEX_WRITE to CREATE_INDEX add coments to tests * use Likst instead of list to support older python versopns * remove type from function args * fix --------- Co-authored-by: GuyAv46 <[email protected]> (cherry picked from commit 357d356) * !! align FGC_parentHandleMissingDocs with master !! update gc stats after removing the invecrted index

MOD-8097, MOD-8114 Fix memory counting (#5204) * expose openNumericKeysDict, requires a write/read argument. replaces all OpenNumericIndex calls. * fork gc -initialiize index pointer only once rename createIndex to createPopulateTermsInvIndex add function to tests/index_utils generalize FGCTest suit add inheriting suite for tags some cosmetics in cpp_range ExtTest.testDynamicLoading: log when assert fails (the test halts upon assertion and the log will be nerver pronted) in IndexTest and QueryTest use IndexSpec_RemoveFromGlobals instead of StrongRef_Release to ensure the spec is properly cleaned otherwise some structs like the global prefix trie nodes, potentially accessed by following tests, might hold a copy of the spec refrence that was already released and cuase a crash. * add inverted index total size to the numeric tree root update in gc and in addition add cpp_range test and cpp_forkgc test * FIX: decrease gc->stats.totalCollected by bytesAdded to not count bytes added by us remove reseting currNode->range->invertedIndexSize in applyNumIdx called upon each node recieved from the child process instead, we will count the memory released when trimming the empty leaves call IndexSpec_TotalMemUsage from RediSearch_MemUsage instead of code duplication add IndexSpec_collect_numeric_overhead that counts the memory of NumericRangeTree in a given spec call this function to add it to IndexSpec_TotalMemUsage update LLApiTest testInfo* accordingly * fixes * fix new cpp forkgc test * fix unsafe read * fix status check * numeric tree debug commands - if tree was not intizlied yet, reply as if it was an empty tree * fix reset range->invertedIndexSize in applyNumIdx when index is empty. it will be counted when the range is deleted * add initial inverted index size if numeric to test don't intialize the numeric inverted index in the child if it's empty * fix gc warning * little fix * changes rt->invertedIndexSize to rt->invertedIndexesSize fix IndexSpec_collect_numeric_overhead and add test fix sending NULL to FGC_sendFixed * CalculateNumericInvertedIndexMemory fix identation and break when failing * rename openNumericKeysDict variable write to initialize and change type to bool * change OPEN_INDEX_READ to DONT_CREATE_INDEX and OPEN_INDEX_WRITE to CREATE_INDEX add coments to tests * use Likst instead of list to support older python versopns * remove type from function args * fix --------- Co-authored-by: GuyAv46 <[email protected]> (cherry picked from commit 357d356)

meiravgri · 2024-11-25T08:20:01Z

Won't backport to 2.6 as #4218, introducing improvements of inverted indexes memory tracking, essential for this pr, wasn't merged into 2.6

In addition, to avoid potential breaking changes, opening the numeric index for read won't be merged into 2.6 as well, since the mechanism of opening the numeric index is a bit different.

This reverts commit a97e925. Revert "[2.10] MOD-8601: Fix error message for LOAD (#5531) (#5550)" This reverts commit 5748233. Revert "[2.10] MOD-8561: Fix Inverted Index SeekTo Edge Case (#5536)" This reverts commit f9923b7. Revert "[2.10] Fix legacy geofilter leak - [MOD-8568] (#5516)" This reverts commit 8b75285. Revert "[2.10] MOD-7872: Add Types To Hide User Input (#5485)" This reverts commit d6acfed. Revert "[2.10] MOD-7864: Obfuscation API (#5487)" This reverts commit 259d417. Revert "MOD-7570 Backport 5422 to 2.10 (#5472)" This reverts commit e699dbc. Revert "[2.10] [MOD-8462] rename total_active_writes -> total_active_write_threads (#5446)" This reverts commit a90cfc3. Revert "[2.10] MOD-8058: Avoid error with _NUM_SSTRING (#5424)" This reverts commit 885cffb. Revert "[2.10] MOD-7634: Fix NOSTEM (#5392)" This reverts commit 4df508d. Revert "[2.10] Fixes for inverted indexes encoding - [MOD-8248] (#5388)" This reverts commit bc003c1. Revert "[2.10] Improve "Raw doc id" encoding - [MOD-8255] (#5372)" This reverts commit 9789e30. Revert "[2.10] Improve SkipToBlock logic - [MOD-8255] (#5370)" This reverts commit 56b16ea. Revert "[2.10] Improve counting active writes - [MOD-8151] (#5362)" This reverts commit a3d447d. Revert "[2.10] [MOD-8039] Reorganize stats files (#5356)" This reverts commit a1b8b29. Revert "[2.10] Improve text indexing performance - [MOD-8249] (#5352)" This reverts commit c5c9245. Revert "[2.10] [MOD-8115] Free spec resources in the main thread (#5327)" This reverts commit 42bb520. Revert "[2.10] Sorting Vector - [MOD-6783] (#5314)" This reverts commit 166a00e. Revert "[2.10] Fix long PREFIX/SUFFIX/INFIX/WILDCARD queries - [MOD-7882] (#5300)" This reverts commit 5c4ed74. Revert "[2.10] Remove assertion in optimizer (#5292)" This reverts commit ba55efb. Revert "[2.10] Fix missing expansions in text queries - [MOD-8142] (#5284)" This reverts commit 10f32d3. Revert "[2.10] Fix Tag OnReopen Callback - [MOD-8011] (#5282)" This reverts commit 2c029ee. Revert "[2.10] Fix MRIterator ownership mechanism - [MOD-8108] (#5266)" This reverts commit 457536e. Revert "[2.10] [MOD-8125] Create inverted index in write operations only (#5257)" This reverts commit 555557c. Revert "[2.10] [MOD-8152] rename and move new metrics (#5255)" This reverts commit a1f460e. Revert "[2.10] MOD-8129: Fix indexed_percent info stat (#5247)" This reverts commit d55afb5. Revert "[2.10] MOD-8097, MOD-8114 Fix memory counting (#5204) (#5241)" This reverts commit 52e58af. Revert "[MOD-7949] Avoid lazy expire upon scan keys in background [2.10] (#5294)" This reverts commit 6dfb503. Revert "[2.10] Cleanup for the trimming tree logic (#5226)" This reverts commit 7dfe6a2. Revert "[2.10] Add global marked deleted to stats [MOD-6053] (#5208)" This reverts commit f396732. Revert "Fix and add skips in pytest (#5210)" This reverts commit a8b505c. Revert "[2.10] [MOD-8035] Add field index error for JSON field indexing failures (#5202)" This reverts commit ed1919e. Revert "[2.10] Improve usage of RedisModuleCtx (#5181)" This reverts commit 8e686c7. Revert "[2.10] Add Active Indices Metrics - [MOD-7952] (#5179)" This reverts commit b72fc38. Revert "[2.10] MOD-8009: Allow Users To Configure Cursor Index Limitation Through Global Config (#5166)" This reverts commit 92e1dd8. Revert "[2.10] Send `SCORER` to the shards with `FT.AGGREGATE` on cluster mode - [MOD-8060] (#5163)" This reverts commit 4982ea2. Revert "[2.10] [MOD-8034] expose indexing failures to INFO modules (#5158)" This reverts commit d03f8de. Revert "[2.10] Add a counter of total query execution time - [MOD-7961] (#5153)" This reverts commit d45365d. Revert "[2.10] Add metrics of min and max memory consumption of indexes - [MOD-7584] (#5146)" This reverts commit 094318f. Revert "[2.10] Remove an unnecessary assertion - [MOD-8008] (#5142)" This reverts commit 3156f08. Revert "[2.10] Skip test_multithread:test_async_updates_sanity with sanitizer (#5125)" This reverts commit 5339033. Revert "Count Queries - [MOD-6492] (#5072)" This reverts commit 1006065. Revert "[2.10] Improve Error Message on VecSim Syntax Error - [MOD-7887] (#5109)" This reverts commit 9621fdc. Revert "[2.10] Fix flaky test (#5086)" This reverts commit 5aff118. Revert "[2.10] Fix flaky test_async_updates_sanity - avoid timeout in force invoke GC [MOD-7867] (#5081)" This reverts commit 2b85289. Revert "[2.10] MOD-7745: Add Benchmarks For Document And Field Expiration (#5063)" This reverts commit f99b0cf. Revert "[2.10] MOD-7863: Add HPEXPIRETIME Mock Support (#5028)" This reverts commit 79dfd0f.

meiravgri added 5 commits November 12, 2024 16:07

expose openNumericKeysDict, requires a write/read argument.

ba40a31

replaces all OpenNumericIndex calls.

add inverted index total size to the numeric tree root

3e3f6af

update in gc and in addition add cpp_range test and cpp_forkgc test

fixes

6fcb2a6

meiravgri and others added 3 commits November 13, 2024 14:02

Merge remote-tracking branch 'origin/master' into meiravg_memory_trac…

8336489

…king_fixes

fix new cpp forkgc test

65928fc

fix unsafe read

2df284b

meiravgri changed the title ~~expose openNumericKeysDict, requires a write/read argument.~~ MOD-8097 expose openNumericKeysDict, requires a write/read argument. Nov 17, 2024

GuyAv46 mentioned this pull request Nov 17, 2024

fix unsafe read #5217

Closed

2 tasks

GuyAv46 and others added 4 commits November 17, 2024 10:44

fix status check

9e71cfe

numeric tree debug commands - if tree was not intizlied yet, reply as…

b9c247c

… if it was an empty tree

fix reset range->invertedIndexSize in applyNumIdx when index is empty.

b98cccf

it will be counted when the range is deleted

add initial inverted index size if numeric to test

f54e808

don't intialize the numeric inverted index in the child if it's empty

GuyAv46 changed the title ~~MOD-8097 expose openNumericKeysDict, requires a write/read argument.~~ MOD-8097, MOD-8114 expose openNumericKeysDict, requires a write/read argument. Nov 17, 2024

meiravgri added 3 commits November 17, 2024 16:38

Merge branch 'master' into meiravg_memory_tracking_fixes

01ecb91

fix gc warning

74fe0da

little fix

3820cad

meiravgri requested review from GuyAv46 and alonre24 November 18, 2024 08:45

GuyAv46 reviewed Nov 18, 2024

View reviewed changes

GuyAv46 added bug backport 2.8 backport 2.10 backport 8.0 labels Nov 18, 2024

changes rt->invertedIndexSize to rt->invertedIndexesSize

cefae4c

fix IndexSpec_collect_numeric_overhead and add test fix sending NULL to FGC_sendFixed

meiravgri changed the title ~~MOD-8097, MOD-8114 expose openNumericKeysDict, requires a write/read argument.~~ MOD-8097, MOD-8114 Fix memory counting Nov 18, 2024

meiravgri added the backport 2.6 label Nov 18, 2024

GuyAv46 previously approved these changes Nov 18, 2024

View reviewed changes

meiravgri enabled auto-merge November 21, 2024 05:23

meiravgri requested a review from GuyAv46 November 21, 2024 05:23

GuyAv46 previously approved these changes Nov 21, 2024

View reviewed changes

meiravgri added this pull request to the merge queue Nov 21, 2024

fix

5a3e205

meiravgri removed this pull request from the merge queue due to a manual request Nov 21, 2024

meiravgri dismissed GuyAv46’s stale review via 5a3e205 November 21, 2024 05:59

meiravgri requested a review from GuyAv46 November 21, 2024 05:59

meiravgri enabled auto-merge November 21, 2024 06:00

GuyAv46 approved these changes Nov 21, 2024

View reviewed changes

meiravgri added this pull request to the merge queue Nov 21, 2024

Merged via the queue into master with commit 357d356 Nov 21, 2024

meiravgri deleted the meiravg_memory_tracking_fixes branch November 21, 2024 07:39

meiravgri mentioned this pull request Nov 21, 2024

[2.10] MOD-8097, MOD-8114 Fix memory counting (#5204) #5241

Merged

meiravgri mentioned this pull request Nov 21, 2024

[2.8] MOD-8097, MOD-8114 Fix memory counting #5242

Merged

meiravgri mentioned this pull request Nov 21, 2024

[8.0] MOD-8097, MOD-8114 Fix memory counting #5243

Merged

raz-mon mentioned this pull request Jan 30, 2025

[2.10] Revert 2.10 up to 2.10.8 #5586

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

MOD-8097, MOD-8114 Fix memory counting#5204

MOD-8097, MOD-8114 Fix memory counting#5204
meiravgri merged 22 commits intomasterfrom
meiravg_memory_tracking_fixes

meiravgri commented Nov 12, 2024 •

edited by GuyAv46

Loading

Uh oh!

codecov bot commented Nov 12, 2024 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GuyAv46 Nov 21, 2024

Uh oh!

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

meiravgri commented Nov 25, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

meiravgri commented Nov 12, 2024 • edited by GuyAv46 Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

wrong update of range->invertedIndexSize when returns empty from the fork process

fix overflow in spec->stats.invertedSize during FGC_updateStats

Fix bytes collected by the gc statistics

Additional changes in this PR

Additional bug fixes

Uh oh!

codecov bot commented Nov 12, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GuyAv46 Nov 21, 2024

Choose a reason for hiding this comment

Uh oh!

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

redisearch-backport-pull-request bot commented Nov 21, 2024

Uh oh!

meiravgri commented Nov 25, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

meiravgri commented Nov 12, 2024 •

edited by GuyAv46

Loading

wrong update of `range->invertedIndexSize` when returns empty from the fork process

fix overflow in spec->stats.invertedSize during `FGC_updateStats`

codecov bot commented Nov 12, 2024 •

edited

Loading