Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Narek/release0.0.8 #235

Merged
merged 4 commits into from
Dec 4, 2023
Merged

Narek/release0.0.8 #235

merged 4 commits into from
Dec 4, 2023

Conversation

Ngalstyan4
Copy link
Contributor

No description provided.

Copy link

github-actions bot commented Dec 3, 2023

Benchmarks

metric old new pct change
recall (after create) 0.740 0.740 -
recall (after insert) 0.728 0.700 -3.85%
select bulk tps 507.666 505.953 -0.34%
select bulk latency (ms) 15.486 15.044 -2.85%
select bulk latency (stddev ms) 0.933 2.658 +184.89%
create latency (ms) 1204.685 1189.445 -1.27%
insert bulk tps 11.539 11.306 -2.02%
insert bulk latency (ms) 86.656 88.442 +2.06%
insert bulk latency (stddev ms) 3.368 2.277 -32.39%
disk usage (bytes) 6348800.000 6348800.000 -

Copy link

codecov bot commented Dec 3, 2023

Codecov Report

Merging #235 (ee52249) into main (1a6ba51) will decrease coverage by 0.35%.
The diff coverage is 62.50%.

Additional details and impacted files
@@            Coverage Diff             @@
##             main     #235      +/-   ##
==========================================
- Coverage   78.20%   77.85%   -0.35%     
==========================================
  Files          23       23              
  Lines        1647     1653       +6     
  Branches      405      410       +5     
==========================================
- Hits         1288     1287       -1     
- Misses        178      185       +7     
  Partials      181      181              
Files Coverage Δ
src/hooks/post_parse.c 88.75% <100.00%> (+0.14%) ⬆️
src/hooks/executor_start.c 88.37% <75.00%> (-1.63%) ⬇️
src/hnsw/options.c 85.00% <0.00%> (-2.94%) ⬇️

... and 2 files with indirect coverage changes

@Ngalstyan4 Ngalstyan4 force-pushed the narek/release0.0.8 branch 4 times, most recently from 5a92483 to 3fffb3e Compare December 4, 2023 02:35
So far update scripts only made changes in the SQL layer
This works so long as the data structures in the old Lantern and
new lantern versions are compatible.

This is not the case when we change storage format, or header page,
for example.

To mitigate this,we can either:
1. Write binary level update routines that patch necessary DB and lanter
 index pages
2. Recreate all binary data structures by rerunning equivalent SQL-level
 statements

This commit implements the second approach. We can stil have one-off
cases of the first. But, whenever something in the binary formats of
lantern chagnes and a specific migration script (1) does not exist,
the 2nd approach must appropriately be integrated, as demonstrated in
this commit
Running it has no benefit as it always runs on a single version
of the extension - the test suit drops and re create the extension
so the fact that we just ran an update does not make a difference
we can add this back when we actually restructure the tests
so CREATE routines can run on older version and the rest - on newer
updated version
Now that we preload lantern shared library, our hooks see more
statements than before.
With this we ignore create/update statements in hooks because
1) they are irrelevant
2) our hooks assume DB state that is not satisfied when these
statements are running

Warn when lantern is preloaded.

Note: These checks were added when considering a move to preloaded
lantern. We ultimately abandoned the move, so, the code is only relevant
in self-hosting if one decides to preload lantern.
@Ngalstyan4 Ngalstyan4 merged commit 96f50b3 into main Dec 4, 2023
@Ngalstyan4 Ngalstyan4 deleted the narek/release0.0.8 branch December 4, 2023 03:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant