Use separate services for storing job queues and cache #7245
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
/check |
|
❌ Some checks failed |
|
/check |
|
❌ Some checks failed |
4 tasks
SpecLad
force-pushed
the
split-redis
branch
2 times, most recently
from
b2a2b9f to
0702584
Compare
|
/check |
|
❌ Some checks failed |
4 tasks
azhavoro
pushed a commit
that referenced
this pull request
Dec 15, 2023
<!-- Raise an issue to propose your change (https://github.com/opencv/cvat/issues). It helps to avoid duplication of efforts from multiple independent contributors. Discuss your ideas with maintainers to be sure that changes will be approved and merged. Read the [Contribution guide](https://opencv.github.io/cvat/docs/contributing/). --> <!-- Provide a general summary of your changes in the Title above --> ### Motivation and context <!-- Why is this change required? What problem does it solve? If it fixes an open issue, please link to the issue here. Describe your changes in detail, add screenshots. --> The main reason for this is to update the Helm chart, because the current one conflicts with the Redis chart I want to add (in #7245). But also the version we have is quite old and out of support, so it's due for an update anyway. The change in `objects.py` is due to the fact that with the new version of Clickhouse, the value of `date` somehow became a timezone-aware datetime object, so the output of `isoparse` now includes a UTC offset, and appending a "Z" now creates a malformed date. I chose to replace this with a custom format string, which is consistent with other date formatting code in the `analytics_report` app. ### How has this been tested? <!-- Please describe in detail how you tested your changes. Include details of your testing environment, and the tests you ran to see how your change affects other areas of the code, etc. --> I manually tried both the Helm chart and the Compose file, making sure that the analytics are still recorded and displayed. I also checked that a database created with the previous Clickhouse version can be loaded with the new version (it can). ### Checklist <!-- Go over all the following points, and put an `x` in all the boxes that apply. If an item isn't applicable for some reason, then ~~explicitly strikethrough~~ the whole line. If you don't do that, GitHub will show incorrect progress for the pull request. If you're unsure about any of these, don't hesitate to ask. We're here to help! --> - [x] I submit my changes into the `develop` branch - [ ] I have created a changelog fragment <!-- see top comment in CHANGELOG.md --> - ~~[ ] I have updated the documentation accordingly~~ - ~~[ ] I have added tests to cover my changes~~ - [x] I have linked related issues (see [GitHub docs]( https://help.github.com/en/github/managing-your-work-on-github/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword)) - ~~[ ] I have increased versions of npm packages if it is necessary ([cvat-canvas](https://github.com/opencv/cvat/tree/develop/cvat-canvas#versioning), [cvat-core](https://github.com/opencv/cvat/tree/develop/cvat-core#versioning), [cvat-data](https://github.com/opencv/cvat/tree/develop/cvat-data#versioning) and [cvat-ui](https://github.com/opencv/cvat/tree/develop/cvat-ui#versioning))~~ ### License - [x] I submit _my code changes_ under the same [MIT License]( https://github.com/opencv/cvat/blob/develop/LICENSE) that covers the project. Feel free to contact the maintainers if that's a concern.
azhavoro
pushed a commit
that referenced
this pull request
Dec 15, 2023
…les (#7254) The advantages of this are as follows: * It's much easier for a developer to use one `docker compose up` command to bring everything up than to run a custom command for each service. * We eliminate possible divergence of configuration (e.g. versions, command-line parameters) between what we actually use and what's listed in the documentation. * It makes it easier to update the developer guide if new dependencies are introduced. * And speaking of new dependencies, we have KeyDB now, which hasn't been added to the dev guide. The disadvantage is that we have to run an extra copy of the CVAT server, because otherwise OPA can't fetch its rules. I don't think it's a significant issue, since it doesn't prevent you from debugging anything. <!-- Raise an issue to propose your change (https://github.com/opencv/cvat/issues). It helps to avoid duplication of efforts from multiple independent contributors. Discuss your ideas with maintainers to be sure that changes will be approved and merged. Read the [Contribution guide](https://opencv.github.io/cvat/docs/contributing/). --> <!-- Provide a general summary of your changes in the Title above --> ### Motivation and context <!-- Why is this change required? What problem does it solve? If it fixes an open issue, please link to the issue here. Describe your changes in detail, add screenshots. --> Working on #7245, I realized that I don't want to add another custom command for running Redis in the development environment to the dev guide. So I wanted to remove the custom commands entirely. ### How has this been tested? <!-- Please describe in detail how you tested your changes. Include details of your testing environment, and the tests you ran to see how your change affects other areas of the code, etc. --> By manually following the updated instructions. ### Checklist <!-- Go over all the following points, and put an `x` in all the boxes that apply. If an item isn't applicable for some reason, then ~~explicitly strikethrough~~ the whole line. If you don't do that, GitHub will show incorrect progress for the pull request. If you're unsure about any of these, don't hesitate to ask. We're here to help! --> - [x] I submit my changes into the `develop` branch - ~~[ ] I have created a changelog fragment~~ <!-- see top comment in CHANGELOG.md --> - [x] I have updated the documentation accordingly - ~~[ ] I have added tests to cover my changes~~ - [x] I have linked related issues (see [GitHub docs]( https://help.github.com/en/github/managing-your-work-on-github/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword)) - ~~[ ] I have increased versions of npm packages if it is necessary ([cvat-canvas](https://github.com/opencv/cvat/tree/develop/cvat-canvas#versioning), [cvat-core](https://github.com/opencv/cvat/tree/develop/cvat-core#versioning), [cvat-data](https://github.com/opencv/cvat/tree/develop/cvat-data#versioning) and [cvat-ui](https://github.com/opencv/cvat/tree/develop/cvat-ui#versioning))~~ ### License - [x] I submit _my code changes_ under the same [MIT License]( https://github.com/opencv/cvat/blob/develop/LICENSE) that covers the project. Feel free to contact the maintainers if that's a concern.
These types of data have different characteristics and we have different expectations on them: * job queues are small and we'd rather not lose them (although losing them is not fatal); * cached chunks are large and we don't care if we lose them. We currently store both in KeyDB, which has shown itself to not be especially reliable. A few times we've had to clear the KeyDB store due to data corruption, which destroyed the queues as well. While we'll probably end up replacing KeyDB with something else, it would still be useful to have the ability to just clear the cache volume without taking out the job queues in the process. As a solution to this, add a Redis service to be used only for the queues (and potentially for other small data items). Using the original Redis instead of KeyDB should also help with reliability (at least as far as the job queues are concerned).
SpecLad
requested review from
azhavoro,
mdacoca,
Marishka17 and
nmanovic
as code owners
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## develop #7245 +/- ##
===========================================
+ Coverage 81.74% 81.79% +0.04%
===========================================
Files 367 367
Lines 39375 39375
Branches 3644 3644
===========================================
+ Hits 32188 32206 +18
+ Misses 7187 7169 -18
|
azhavoro
approved these changes
Dec 18, 2023
2 tasks
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation and context
These types of data have different characteristics and we have different expectations on them:
job queues are small and we'd rather not lose them (although losing them is not fatal);
cached chunks are large and we don't care if we lose them.
We currently store both in KeyDB, which has shown itself to not be especially reliable. A few times we've had to clear the KeyDB store due to data corruption, which destroyed the queues as well. While we'll probably end up replacing KeyDB with something else, it would still be useful to have the ability to just clear the cache volume without taking out the job queues in the process.
As a solution to this, add a Redis service to be used only for the queues (and potentially for other small data items). Using the original Redis instead of KeyDB should also help with reliability (at least as far as the job queues are concerned).
How has this been tested?
I checked the CVAT can still start using the development environment instructions, the Compose file and the Helm chart.
Checklist
developbranch[ ] I have updated the documentation accordingly[ ] I have added tests to cover my changes[ ] I have linked related issues (see GitHub docs)[ ] I have increased versions of npm packages if it is necessary(cvat-canvas,
cvat-core,
cvat-data and
cvat-ui)
License
Feel free to contact the maintainers if that's a concern.