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

Add cache to stf context #19875

Closed
wants to merge 24 commits into from

Conversation

hieuvubk
Copy link
Contributor

Description

Closes: #19223


Author Checklist

All items are required. Please add a note to the item if the item is not applicable and
please add links to any relevant follow up issues.

I have...

  • Add object cache to stf context

  • get, set, remove of items affects cache

  • add tests

Reviewers Checklist

All items are required. Please add a note if the item is not applicable and please add
your handle next to the items reviewed if you only reviewed selected items.

I have...

  • confirmed the correct type prefix in the PR title
  • confirmed all author checklist items have been addressed
  • reviewed state machine logic, API design and naming, documentation is accurate, tests and test coverage

@hieuvubk hieuvubk requested a review from a team as a code owner March 26, 2024 19:14
Copy link
Contributor

coderabbitai bot commented Mar 26, 2024

Important

Auto Review Skipped

Auto reviews are disabled on base/target branches other than the default branch. Please add the base/target branch pattern to the list of additional branches to be reviewed in the settings.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.


Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media?

Share
Tips

Chat

There are 3 ways to chat with CodeRabbit:

  • Review comments: Directly reply to a review comment made by CodeRabbit. Example:
    • I pushed a fix in commit <commit_id>.
    • Generate unit testing code for this file.
    • Open a follow-up GitHub issue for this discussion.
  • Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
    • @coderabbitai generate unit testing code for this file.
    • @coderabbitai modularize this function.
  • PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
    • @coderabbitai generate interesting stats about this repository and render them as a table.
    • @coderabbitai show all the console.log statements in this repository.
    • @coderabbitai read src/utils.ts and generate unit testing code.
    • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (invoked as PR comments)

  • @coderabbitai pause to pause the reviews on a PR.
  • @coderabbitai resume to resume the paused reviews.
  • @coderabbitai review to trigger a review. This is useful when automatic reviews are disabled for the repository.
  • @coderabbitai resolve resolve all the CodeRabbit review comments.
  • @coderabbitai help to get help.

Additionally, you can add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.

CodeRabbit Configration File (.coderabbit.yaml)

  • You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
  • Please see the configuration documentation for more information.
  • If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

  • Visit our Documentation for detailed information on how to use CodeRabbit.
  • Join our Discord Community to get help, request features, and share feedback.
  • Follow us on X/Twitter for updates and announcements.

@@ -391,6 +391,7 @@ type executionContext struct {
meter gas.Meter
events []event.Event
sender []transaction.Identity
Cache ModuleContainer
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

will this be a source of concurrency issues, when parrallized txs become a thing?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

do you mean by using the same ctx for many txs?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

u right we use clone ctx through makeContext, need a way to update actual ctx after exec tx.
Let me adjust

return kc
}

func unsafeString(b []byte) string { return *(*string)(unsafe.Pointer(&b)) }

Check warning

Code scanning / gosec

Use of unsafe calls should be audited

Use of unsafe calls should be audited
@hieuvubk hieuvubk marked this pull request as draft March 26, 2024 19:37
server/v2/stf/cache.go Fixed Show fixed Hide fixed
@@ -349,6 +357,7 @@
if err != nil {
return nil, nil, err
}
applyContextCache(ctx, ebCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -338,6 +345,7 @@
event.Attribute{Key: "mode", Value: "BeginBlock"},
)
}
applyContextCache(ctx, ebCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -314,6 +320,7 @@
event.Attribute{Key: "mode", Value: "BeginBlock"},
)
}
applyContextCache(ctx, bbCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -297,6 +302,7 @@
event.Attribute{Key: "mode", Value: "PreBlock"},
)
}
applyContextCache(ctx, pbCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -281,6 +285,7 @@
}
msgResps[i] = resp
}
applyContextCache(ctx, execCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -258,6 +261,7 @@
if applyErr != nil {
return nil, 0, nil, applyErr
}
applyContextCache(ctx, postTxCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -240,6 +242,7 @@
if applyErr != nil {
return nil, 0, nil, applyErr
}
applyContextCache(ctx, postTxCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@@ -205,6 +206,7 @@
if err != nil {
return 0, nil, err
}
applyContextCache(ctx, validateCtx)

Check warning

Code scanning / gosec

Errors unhandled.

Errors unhandled.
@hieuvubk
Copy link
Contributor Author

I have initialized the empty cache for ctx with each DeliverBlock. ctx.Cache will be updated during block execution (via the applyCache function).

One problem I'm facing is that the cache is only written by item.Set. Imagine a block that only executes Get functions (cache is empty so optimization is meaningless).

What if we modified the item.Get() function a bit? If the item is present but not in the cache, save it to the cache for later use

wdut? @tac0turtle

@tac0turtle
Copy link
Member

What if we modified the item.Get() function a bit? If the item is present but not in the cache, save it to the cache for later use

wdut? @tac0turtle

Yea this is needed!!

@hieuvubk
Copy link
Contributor Author

What if we modified the item.Get() function a bit? If the item is present but not in the cache, save it to the cache for later use
wdut? @tac0turtle

Yea this is needed!!

Adjusted in 29eb7d2

@hieuvubk hieuvubk marked this pull request as ready for review March 30, 2024 08:41
@@ -6,11 +6,15 @@ import (
"fmt"

"cosmossdk.io/collections/codec"
"cosmossdk.io/server/v2/stf"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems problematic, collections should not depend on server v2 and STF.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ty, Wrapped in a container.Service. Now collections is not depended on stf.

Copy link
Member

@kocubinski kocubinski left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

collections should not depend on server/v2

server/v2/stf/cache.go Dismissed Show dismissed Hide dismissed

require (
cosmossdk.io/core v0.11.0
cosmossdk.io/collections v0.4.0
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can this be removed?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hmm it's using for the test

Copy link
Contributor

@alpe alpe left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have very limited context on the new stf but I am bit worried about dirty reads in the caching approach. I could not find where state is reverted when a CacheMultiStore is discarded. Also the cache seems to be reused via applyContextCache by different methods. But I may got that wrong.

I found caching very hard to get right in the past. Especially on the app level. There should be caching on the store level already. It would be good to see some bechmarks how much another cache adds on the app level to this.
Ideally, caches come with configuration options so that people can trade memory for cpu and some metrics that show internals like usage size and hit rate.

server/v2/stf/cache.go Outdated Show resolved Hide resolved
// Package branch contains the core branch service interface.
package container

type Service interface {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

personal preference: Service is a very overused name. Naming is hard but In this context, I assume it is just an abstract cache?
Would it make sense to move the interface to collections package?

type Item[V any] Map[noKey, V]
type Item[V any] struct {
m Map[noKey, V]
getContainer func(ctx context.Context) container.Service
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

q: why is it called container and not cache?


func NewModuleContainer() ModuleContainer {
return ModuleContainer{
m: make(map[string]Container),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

note: this cache is not limited. In address package for example simplelru.NewLRU is used. Would it make sense here, too?
The cache is not safe for concurrent access. Looking into the future, would it make sense to support parallel execution?

func applyContextCache(dst, src context.Context) error {
srcExecutionCtx, ok := src.(*executionContext)
if !ok {
return fmt.Errorf("Can not convert ctx to executionContext")
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like the errors are not handled by the callers. Would it make sense to drop them here already?

@tac0turtle
Copy link
Member

I have very limited context on the new stf but I am bit worried about dirty reads in the caching approach. I could not find where state is reverted when a CacheMultiStore is discarded. Also the cache seems to be reused via applyContextCache by different methods. But I may got that wrong.

I found caching very hard to get right in the past. Especially on the app level. There should be caching on the store level already. It would be good to see some bechmarks how much another cache adds on the app level to this. Ideally, caches come with configuration options so that people can trade memory for cpu and some metrics that show internals like usage size and hit rate.

Thanks for reviewing. I agree this is hard. The idea here is to cache decoded values. You are right there are caches at the store level and within the database. We have found for certain items decoding is a large overhead and done many times during a block. This is meant to reduce this

collections/go.mod Outdated Show resolved Hide resolved
Copy link
Member

@tac0turtle tac0turtle left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should make sure this works in the parallel execution case as well.

return ctx.(*executionContext)
}

func NewExecutionContext() *executionContext {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

these things should not be exported

return NewSchemaBuilderFromAccessor(service.OpenKVStore)
sb := NewSchemaBuilderFromAccessor(service.OpenKVStore)
kl, ok := service.(interface {
OpenContainer(ctx context.Context) container.Service
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this interface should be in core

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

resolved

return item
}

// Get gets the item, if it is not set it returns an ErrNotFound error.
// If value decoding fails then an ErrEncoding is returned.
func (i Item[V]) Get(ctx context.Context) (V, error) {
return (Map[noKey, V])(i).Get(ctx, noKey{})
var toCache bool
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is this boolean really needed?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Its used for case cache is empty when we get value from Item. So we will cache value of 1st get.

@tac0turtle
Copy link
Member

closing this as we are upstreaming server modular to main and should add a quick adr on the design

@tac0turtle tac0turtle closed this Jun 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants