This repository was archived by the owner on Aug 2, 2022. It is now read-only.
remove speculation when stale#4649
Merged
heifner merged 3 commits intoJul 14, 2018
Merged
Conversation
in the current default mode, a node attempts to act as a good p2p citizen and relay transactions. This requires creating a speculative block. However, in times when a network has steady transactions and the node is stale, this creates a storm of database activity managing our transaction deduplication list as "time" bounces from current for speculation to stale for applying incoming blocks. The new default is to only open a speculative block when the node is in sync. If transactions come in while the node is out of sync they are queued instead of immidiately executed and relayed. The result is a significant speed up when syncing in to a live network with some traffic on it. The old default can be restored by using the new `enable-stale-speculation` config parameter
heifner
approved these changes
Jul 13, 2018
heifner
reviewed
Jul 14, 2018
|
|
||
| producer_options.add_options() | ||
| ("enable-stale-production,e", boost::program_options::bool_switch()->notifier([this](bool e){my->_production_enabled = e;}), "Enable block production, even if the chain is stale.") | ||
| ("enable-stale-speculation", boost::program_options::bool_switch()->notifier([this](bool e){my->_enable_stale_speculation = e;}), "Enable speculative execution and relay, even if the chain is stale.") |
Contributor
There was a problem hiding this comment.
I don't see any reason to include this option.
Contributor
Author
There was a problem hiding this comment.
I'm leaning towards agreeing. It is only present to maintain the original default functionality however, there seems to be almost zero value in that mode of operation.
…able-stale-speculation
…now misleading ¯\_(ツ)_/¯
wanderingbort
pushed a commit
to wanderingbort/eos
that referenced
this pull request
Jul 17, 2018
Previously, we based the stale speculation check on the `_production_enabled` flag which is set, at most, once per process. This lead to a situation where a node which fell behind would still speculate. This changes that check to be based on the age of the head block so, nodes which fall behind will stop speculating. NOTE, this only affects speculative modes, actual production should attempt to proceed despite the age of the head block. relates to EOSIO#4649
This file contains hidden or 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
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
in the current default mode, a node attempts to act as a good p2p citizen and relay transactions. This requires creating a speculative block. However, in times when a network has steady transactions and the node is stale, this creates a storm of database activity managing our transaction deduplication list as "time" bounces from current for speculation to stale for applying incoming blocks.
The new default is to only open a speculative block when the node is in sync. If transactions come in while the node is out of sync they are queued instead of immidiately executed and relayed. The result is a significant speed up when syncing in to a live network with some traffic on it.
The old default can be restored by using the new
enable-stale-speculationconfig parameter