ChainDB: store a ledger snapshot on startup after a long replay #1956
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
The LgrDB decides to make snapshots based on the time passed and the number of blocks advanced since the last snapshot. When it had to replay blocks on startup because the on-disk snapshots were out of date or missing, that number of replayed blocks is taken into account for the decision to make the next snapshot (as the starting point for the number of blocks advanced). However, the decision making logic is only triggered *after* copying a block to the ImmutableDB. So, for example, when starting a node with lots of blocks, but without on-disk ledger snapshots, it will replay all blocks to reconstruct an up-to-date ledger state, but it won't write it to disk until a block has been copied to the ImmutableDB. When experimenting, one often has a node that no longer downloads blocks acting as a server node, so that node will never write its snapshot to disk. This means a full ledger replay has to be done on every startup. Fix this by triggering the snapshot making logic before starting the loop that waits for blocks being copied to the ImmutableDB.
edsko
approved these changes
Apr 15, 2020
|
bors merge |
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Outdated suggestions cannot be applied.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.
The LgrDB decides to make snapshots based on the time passed and the number of
blocks advanced since the last snapshot. When it had to replay blocks on
startup because the on-disk snapshots were out of date or missing, that number
of replayed blocks is taken into account for the decision to make the next
snapshot (as the starting point for the number of blocks advanced).
However, the decision making logic is only triggered after copying a block
to the ImmutableDB. So, for example, when starting a node with lots of blocks,
but without on-disk ledger snapshots, it will replay all blocks to reconstruct
an up-to-date ledger state, but it won't write it to disk until a block has
been copied to the ImmutableDB. When experimenting, one often has a node that
no longer downloads blocks acting as a server node, so that node will never
write its snapshot to disk. This means a full ledger replay has to be done on
every startup.
Fix this by triggering the snapshot making logic before starting the loop that
waits for blocks being copied to the ImmutableDB.