Skip to content

eth/downloader: fix skeleton cleanup #28581

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.

Sign up for GitHub

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

Merged
merged 4 commits into from
Jan 31, 2024
Merged

eth/downloader: fix skeleton cleanup #28581

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
75929ca
eth/downloader: fix skeleton cleanup
rjl493456442 Nov 22, 2023
de234df
eth/downloader: short circuit if nothing to delete
rjl493456442 Dec 12, 2023
5b6571f
eth/downloader: polish the logic in cleanup
rjl493456442 Dec 12, 2023
eead5a2
eth/downloader: address comments
rjl493456442 Jan 31, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion eth/downloader/beaconsync.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,8 @@ func newBeaconBackfiller(dl *Downloader, success func()) backfiller {
}

// suspend cancels any background downloader threads and returns the last header
// that has been successfully backfilled.
// that has been successfully backfilled (potentially in a previous run), or the
// genesis.
func (b *beaconBackfiller) suspend() *types.Header {
// If no filling is running, don't waste cycles
b.lock.Lock()
Expand Down
1 change: 1 addition & 0 deletions eth/downloader/downloader.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -611,6 +611,7 @@ func (d *Downloader) syncWithPeer(p *peerConnection, hash common.Hash, td, ttd *
if err := d.lightchain.SetHead(origin); err != nil {
return err
}
log.Info("Truncated excess ancient chain segment", "oldhead", frozen-1, "newhead", origin)
}
}
// Initiate the sync using a concurrent header and content retrieval algorithm
Expand Down
78 changes: 47 additions & 31 deletions eth/downloader/skeleton.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -161,7 +161,7 @@ type backfiller interface {
// on initial startup.
//
// The method should return the last block header that has been successfully
// backfilled, or nil if the backfiller was not resumed.
// backfilled (in the current or a previous run), falling back to the genesis.
suspend() *types.Header

// resume requests the backfiller to start running fill or snap sync based on
Expand Down Expand Up @@ -382,14 +382,17 @@ func (s *skeleton) sync(head *types.Header) (*types.Header, error) {
done := make(chan struct{})
go func() {
defer close(done)
if filled := s.filler.suspend(); filled != nil {
// If something was filled, try to delete stale sync helpers. If
// unsuccessful, warn the user, but not much else we can do (it's
// a programming error, just let users report an issue and don't
// choke in the meantime).
if err := s.cleanStales(filled); err != nil {
log.Error("Failed to clean stale beacon headers", "err", err)
}
filled := s.filler.suspend()
if filled == nil {
log.Error("Latest filled block is not available")
return
}
// If something was filled, try to delete stale sync helpers. If
// unsuccessful, warn the user, but not much else we can do (it's
// a programming error, just let users report an issue and don't
// choke in the meantime).
if err := s.cleanStales(filled); err != nil {
log.Error("Failed to clean stale beacon headers", "err", err)
}
}()
// Wait for the suspend to finish, consuming head events in the meantime
Expand Down Expand Up @@ -1120,33 +1123,46 @@ func (s *skeleton) cleanStales(filled *types.Header) error {
number := filled.Number.Uint64()
log.Trace("Cleaning stale beacon headers", "filled", number, "hash", filled.Hash())

// If the filled header is below the linked subchain, something's
// corrupted internally. Report and error and refuse to do anything.
if number < s.progress.Subchains[0].Tail {
// If the filled header is below the linked subchain, something's corrupted
// internally. Report and error and refuse to do anything.
if number+1 < s.progress.Subchains[0].Tail {
Copy link
Member

Choose a reason for hiding this comment

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

I'm unsure that this is correct.

Number is the last item newly filled. When I start a sync loop, lets say I have N blocks in my chain. Then I skeleton sync, and will reconstruct headers from HEAD to N+1, and which point I have my subchain linked to the local chain.

I start backfilling. If I fail to fill anything, cleanStales does not get called. If I successfully fill 1 block, then filled will be N+1 (the newly filled one block).

In that case number == filled == N+1, which is == s.progress.Subchains[0].Tail. The only way to hit this error would be to backfill something we already had in our chain.

With the proposed modification however, backfilling my local head block would make the check pass, but that should not be possible, because I can only backfill stuff from my subchain, not 1 block below it.

Copy link
Member Author

Choose a reason for hiding this comment

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

If I fail to fill anything, cleanStales does not get called

I am not sure about it. cleanStales will be invoked anyway regardless if we successfully fill something or not. The CurrentSnapBlock() will be returned as the newly filled block.

Thus it's theoretically possible (1) the tail is N+1 and (2) filled is N, the error will be reported.

Copy link
Member

Choose a reason for hiding this comment

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

According the below, the will only call cleanStales if something was filled.

			if filled := s.filler.suspend(); filled != nil {
				// If something was filled, try to delete stale sync helpers. If
				// unsuccessful, warn the user, but not much else we can do (it's
				// a programming error, just let users report an issue and don't
				// choke in the meantime).
				if err := s.cleanStales(filled); err != nil {
					log.Error("Failed to clean stale beacon headers", "err", err)
				}
			}

That said, you are also right that resuming the filler will always return the head snap block (or genesis I guess in case we're brand new).

		defer func() {
			b.lock.Lock()
			b.filling = false
			b.filled = b.downloader.blockchain.CurrentSnapBlock()
			b.lock.Unlock()
		}()

Interesting, I guess this was a refactor after I designed the original idea. It kind of makes the filled check in my first code segment moot.

Copy link
Member

Choose a reason for hiding this comment

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

Ok, I agree with this +1, but we need to fix some comments that mention that it cannot happen. Namely,

return fmt.Errorf("filled header below beacon header tail: %d < %d", number, s.progress.Subchains[0].Tail)
}
// Subchain seems trimmable, push the tail forward up to the last
// filled header and delete everything before it - if available. In
// case we filled past the head, recreate the subchain with a new
// head to keep it consistent with the data on disk.
// If nothing in subchain is filled, don't bother to do cleanup.
if number+1 == s.progress.Subchains[0].Tail {
return nil
}
var (
start = s.progress.Subchains[0].Tail // start deleting from the first known header
end = number // delete until the requested threshold
start uint64
end uint64
batch = s.db.NewBatch()
)
s.progress.Subchains[0].Tail = number
Copy link
Member

Choose a reason for hiding this comment

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

Ah hmm, indeed this is wonky. If the tail is the first "missing" header, then it should be filled+1

s.progress.Subchains[0].Next = filled.ParentHash

if s.progress.Subchains[0].Head < number {
// If more headers were filled than available, push the entire
// subchain forward to keep tracking the node's block imports
end = s.progress.Subchains[0].Head + 1 // delete the entire original range, including the head
s.progress.Subchains[0].Head = number // assign a new head (tail is already assigned to this)

// The entire original skeleton chain was deleted and a new one
// defined. Make sure the new single-header chain gets pushed to
// disk to keep internal state consistent.
rawdb.WriteSkeletonHeader(batch, filled)
if number < s.progress.Subchains[0].Head {
// The skeleton chain is partially consumed, set the new tail as filled+1.
tail := rawdb.ReadSkeletonHeader(s.db, number+1)
if tail.ParentHash != filled.Hash() {
return fmt.Errorf("filled header is discontinuous with subchain: %d %s, please file an issue", number, filled.Hash())
}
start, end = s.progress.Subchains[0].Tail, number+1 // remove headers in [tail, filled]
s.progress.Subchains[0].Tail = tail.Number.Uint64()
s.progress.Subchains[0].Next = tail.ParentHash
} else {
// The skeleton chain is fully consumed, set both head and tail as filled.
start, end = s.progress.Subchains[0].Tail, filled.Number.Uint64() // remove headers in [tail, filled)
s.progress.Subchains[0].Tail = filled.Number.Uint64()
s.progress.Subchains[0].Next = filled.ParentHash

// If more headers were filled than available, push the entire subchain
// forward to keep tracking the node's block imports.
if number > s.progress.Subchains[0].Head {
end = s.progress.Subchains[0].Head + 1 // delete the entire original range, including the head
s.progress.Subchains[0].Head = number // assign a new head (tail is already assigned to this)

// The entire original skeleton chain was deleted and a new one
// defined. Make sure the new single-header chain gets pushed to
// disk to keep internal state consistent.
rawdb.WriteSkeletonHeader(batch, filled)
}
}
// Execute the trimming and the potential rewiring of the progress
s.saveSyncStatus(batch)
Expand Down