Persist ChannelMonitors after new blocks are connected #1108

TheBlueMatt · 2021-10-05T17:34:40Z

This resolves several user complaints (and issues in the sample
node) where startup is substantially delayed as we're always
waiting for the chain data to sync.

Further, in an upcoming PR, we'll be reloading pending payments
from ChannelMonitors on restart, at which point we'll need the
change here which avoids handling events until after the user
has confirmed the ChannelMonitor has been persisted to disk.
It will avoid a race where we

send a payment/HTLC (persisting the monitor to disk with the
HTLC pending),
force-close the channel, removing the channel entry from the
ChannelManager entirely,
persist the ChannelManager,
connect a block which contains a fulfill of the HTLC, generating
a claim event,
handle the claim event while the ChannelMonitor is being
persisted,
persist the ChannelManager (before the CHannelMonitor is
persisted fully),
restart, reloading the HTLC as a pending payment in the
ChannelManager, which now has no references to it except from
the ChannelMonitor which still has the pending HTLC,
replay the block connection, generating a duplicate PaymentSent
event.

jkczyz · 2021-10-05T17:45:41Z

lightning/src/chain/channelmonitor.rs

+	/// Failures here do not imply the channel will be force-closed, however any future calls to
+	/// [`update_persisted_channel`] after an error is returned here MUST either persist the full,
+	/// updated [`ChannelMonitor`] provided to [`update_persisted_channel`] or return
+	/// [`ChannelMonitorUpdateErr::PermanentFailure`], force-closing the channel. In other words,
+	/// any future calls to [`update_persisted_channel`] after an error here MUST NOT persist the
+	/// [`ChannelMonitorUpdate`] alone.


Could we simplify this by making update_persisted_channel take an Option<&ChannelMonitorUpdate> and forgoing adding sync_persisted_channel method to the trait? Then when the user sees None they must write the full ChannelMonitor.

The return value distinction is the harder part to capture - users don't really have the option to return a PermanentFailure here as the call site (the ChainMonitor) doesn't have the ability to force-close a channel.

Ok, I think I understand the subtly here. But shouldn't we not call update_persisted_channel once sync_persisted_channel returns an error rather than rely on the user to implement update_persisted_channel correctly? Or at very least never pass them a ChannelMonitorUpdate there (making it an Option) even if we still need the extra method.

Yea, that's a good point, I guess I hadn't thought about making the update an option and skipping it if we've marked a channel as "previous sync didn't persist". The "don't call update_persisted_channel" change, though feels wrong - technically if the user complies with the API docs we can continue operating fine, after a restart, but if we don't call update_persisted_channel at all we have to force-close the channel immediately.

What if after a block comes in, we set a bool in ChannelMonitor indicating there’s been a new block? Then we can change update_persisted_channel to pass in whether a full sync is needed to avoid a big chain resync.

Can also check the new bool in timer_tick and call update_persisted_channel then, if we also make the ChannelMonitorUpdate an Option maybe

We still need to sync the monitor before we can hand any chain-generated events back to the ChannelManager, so I don't think we can avoid a sync at all.

codecov · 2021-10-05T17:56:12Z

Codecov Report

Merging #1108 (240fc03) into main (dda86a0) will decrease coverage by 0.18%.
The diff coverage is 90.27%.

❗ Current head 240fc03 differs from pull request most recent head 6fb5bd3. Consider uploading reports for the commit 6fb5bd3 to get more accurate results

@@            Coverage Diff             @@
##             main    #1108      +/-   ##
==========================================
- Coverage   90.58%   90.40%   -0.19%     
==========================================
  Files          66       67       +1     
  Lines       34459    34658     +199     
==========================================
+ Hits        31215    31332     +117     
- Misses       3244     3326      +82

Impacted Files	Coverage Δ
lightning/src/chain/mod.rs	`61.11% <ø> (ø)`
lightning/src/util/byte_utils.rs	`100.00% <ø> (ø)`
lightning/src/chain/channelmonitor.rs	`90.94% <64.70%> (-0.32%)`	⬇️
lightning/src/ln/channelmanager.rs	`83.61% <77.77%> (-1.52%)`	⬇️
lightning/src/chain/chainmonitor.rs	`90.87% <85.71%> (-5.49%)`	⬇️
lightning-persister/src/lib.rs	`94.30% <100.00%> (+0.09%)`	⬆️
lightning/src/ln/chanmon_update_fail_tests.rs	`97.65% <100.00%> (ø)`
lightning/src/ln/functional_tests.rs	`97.30% <100.00%> (-0.03%)`	⬇️
lightning/src/ln/peer_handler.rs	`45.67% <100.00%> (-0.21%)`	⬇️
lightning/src/util/atomic_counter.rs	`100.00% <100.00%> (ø)`
... and 9 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dda86a0...6fb5bd3. Read the comment docs.

TheBlueMatt · 2021-10-05T18:11:20Z

Also shoved in two commtis to move Persist and ChannelMonitorUpdateErr cause their locations was just wrong.

jkczyz · 2021-10-05T19:04:45Z

lightning/src/chain/channelmonitor.rs

+	/// Failures here do not imply the channel will be force-closed, however any future calls to
+	/// [`update_persisted_channel`] after an error is returned here MUST either persist the full,
+	/// updated [`ChannelMonitor`] provided to [`update_persisted_channel`] or return
+	/// [`ChannelMonitorUpdateErr::PermanentFailure`], force-closing the channel. In other words,
+	/// any future calls to [`update_persisted_channel`] after an error here MUST NOT persist the
+	/// [`ChannelMonitorUpdate`] alone.


Ok, I think I understand the subtly here. But shouldn't we not call update_persisted_channel once sync_persisted_channel returns an error rather than rely on the user to implement update_persisted_channel correctly? Or at very least never pass them a ChannelMonitorUpdate there (making it an Option) even if we still need the extra method.

jkczyz · 2021-10-05T19:12:09Z

lightning/src/chain/chainmonitor.rs

+					let mut ev_lock = self.event_mutex.lock().unwrap();
+					txn_outputs = process(monitor, txdata);
+					log_trace!(self.logger, "Syncing Channel Monitor for channel {}", log_funding_info!(monitor));
+					if let Err(()) = self.persister.sync_persisted_channel(*funding_outpoint, monitor) {


Do you have plans to aggregate updates across multiple blocks (in case of a restart)? Or maybe that is also problematic if offline for awhile and wouldn't persist until the end of the sync. 😕

I think we're "just fine" here, somewhat due to an API quirk. The on-load replay is supposed to be per-channelmonitor, and not via the chain::Watch/ChainMonitor, or at least we recommend it. Technically I think a user could use ChainMontior during that time, but at least we dont?

lightning/src/chain/channelmonitor.rs

valentinewallace · 2021-10-06T07:58:22Z

Re: the first commit, would it be possible to separate out the fix for the redundant chain sync from the fix for the duplicate PaymentSent scenario? Seems that might ease review

TheBlueMatt · 2021-10-08T18:57:23Z

Making this a draft for now, I think it'll end up depending on a new test-refactor/ChainMonitor-api-refactor pr.

TheBlueMatt · 2021-10-08T23:35:09Z

This is now based on #1112 and should have largely addressed the feedback.

lightning/src/chain/chainmonitor.rs

lightning/src/chain/channelmonitor.rs

lightning/src/chain/chainmonitor.rs

TheBlueMatt · 2021-10-17T22:29:18Z

Tested as a part of #1130.

jkczyz · 2021-10-18T15:30:09Z

lightning/src/chain/chainmonitor.rs

+					monitor_state.last_chain_persist_height.load(Ordering::Acquire) + LATENCY_GRACE_PERIOD_BLOCKS as usize
+						> self.highest_chain_height.load(Ordering::Acquire)


Can there be an edge case around reorgs here where highest_chain_height is reduced while blocks are disconnected, release_pending_monitor_events is called, and then the new blocks are connected?

Hmm, yea, was thinking it wasn't worth worrying about, but its easy enough to just fetch_max, so I did that.

valentinewallace

Thanks for adding the testing! Almost ready to sign off

lightning/src/chain/chainmonitor.rs

valentinewallace · 2021-10-18T21:32:53Z

lightning/src/chain/chainmonitor.rs

+		entry.insert(MonitorHolder {
+			monitor,
+			pending_monitor_updates: Mutex::new(pending_monitor_updates),
+			channel_perm_failed: AtomicBool::new(false),


Could we be a bit smarter and check monitor.locked_from_offchain? Or, is there any way to inquire the monitor rather than defaulting to false?

I don't think we care? If we get locked_from_offchain that means ChannelManager told us the channel is closed, so we're probably not gonna get into watch_channel at that point. The only case where it can be perm-failed at this point is if we failed to persist above, but then we'd have returned early already.

We also take this opportunity to drop byte_utils::le64_to_array, as our MSRV now supports the native to_le_bytes() call.

TheBlueMatt · 2021-10-18T22:59:28Z

Squashed all fixups except for new ones from the latest round of feedback.

jkczyz · 2021-10-19T04:59:44Z

lightning/src/chain/chainmonitor.rs

+			let mut old_height = self.highest_chain_height.load(Ordering::Relaxed);
+			while self.highest_chain_height
+				.compare_exchange(old_height, height as usize, Ordering::AcqRel, Ordering::Relaxed).is_err()
+			{
+				old_height = self.highest_chain_height.load(Ordering::Acquire);
+			}


So taking a closer look, we actually never call process_chain_data when blocks are disconnected, so highest_chain_height won't be reduced. I suppose a user could call best_block_updated with a smaller height, though, and there's some logic in ChannelMonitor to handle this case.

But when given a smaller height, won't this not be equivalent to fetch_max?

I believe its still an issue in 2-block reorgs - you may disconnect two blocks and then connect one and see a height that is one lower than the target.

And, oops, yes, added the requisite compare :)

valentinewallace

Looks good (mod Jeff outstanding comments)! 🚀

If anything, I wonder if testing is a touch light given the edge case-yness, but fine to chat more about that in #1130.

I'd be happy to check out #1104 next if that can get a rebase and CI fix :)

jkczyz

This should be pretty much good to go, a couple comments outstanding.

lightning/src/chain/chainmonitor.rs

jkczyz · 2021-10-19T20:17:52Z

lightning/src/chain/chainmonitor.rs

-	/// it is up to you to maintain a correct mapping between the outpoint and the
-	/// stored channel data). Note that you **must** persist every new monitor to
-	/// disk. See the `Persist` trait documentation for more details.
+	/// Persist a new channel's data. The data can be stored any way you want, but the identifier


Just wanted to circle back on this. We could possibly rename persist_new_channel to persist_channel and pass it enum stating the type of persistence: new channel, re-persistence, etc. Then get rid of the Option in the update_persisted_channel parameter. But that may involve piping the enum through chain::Watch? I don't feel strongly about it, but the docs may need to be clarified here.

lightning/src/chain/chainmonitor.rs

jkczyz · 2021-10-19T21:25:21Z

lightning/src/chain/chainmonitor.rs

-	/// it is up to you to maintain a correct mapping between the outpoint and the
-	/// stored channel data). Note that you **must** persist every new monitor to
-	/// disk. See the `Persist` trait documentation for more details.
+	/// Persist a new channel's data. The data can be stored any way you want, but the identifier


Right, my concern is also that the chain::Watch should/should be able to reject new channelmonitors for the same outpoint twice. I believe we actually rely on this behavior somewhat to avoid some attacks. If we have an enum that controls whether the user is supposed to return an error or not, I'm very worried they just won't do it.

Good point. Happy to leave it as is.

Which docs do you think need to be updated?

Ah, just that this is not necessarily a new channel since it may be called by watch_channel upon restart as per ChannelManagerReadArgs' docs.

valentinewallace · 2021-10-19T22:04:52Z

lightning/src/chain/chainmonitor.rs

+			let mut old_height = self.highest_chain_height.load(Ordering::Relaxed);
+			let new_height = height as usize;
+			while new_height > old_height && self.highest_chain_height
+				.compare_exchange(old_height, new_height, Ordering::AcqRel, Ordering::Relaxed).is_err()
+			{
+				old_height = self.highest_chain_height.load(Ordering::Acquire);
+			}


I'd find this simpler/more readable:

let old_height = .. let new_height = .. if new_height > old_height { self.highest_chain_height.store(new_height) }

maybe I'm missing something here?

Duh, yes, can do that as long as its under the lock. I've moved the code down and simplified.

jkczyz

ACK 240fc03

TheBlueMatt · 2021-10-19T22:33:20Z

Squashed without changes, diff from Val's ACK (will land after CI):

$ git diff-tree -U1 1816e2b96884961e04e1041a9d0f88ac347b6a33 a65686cc4
diff --git a/lightning/src/chain/chainmonitor.rs b/lightning/src/chain/chainmonitor.rs
index 0e66c352d..b52eb20a7 100644
--- a/lightning/src/chain/chainmonitor.rs
+++ b/lightning/src/chain/chainmonitor.rs
@@ -264,15 +264,14 @@ where C::Target: chain::Filter,
 		let mut dependent_txdata = Vec::new();
-		if let Some(height) = best_height {
-			// Sadly AtomicUsize::fetch_max wasn't stabilized until 1.45, so until then we have to
-			// manually CAS.
-			let mut old_height = self.highest_chain_height.load(Ordering::Relaxed);
-			let new_height = height as usize;
-			while new_height > old_height && self.highest_chain_height
-				.compare_exchange(old_height, new_height, Ordering::AcqRel, Ordering::Relaxed).is_err()
-			{
-				old_height = self.highest_chain_height.load(Ordering::Acquire);
-			}
-		}
 		{
 			let monitor_states = self.monitors.write().unwrap();
+			if let Some(height) = best_height {
+				// If the best block height is being updated, update highest_chain_height under the
+				// monitors write lock.
+				let old_height = self.highest_chain_height.load(Ordering::Acquire);
+				let new_height = height as usize;
+				if new_height > old_height {
+					self.highest_chain_height.store(new_height, Ordering::Release);
+				}
+			}
+
 			for (funding_outpoint, monitor_state) in monitor_states.iter() {
$

In the next commit we'll need ChainMonitor to "see" when a monitor persistence completes, which means `monitor_updated` needs to move to `ChainMonitor`. The simplest way to then communicate that information to `ChannelManager` is via `MonitorEvet`s, which seems to line up ok, even if they're now constructed by multiple different places.

In the next commit, we'll be originating monitor updates both from the ChainMonitor and from the ChannelManager, making simple sequential update IDs impossible. Further, the existing async monitor update API was somewhat hard to work with - instead of being able to generate monitor_updated callbacks whenever a persistence process finishes, you had to ensure you only did so at least once all previous updates had also been persisted. Here we eat the complexity for the user by moving to an opaque type for monitor updates, tracking which updates are in-flight for the user and only generating monitor-persisted events once all pending updates have been committed.

TheBlueMatt · 2021-10-19T23:49:32Z

Oops, added missing doclinks and dropped a link in private docs that needed linking, will land after CI:

$ git diff-tree -U1 a65686cc 6fb5bd36a
diff --git a/lightning/src/chain/chainmonitor.rs b/lightning/src/chain/chainmonitor.rs
index b52eb20a7..71b0b3e50 100644
--- a/lightning/src/chain/chainmonitor.rs
+++ b/lightning/src/chain/chainmonitor.rs
@@ -50,3 +50,3 @@ use core::sync::atomic::{AtomicBool, AtomicUsize, Ordering};
 enum UpdateOrigin {
-	/// An update that was generated by the [`ChannelManager`] (via our `chain::Watch`
+	/// An update that was generated by the `ChannelManager` (via our `chain::Watch`
 	/// implementation). This corresponds to an actual [`ChannelMonitorUpdate::update_id`] field
@@ -107,2 +107,3 @@ pub trait Persist<ChannelSigner: Sign> {
 	///
+	/// [`ChannelManager`]: crate::ln::channelmanager::ChannelManager
 	/// [`Writeable::write`]: crate::util::ser::Writeable::write
diff --git a/lightning/src/chain/mod.rs b/lightning/src/chain/mod.rs
index fbe22e6ed..25e5a97d2 100644
--- a/lightning/src/chain/mod.rs
+++ b/lightning/src/chain/mod.rs
@@ -216,2 +216,4 @@ pub enum ChannelMonitorUpdateErr {
 	/// updates will return TemporaryFailure until the remote copies could be updated.
+	///
+	/// [`ChainMonitor::channel_monitor_updated`]: chainmonitor::ChainMonitor::channel_monitor_updated
 	TemporaryFailure,
$

This resolves several user complaints (and issues in the sample node) where startup is substantially delayed as we're always waiting for the chain data to sync. Further, in an upcoming PR, we'll be reloading pending payments from ChannelMonitors on restart, at which point we'll need the change here which avoids handling events until after the user has confirmed the `ChannelMonitor` has been persisted to disk. It will avoid a race where we * send a payment/HTLC (persisting the monitor to disk with the HTLC pending), * force-close the channel, removing the channel entry from the ChannelManager entirely, * persist the ChannelManager, * connect a block which contains a fulfill of the HTLC, generating a claim event, * handle the claim event while the `ChannelMonitor` is being persisted, * persist the ChannelManager (before the CHannelMonitor is persisted fully), * restart, reloading the HTLC as a pending payment in the ChannelManager, which now has no references to it except from the ChannelMonitor which still has the pending HTLC, * replay the block connection, generating a duplicate PaymentSent event.

ChannelMonitors now require that they be re-persisted before MonitorEvents be provided to the ChannelManager, the exact thing that test_dup_htlc_onchain_fails_on_reload was testing for when it *didn't* happen. As such, test_dup_htlc_onchain_fails_on_reload is now testing that we bahve correctly when the API guarantees are not met, something we don't need to do. Here, we adapt it to test the new API requirements through ChainMonitor's calls to the Persist trait instead.

If we have a `ChannelMonitor` update from an on-chain event which returns a `TemporaryFailure`, we block `MonitorEvent`s from that `ChannelMonitor` until the update is persisted. This prevents duplicate payment send events to the user after payments get reloaded from monitors on restart. However, if the event being avoided isn't going to generate a PaymentSent, but instead result in us claiming an HTLC from an upstream channel (ie the HTLC was forwarded), then the result of a user delaying the event is that we delay getting our money, not a duplicate event. Because user persistence may take an arbitrary amount of time, we need to bound the amount of time we can possibly wait to return events, which we do here by bounding it to 3 blocks. Thanks to Val for catching this in review.

Its somewhat confusing that `persist_new_channel` is called on startup for an existing channel in common deployments, so we call it out explicitly.

TheBlueMatt added this to the 0.0.102 milestone Oct 5, 2021

TheBlueMatt mentioned this pull request Oct 5, 2021

Reload pending outbound payments from ChannelMonitors on startup #1104

Merged

jkczyz reviewed Oct 5, 2021

View reviewed changes

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from 29c7f1e to 60a8a70 Compare October 5, 2021 18:11

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch 2 times, most recently from 322bc11 to bde998c Compare October 5, 2021 18:38

jkczyz reviewed Oct 5, 2021

View reviewed changes

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from bde998c to 96d9895 Compare October 7, 2021 20:38

TheBlueMatt mentioned this pull request Oct 7, 2021

Auto-Detect ChannelMonitor Persistence on Startup #1111

Closed

TheBlueMatt marked this pull request as draft October 8, 2021 18:57

TheBlueMatt mentioned this pull request Oct 8, 2021

Hide internals of ChainMonitor behind getter #1112

Merged

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from 96d9895 to 66cc962 Compare October 8, 2021 23:34

TheBlueMatt marked this pull request as ready for review October 8, 2021 23:35

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch 6 times, most recently from 374e622 to d3f0772 Compare October 10, 2021 23:34

TheBlueMatt mentioned this pull request Oct 13, 2021

Pre-scoring test clean-ups #1120

Merged

valentinewallace reviewed Oct 13, 2021

View reviewed changes

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from d3f0772 to 33c37f0 Compare October 13, 2021 18:44

valentinewallace reviewed Oct 13, 2021

View reviewed changes

lightning/src/chain/chainmonitor.rs Outdated Show resolved Hide resolved

lightning/src/chain/chainmonitor.rs Outdated Show resolved Hide resolved

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch 2 times, most recently from 895d4ce to f73999d Compare October 13, 2021 20:07

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from 4725103 to f0aa603 Compare October 17, 2021 22:59

jkczyz reviewed Oct 18, 2021

View reviewed changes

valentinewallace removed the MOAR TEST PLZ NEEDS MOAR TEST label Oct 18, 2021

valentinewallace reviewed Oct 18, 2021

View reviewed changes

TheBlueMatt added 2 commits October 18, 2021 22:04

Move MonitorEvent serialization to TLV-enum-upgradable from custom

e9f1934

Move the two-AtomicUsize counter in peer_handler to a util struct

df8bde9

We also take this opportunity to drop byte_utils::le64_to_array, as our MSRV now supports the native to_le_bytes() call.

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch 2 times, most recently from 3eb7879 to 2a22a16 Compare October 18, 2021 23:27

jkczyz reviewed Oct 19, 2021

View reviewed changes

valentinewallace approved these changes Oct 19, 2021

View reviewed changes

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from 8f47cf8 to c1c5462 Compare October 19, 2021 17:52

jkczyz reviewed Oct 19, 2021

View reviewed changes

jkczyz approved these changes Oct 19, 2021

View reviewed changes

valentinewallace approved these changes Oct 19, 2021

View reviewed changes

jkczyz approved these changes Oct 19, 2021

View reviewed changes

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from 240fc03 to a65686c Compare October 19, 2021 22:32

TheBlueMatt added 2 commits October 19, 2021 23:49

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from a65686c to a3d07e6 Compare October 19, 2021 23:49

TheBlueMatt added 4 commits October 20, 2021 00:06

Clarify the contexts in which persist_new_channel may be called

6fb5bd3

Its somewhat confusing that `persist_new_channel` is called on startup for an existing channel in common deployments, so we call it out explicitly.

TheBlueMatt force-pushed the 2021-10-persist-mon-blocks branch from a3d07e6 to 6fb5bd3 Compare October 20, 2021 00:06

TheBlueMatt merged commit 107c6c7 into lightningdevkit:main Oct 20, 2021

G8XSU mentioned this pull request Mar 21, 2024

Don't pause events for chainsync persistence #2957

Merged

11 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Persist ChannelMonitors after new blocks are connected #1108

Persist ChannelMonitors after new blocks are connected #1108

TheBlueMatt commented Oct 5, 2021

jkczyz Oct 5, 2021

TheBlueMatt Oct 5, 2021

jkczyz Oct 5, 2021

TheBlueMatt Oct 5, 2021

valentinewallace Oct 6, 2021

TheBlueMatt Oct 6, 2021

codecov bot commented Oct 5, 2021 •

edited

Loading

TheBlueMatt commented Oct 5, 2021

jkczyz Oct 5, 2021

jkczyz Oct 5, 2021

TheBlueMatt Oct 5, 2021

valentinewallace commented Oct 6, 2021

TheBlueMatt commented Oct 8, 2021

TheBlueMatt commented Oct 8, 2021

TheBlueMatt commented Oct 17, 2021

jkczyz Oct 18, 2021

TheBlueMatt Oct 18, 2021

valentinewallace left a comment

valentinewallace Oct 18, 2021

TheBlueMatt Oct 18, 2021

TheBlueMatt commented Oct 18, 2021

jkczyz Oct 19, 2021

TheBlueMatt Oct 19, 2021

valentinewallace left a comment •

edited

Loading

jkczyz left a comment

jkczyz Oct 19, 2021

jkczyz Oct 19, 2021

valentinewallace Oct 19, 2021

TheBlueMatt Oct 19, 2021

jkczyz left a comment

TheBlueMatt commented Oct 19, 2021

TheBlueMatt commented Oct 19, 2021 •

edited

Loading

		monitor_state.last_chain_persist_height.load(Ordering::Acquire) + LATENCY_GRACE_PERIOD_BLOCKS as usize
		> self.highest_chain_height.load(Ordering::Acquire)

Persist ChannelMonitors after new blocks are connected #1108

Persist ChannelMonitors after new blocks are connected #1108

Conversation

TheBlueMatt commented Oct 5, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov bot commented Oct 5, 2021 • edited Loading

Codecov Report

TheBlueMatt commented Oct 5, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

valentinewallace commented Oct 6, 2021

TheBlueMatt commented Oct 8, 2021

TheBlueMatt commented Oct 8, 2021

TheBlueMatt commented Oct 17, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

valentinewallace left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

TheBlueMatt commented Oct 18, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

valentinewallace left a comment • edited Loading

Choose a reason for hiding this comment

jkczyz left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jkczyz left a comment

Choose a reason for hiding this comment

TheBlueMatt commented Oct 19, 2021

TheBlueMatt commented Oct 19, 2021 • edited Loading

codecov bot commented Oct 5, 2021 •

edited

Loading

valentinewallace left a comment •

edited

Loading

TheBlueMatt commented Oct 19, 2021 •

edited

Loading