1 package org.ldk.structs;
3 import org.ldk.impl.bindings;
4 import org.ldk.enums.*;
6 import java.util.Arrays;
7 import java.lang.ref.Reference;
8 import javax.annotation.Nullable;
12 * Implements [`Persist`] in a way that writes and reads both [`ChannelMonitor`]s and
13 * [`ChannelMonitorUpdate`]s.
17 * The main benefit this provides over the [`KVStore`]'s [`Persist`] implementation is decreased
18 * I/O bandwidth and storage churn, at the expense of more IOPS (including listing, reading, and
19 * deleting) and complexity. This is because it writes channel monitor differential updates,
20 * whereas the other (default) implementation rewrites the entire monitor on each update. For
21 * routing nodes, updates can happen many times per second to a channel, and monitors can be tens
22 * of megabytes (or more). Updates can be as small as a few hundred bytes.
24 * Note that monitors written with `MonitorUpdatingPersister` are _not_ backward-compatible with
25 * the default [`KVStore`]'s [`Persist`] implementation. They have a prepended byte sequence,
26 * [`MONITOR_UPDATING_PERSISTER_PREPEND_SENTINEL`], applied to prevent deserialization with other
27 * persisters. This is because monitors written by this struct _may_ have unapplied updates. In
28 * order to downgrade, you must ensure that all updates are applied to the monitor, and remove the
33 * Monitors are stored by implementing the [`Persist`] trait, which has two functions:
35 * - [`Persist::persist_new_channel`], which persists whole [`ChannelMonitor`]s.
36 * - [`Persist::update_persisted_channel`], which persists only a [`ChannelMonitorUpdate`]
38 * Whole [`ChannelMonitor`]s are stored in the [`CHANNEL_MONITOR_PERSISTENCE_PRIMARY_NAMESPACE`],
39 * using the familiar encoding of an [`OutPoint`] (for example, `[SOME-64-CHAR-HEX-STRING]_1`).
41 * Each [`ChannelMonitorUpdate`] is stored in a dynamic secondary namespace, as follows:
43 * - primary namespace: [`CHANNEL_MONITOR_UPDATE_PERSISTENCE_PRIMARY_NAMESPACE`]
44 * - secondary namespace: [the monitor's encoded outpoint name]
46 * Under that secondary namespace, each update is stored with a number string, like `21`, which
47 * represents its `update_id` value.
49 * For example, consider this channel, named for its transaction ID and index, or [`OutPoint`]:
51 * - Transaction ID: `deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef`
54 * Full channel monitors would be stored at a single key:
56 * `[CHANNEL_MONITOR_PERSISTENCE_PRIMARY_NAMESPACE]/deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef_1`
58 * Updates would be stored as follows (with `/` delimiting primary_namespace/secondary_namespace/key):
61 * [CHANNEL_MONITOR_UPDATE_PERSISTENCE_PRIMARY_NAMESPACE]/deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef_1/1
62 * [CHANNEL_MONITOR_UPDATE_PERSISTENCE_PRIMARY_NAMESPACE]/deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef_1/2
63 * [CHANNEL_MONITOR_UPDATE_PERSISTENCE_PRIMARY_NAMESPACE]/deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef_1/3
67 * # Reading channel state from storage
69 * Channel state can be reconstructed by calling
70 * [`MonitorUpdatingPersister::read_all_channel_monitors_with_updates`]. Alternatively, users can
71 * list channel monitors themselves and load channels individually using
72 * [`MonitorUpdatingPersister::read_channel_monitor_with_updates`].
74 * ## EXTREMELY IMPORTANT
76 * It is extremely important that your [`KVStore::read`] implementation uses the
77 * [`io::ErrorKind::NotFound`] variant correctly: that is, when a file is not found, and _only_ in
78 * that circumstance (not when there is really a permissions error, for example). This is because
79 * neither channel monitor reading function lists updates. Instead, either reads the monitor, and
80 * using its stored `update_id`, synthesizes update storage keys, and tries them in sequence until
81 * one is not found. All _other_ errors will be bubbled up in the function's [`Result`].
83 * # Pruning stale channel updates
85 * Stale updates are pruned when a full monitor is written. The old monitor is first read, and if
86 * that succeeds, updates in the range between the old and new monitors are deleted. The `lazy`
87 * flag is used on the [`KVStore::remove`] method, so there are no guarantees that the deletions
88 * will complete. However, stale updates are not a problem for data integrity, since updates are
89 * only read that are higher than the stored [`ChannelMonitor`]'s `update_id`.
91 * If you have many stale updates stored (such as after a crash with pending lazy deletes), and
92 * would like to get rid of them, consider using the
93 * [`MonitorUpdatingPersister::cleanup_stale_updates`] function.
95 @SuppressWarnings("unchecked") // We correctly assign various generic arrays
96 public class MonitorUpdatingPersister extends CommonBase {
97 MonitorUpdatingPersister(Object _dummy, long ptr) { super(ptr); }
98 @Override @SuppressWarnings("deprecation")
99 protected void finalize() throws Throwable {
101 if (ptr != 0) { bindings.MonitorUpdatingPersister_free(ptr); }
105 * Constructs a new [`MonitorUpdatingPersister`].
107 * The `maximum_pending_updates` parameter controls how many updates may be stored before a
108 * [`MonitorUpdatingPersister`] consolidates updates by writing a full monitor. Note that
109 * consolidation will frequently occur with fewer updates than what you set here; this number
110 * is merely the maximum that may be stored. When setting this value, consider that for higher
111 * values of `maximum_pending_updates`:
113 * - [`MonitorUpdatingPersister`] will tend to write more [`ChannelMonitorUpdate`]s than
114 * [`ChannelMonitor`]s, approaching one [`ChannelMonitor`] write for every
115 * `maximum_pending_updates` [`ChannelMonitorUpdate`]s.
116 * - [`MonitorUpdatingPersister`] will issue deletes differently. Lazy deletes will come in
117 * \"waves\" for each [`ChannelMonitor`] write. A larger `maximum_pending_updates` means bigger,
118 * less frequent \"waves.\"
119 * - [`MonitorUpdatingPersister`] will potentially have more listing to do if you need to run
120 * [`MonitorUpdatingPersister::cleanup_stale_updates`].
122 public static MonitorUpdatingPersister of(org.ldk.structs.KVStore kv_store, org.ldk.structs.Logger logger, long maximum_pending_updates, org.ldk.structs.EntropySource entropy_source, org.ldk.structs.SignerProvider signer_provider) {
123 long ret = bindings.MonitorUpdatingPersister_new(kv_store.ptr, logger.ptr, maximum_pending_updates, entropy_source.ptr, signer_provider.ptr);
124 Reference.reachabilityFence(kv_store);
125 Reference.reachabilityFence(logger);
126 Reference.reachabilityFence(maximum_pending_updates);
127 Reference.reachabilityFence(entropy_source);
128 Reference.reachabilityFence(signer_provider);
129 if (ret >= 0 && ret <= 4096) { return null; }
130 org.ldk.structs.MonitorUpdatingPersister ret_hu_conv = null; if (ret < 0 || ret > 4096) { ret_hu_conv = new org.ldk.structs.MonitorUpdatingPersister(null, ret); }
131 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(ret_hu_conv); };
132 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(kv_store); };
133 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(logger); };
134 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(entropy_source); };
135 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(signer_provider); };
140 * Reads all stored channel monitors, along with any stored updates for them.
142 * It is extremely important that your [`KVStore::read`] implementation uses the
143 * [`io::ErrorKind::NotFound`] variant correctly. For more information, please see the
144 * documentation for [`MonitorUpdatingPersister`].
146 public Result_CVec_C2Tuple_ThirtyTwoBytesChannelMonitorZZIOErrorZ read_all_channel_monitors_with_updates(org.ldk.structs.BroadcasterInterface broadcaster, org.ldk.structs.FeeEstimator fee_estimator) {
147 long ret = bindings.MonitorUpdatingPersister_read_all_channel_monitors_with_updates(this.ptr, broadcaster.ptr, fee_estimator.ptr);
148 Reference.reachabilityFence(this);
149 Reference.reachabilityFence(broadcaster);
150 Reference.reachabilityFence(fee_estimator);
151 if (ret >= 0 && ret <= 4096) { return null; }
152 Result_CVec_C2Tuple_ThirtyTwoBytesChannelMonitorZZIOErrorZ ret_hu_conv = Result_CVec_C2Tuple_ThirtyTwoBytesChannelMonitorZZIOErrorZ.constr_from_ptr(ret);
153 if (this != null) { this.ptrs_to.add(broadcaster); };
154 if (this != null) { this.ptrs_to.add(fee_estimator); };
159 * Read a single channel monitor, along with any stored updates for it.
161 * It is extremely important that your [`KVStore::read`] implementation uses the
162 * [`io::ErrorKind::NotFound`] variant correctly. For more information, please see the
163 * documentation for [`MonitorUpdatingPersister`].
165 * For `monitor_key`, channel storage keys be the channel's transaction ID and index, or
166 * [`OutPoint`], with an underscore `_` between them. For example, given:
168 * - Transaction ID: `deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef`
171 * The correct `monitor_key` would be:
172 * `deadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeefdeadbeef_1`
174 * Loading a large number of monitors will be faster if done in parallel. You can use this
175 * function to accomplish this. Take care to limit the number of parallel readers.
177 public Result_C2Tuple_ThirtyTwoBytesChannelMonitorZIOErrorZ read_channel_monitor_with_updates(org.ldk.structs.BroadcasterInterface broadcaster, org.ldk.structs.FeeEstimator fee_estimator, java.lang.String monitor_key) {
178 long ret = bindings.MonitorUpdatingPersister_read_channel_monitor_with_updates(this.ptr, broadcaster.ptr, fee_estimator.ptr, monitor_key);
179 Reference.reachabilityFence(this);
180 Reference.reachabilityFence(broadcaster);
181 Reference.reachabilityFence(fee_estimator);
182 Reference.reachabilityFence(monitor_key);
183 if (ret >= 0 && ret <= 4096) { return null; }
184 Result_C2Tuple_ThirtyTwoBytesChannelMonitorZIOErrorZ ret_hu_conv = Result_C2Tuple_ThirtyTwoBytesChannelMonitorZIOErrorZ.constr_from_ptr(ret);
185 if (this != null) { this.ptrs_to.add(broadcaster); };
186 if (this != null) { this.ptrs_to.add(fee_estimator); };
191 * Cleans up stale updates for all monitors.
193 * This function works by first listing all monitors, and then for each of them, listing all
194 * updates. The updates that have an `update_id` less than or equal to than the stored monitor
195 * are deleted. The deletion can either be lazy or non-lazy based on the `lazy` flag; this will
196 * be passed to [`KVStore::remove`].
198 public Result_NoneIOErrorZ cleanup_stale_updates(boolean lazy) {
199 long ret = bindings.MonitorUpdatingPersister_cleanup_stale_updates(this.ptr, lazy);
200 Reference.reachabilityFence(this);
201 Reference.reachabilityFence(lazy);
202 if (ret >= 0 && ret <= 4096) { return null; }
203 Result_NoneIOErrorZ ret_hu_conv = Result_NoneIOErrorZ.constr_from_ptr(ret);
208 * Constructs a new Persist which calls the relevant methods on this_arg.
209 * This copies the `inner` pointer in this_arg and thus the returned Persist must be freed before this_arg is
211 public Persist as_Persist() {
212 long ret = bindings.MonitorUpdatingPersister_as_Persist(this.ptr);
213 Reference.reachabilityFence(this);
214 if (ret >= 0 && ret <= 4096) { return null; }
215 Persist ret_hu_conv = new Persist(null, ret);
216 if (ret_hu_conv != null) { ret_hu_conv.ptrs_to.add(this); };