[Tarantool-patches] [PATCH v3 2/4] recovery: allow to ignore rows coming from a certain instance

[Serge Petrenko]

Prerequisite #4739
---
 src/box/box.cc      |  2 +-
 src/box/recovery.cc | 12 +++++++++++-
 src/box/recovery.h  |  7 ++++++-
 src/box/relay.cc    |  4 ++--
 4 files changed, 20 insertions(+), 5 deletions(-)

diff --git a/src/box/box.cc b/src/box/box.cc
index 0290578b2..a4d823df0 100644
--- a/src/box/box.cc
+++ b/src/box/box.cc
@@ -2171,7 +2171,7 @@ local_recovery(const struct tt_uuid *instance_uuid,
 	struct recovery *recovery;
 	recovery = recovery_new(cfg_gets("wal_dir"),
 				cfg_geti("force_recovery"),
-				checkpoint_vclock);
+				checkpoint_vclock, 0);
 
 	/*
 	 * Make sure we report the actual recovery position
diff --git a/src/box/recovery.cc b/src/box/recovery.cc
index 64aa467b1..e6d6b5744 100644
--- a/src/box/recovery.cc
+++ b/src/box/recovery.cc
@@ -81,7 +81,7 @@
  */
 struct recovery *
 recovery_new(const char *wal_dirname, bool force_recovery,
-	     const struct vclock *vclock)
+	     const struct vclock *vclock, unsigned int id_ignore_map)
 {
 	struct recovery *r = (struct recovery *)
 			calloc(1, sizeof(*r));
@@ -101,6 +101,8 @@ recovery_new(const char *wal_dirname, bool force_recovery,
 
 	vclock_copy(&r->vclock, vclock);
 
+	r->id_ignore_map = id_ignore_map;
+
 	/**
 	 * Avoid scanning WAL dir before we recovered
 	 * the snapshot and know instance UUID - this will
@@ -275,6 +277,14 @@ recover_xlog(struct recovery *r, struct xstream *stream,
 		 * failed row anyway.
 		 */
 		vclock_follow_xrow(&r->vclock, &row);
+		/*
+		 * Do not try to apply rows coming from an
+		 * ignored instance, but still follow their lsns
+		 * to make sure recovery vclock stays the same as
+		 * the one emerging from local recovery.
+		 */
+		if (1 << row.replica_id & r->id_ignore_map)
+			continue;
 		if (xstream_write(stream, &row) == 0) {
 			++row_count;
 			if (row_count % 100000 == 0)
diff --git a/src/box/recovery.h b/src/box/recovery.h
index 6e68abc0b..2e10b7064 100644
--- a/src/box/recovery.h
+++ b/src/box/recovery.h
@@ -50,6 +50,11 @@ struct recovery {
 	/** The WAL cursor we're currently reading/writing from/to. */
 	struct xlog_cursor cursor;
 	struct xdir wal_dir;
+	/**
+	 * A map of replica ids whose changes will be silently
+	 * ignored on recovery.
+	 */
+	unsigned int id_ignore_map;
 	/**
 	 * This fiber is used in local hot standby mode.
 	 * It looks for changes in the wal_dir and applies
@@ -62,7 +67,7 @@ struct recovery {
 
 struct recovery *
 recovery_new(const char *wal_dirname, bool force_recovery,
-	     const struct vclock *vclock);
+	     const struct vclock *vclock, unsigned int id_ignore_map);
 
 void
 recovery_delete(struct recovery *r);
diff --git a/src/box/relay.cc b/src/box/relay.cc
index b89632273..741a09201 100644
--- a/src/box/relay.cc
+++ b/src/box/relay.cc
@@ -355,7 +355,7 @@ relay_final_join(int fd, uint64_t sync, struct vclock *start_vclock,
 	});
 
 	relay->r = recovery_new(cfg_gets("wal_dir"), false,
-			       start_vclock);
+			       start_vclock, 0);
 	vclock_copy(&relay->stop_vclock, stop_vclock);
 
 	int rc = cord_costart(&relay->cord, "final_join",
@@ -701,7 +701,7 @@ relay_subscribe(struct replica *replica, int fd, uint64_t sync,
 
 	vclock_copy(&relay->local_vclock_at_subscribe, &replicaset.vclock);
 	relay->r = recovery_new(cfg_gets("wal_dir"), false,
-			        replica_clock);
+			        replica_clock, 0);
 	vclock_copy(&relay->tx.vclock, replica_clock);
 	relay->version_id = replica_version_id;
 
-- 
2.21.1 (Apple Git-122.3)