MDEV-22169 Recovery fails after failing to insert into mlog_init

In a multi-batch recovery, we must ensure that INIT_PAGE and
especially the MDEV-15528 FREE_PAGE records will be taken
properly into account.

Writing a FREE_PAGE record gives the server permission to omit
a page write. If recovery insists on applying log to a page
whose page flush has been omitted, then the consistency checks
in the application of high-level redo log records (appending
an undo log record, inserting or deleting an index record)
will likely fail.

mlog_init_t::add(): Return whether the state was changed.

mlog_init_t::will_avoid_read(): Determine whether a page read
will be avoided and whether older log records can be safely
skipped.

recv_sys_t::parse(): Even if store==STORE_NO, process the records
INIT_PAGE and FREE_PAGE. While processing them, we can delete older
redo log records for the page. If store!=STORE_NO, we can directly
skip redo log recods of other types if mlog_init indicates that the
page will be freed or initialized by at a later LSN.

This fix was developed in cooperation with
Thirunarayanan Balathandayuthapani.

1 files changed, 40 insertions, 12 deletions

diff --git a/storage/innobase/log/log0recv.cc b/storage/innobase/log/log0recv.cc
index 3533cd3058b..cb1b8b31dcb 100644
--- a/storage/innobase/log/log0recv.cc
+++ b/storage/innobase/log/log0recv.cc
@@ -107,11 +107,11 @@ private:
 
   /** @return start of the log records */
   byte *begin() { return reinterpret_cast<byte*>(&len + 1); }
-  /** @return start of the log records */
-  const byte *begin() const { return const_cast<log_phys_t*>(this)->begin(); }
   /** @return end of the log records */
   byte *end() { byte *e= begin() + len; ut_ad(!*e); return e; }
 public:
+  /** @return start of the log records */
+  const byte *begin() const { return const_cast<log_phys_t*>(this)->begin(); }
   /** @return end of the log records */
   const byte *end() const { return const_cast<log_phys_t*>(this)->end(); }
 
@@ -598,17 +598,19 @@ private:
 public:
 	/** Record that a page will be initialized by the redo log.
 	@param[in]	page_id		page identifier
-	@param[in]	lsn		log sequence number */
-	void add(const page_id_t page_id, lsn_t lsn)
+	@param[in]	lsn		log sequence number
+	@return whether the state was changed */
+	bool add(const page_id_t page_id, lsn_t lsn)
 	{
 		ut_ad(mutex_own(&recv_sys.mutex));
 		const init init = { lsn, false };
 		std::pair<map::iterator, bool> p = inits.insert(
 			map::value_type(page_id, init));
 		ut_ad(!p.first->second.created);
-		if (!p.second && p.first->second.lsn < init.lsn) {
-			p.first->second = init;
-		}
+		if (p.second) return true;
+		if (p.first->second.lsn >= init.lsn) return false;
+		p.first->second = init;
+		return true;
 	}
 
 	/** Get the last stored lsn of the page id and its respective
@@ -623,6 +625,17 @@ public:
 		return inits.find(page_id)->second;
 	}
 
+	/** Determine if a page will be initialized or freed after a time.
+	@param page_id      page identifier
+	@param lsn          log sequence number
+	@return whether page_id will be freed or initialized after lsn */
+	bool will_avoid_read(page_id_t page_id, lsn_t lsn) const
+	{
+		ut_ad(mutex_own(&recv_sys.mutex));
+		auto i= inits.find(page_id);
+		return i != inits.end() && i->second.lsn > lsn;
+	}
+
 	/** At the end of each recovery batch, reset the 'created' flags. */
 	void reset()
 	{
@@ -1938,7 +1951,7 @@ same_page:
 
     if (got_page_op)
     {
-      ut_d(const page_id_t id(space_id, page_no));
+      const page_id_t id(space_id, page_no);
       ut_d(if ((b & 0x70) == INIT_PAGE) freed.erase(id));
       ut_ad(freed.find(id) == freed.end());
       switch (b & 0x70) {
@@ -2059,16 +2072,31 @@ same_page:
         ut_ad(modified.emplace(id).second || (b & 0x70) != INIT_PAGE);
       }
 #endif
+      const bool is_init= (b & 0x70) <= INIT_PAGE;
       switch (store) {
-      case STORE_NO:
-        continue;
       case STORE_IF_EXISTS:
         if (!fil_space_get_size(space_id))
           continue;
         /* fall through */
       case STORE_YES:
-        add(page_id_t(space_id, page_no), start_lsn, end_lsn, recs,
-            static_cast<size_t>(l + rlen - recs));
+        if (is_init || !mlog_init.will_avoid_read(id, start_lsn))
+          add(id, start_lsn, end_lsn, recs,
+              static_cast<size_t>(l + rlen - recs));
+        continue;
+      case STORE_NO:
+        if (!is_init)
+          continue;
+        map::iterator i= pages.find(id);
+        if (i == pages.end())
+          continue;
+        if ((*static_cast<const log_phys_t*>(*i->second.log.begin())->begin() &
+             0x70) <= INIT_PAGE)
+        {
+          ut_ad(i->second.state == page_recv_t::RECV_WILL_NOT_READ);
+          continue;
+        }
+        pages.erase(i);
+        mlog_init.add(id, start_lsn);
       }
     }
 #if 1 /* MDEV-14425 FIXME: this must be in the checkpoint file only! */