Import wiredtiger: 90a2282276ccc25fcc662e6ebef6fbc91d0ff0af from branch mongodb-4.4

ref: 4cf221a61f..90a2282276
for: 4.4.0-rc7

WT-6042       Changes to checkpoint04 to remove or debug failures on Windows
WT-6165       Fix cursor operation hung on WT_PREPARE_CONFLICT
WT-6173       Change __wt_time_window_ functions to macros
WT-6174       Fix dhandle->session_inuse counter leak in __evict_walk
WT-6195       Ensure that we don't perform rollback to stable without a history store
WT-6214       format.sh must disassociate from the child process

3 files changed, 76 insertions, 27 deletions

diff --git a/src/third_party/wiredtiger/test/format/format.i b/src/third_party/wiredtiger/test/format/format.i
index 661dd096ae9..e1634ff13d3 100644
--- a/src/third_party/wiredtiger/test/format/format.i
+++ b/src/third_party/wiredtiger/test/format/format.i
@@ -34,27 +34,53 @@ static inline int
 read_op(WT_CURSOR *cursor, read_operation op, int *exactp)
 {
     WT_DECL_RET;
+    uint64_t start, now;
 
     /*
      * Read operations wait out prepare-conflicts. (As part of the snapshot isolation checks, we
      * repeat reads that succeeded before, they should be repeatable.)
      */
+    __wt_seconds(NULL, &start);
     switch (op) {
     case NEXT:
-        while ((ret = cursor->next(cursor)) == WT_PREPARE_CONFLICT)
+        while ((ret = cursor->next(cursor)) == WT_PREPARE_CONFLICT) {
             __wt_yield();
+
+            /* Ignore clock reset. */
+            __wt_seconds(NULL, &now);
+            testutil_assertfmt(now < start || now - start < 60,
+              "%s: timed out with prepare-conflict", "WT_CURSOR.next");
+        }
         break;
     case PREV:
-        while ((ret = cursor->prev(cursor)) == WT_PREPARE_CONFLICT)
+        while ((ret = cursor->prev(cursor)) == WT_PREPARE_CONFLICT) {
             __wt_yield();
+
+            /* Ignore clock reset. */
+            __wt_seconds(NULL, &now);
+            testutil_assertfmt(now < start || now - start < 60,
+              "%s: timed out with prepare-conflict", "WT_CURSOR.prev");
+        }
         break;
     case SEARCH:
-        while ((ret = cursor->search(cursor)) == WT_PREPARE_CONFLICT)
+        while ((ret = cursor->search(cursor)) == WT_PREPARE_CONFLICT) {
             __wt_yield();
+
+            /* Ignore clock reset. */
+            __wt_seconds(NULL, &now);
+            testutil_assertfmt(now < start || now - start < 60,
+              "%s: timed out with prepare-conflict", "WT_CURSOR.search");
+        }
         break;
     case SEARCH_NEAR:
-        while ((ret = cursor->search_near(cursor, exactp)) == WT_PREPARE_CONFLICT)
+        while ((ret = cursor->search_near(cursor, exactp)) == WT_PREPARE_CONFLICT) {
             __wt_yield();
+
+            /* Ignore clock reset. */
+            __wt_seconds(NULL, &now);
+            testutil_assertfmt(now < start || now - start < 60,
+              "%s: timed out with prepare-conflict", "WT_CURSOR.search_near");
+        }
         break;
     }
     return (ret);
diff --git a/src/third_party/wiredtiger/test/format/format.sh b/src/third_party/wiredtiger/test/format/format.sh
index 0919eab9b03..7f853819471 100755
--- a/src/third_party/wiredtiger/test/format/format.sh
+++ b/src/third_party/wiredtiger/test/format/format.sh
@@ -78,7 +78,6 @@ timing_stress_split_test=0
 total_jobs=0
 verbose=0
 format_binary="./t"
-env_var=""
 
 while :; do
 	case "$1" in
@@ -92,7 +91,7 @@ while :; do
 		config="$2"
 		shift ; shift ;;
 	-e)
-		env_var="$2"
+		export "$2"
 		shift ; shift ;;
 	-E)
 		skip_errors=1
@@ -448,10 +447,11 @@ format()
 	# continue to run.
 	# Run format in its own session so child processes are in their own process gorups
 	# and we can individually terminate (and clean up) running jobs and their children.
-	eval $env_var setsid $cmd > $log 2>&1 &
+	nohup setsid $cmd > $log 2>&1 &
 
-	# Check for setsid command failed execution, and forcibly quit.
-	# The RUNDIR is not successfully created in this failure type.
+	# Check for setsid command failed execution, and forcibly quit (setsid exits 0 if the
+	# command execution fails so we can't check the exit status). The RUNDIR directory is
+	# not created in this failure type, check the log file explicitly.
 	sleep 1
 	grep -E -i 'setsid: failed to execute' $log > /dev/null && {
 		failure=$(($failure + 1))
diff --git a/src/third_party/wiredtiger/test/suite/test_checkpoint04.py b/src/third_party/wiredtiger/test/suite/test_checkpoint04.py
index eb6f0313d61..b2af62ce4a1 100755
--- a/src/third_party/wiredtiger/test/suite/test_checkpoint04.py
+++ b/src/third_party/wiredtiger/test/suite/test_checkpoint04.py
@@ -50,13 +50,14 @@ class test_checkpoint04(wttest.WiredTigerTestCase):
             tables[uri] = ds
         return tables
 
-    def add_updates(self, uri, ds, value, nrows, ts):
+    def add_updates(self, uri, ds, value, nrows):
         session = self.session
         cursor = session.open_cursor(uri)
+        self.pr('update: ' + uri + ' for ' + str(nrows) + ' rows')
         for i in range(0, nrows):
             session.begin_transaction()
             cursor[ds.key(i)] = value
-            session.commit_transaction('commit_timestamp=' + timestamp_str(ts))
+            session.commit_transaction()
         cursor.close()
 
     def get_stat(self, stat):
@@ -69,14 +70,11 @@ class test_checkpoint04(wttest.WiredTigerTestCase):
         nrows = 100
         ntables = 5
 
-        self.conn.set_timestamp('oldest_timestamp=' + timestamp_str(10) +
-            ',stable_timestamp=' + timestamp_str(10))
-
         # Create many tables and perform many updates so our checkpoint stats are populated.
         value = "wired" * 100
         tables = self.create_tables(ntables)
         for uri, ds in tables.items():
-            self.add_updates(uri, ds, value, nrows, 20)
+            self.add_updates(uri, ds, value, nrows)
 
         # Perform a checkpoint.
         self.session.checkpoint()
@@ -85,23 +83,48 @@ class test_checkpoint04(wttest.WiredTigerTestCase):
         value = "tiger" * 100
         tables = self.create_tables(ntables)
         for uri, ds in tables.items():
-            self.add_updates(uri, ds, value, nrows, 30)
+            self.add_updates(uri, ds, value, nrows)
 
         # Perform a checkpoint.
         self.session.checkpoint()
 
         # Check the statistics.
-        self.assertEqual(self.get_stat(stat.conn.txn_checkpoint), 2)
-        self.assertEqual(self.get_stat(stat.conn.txn_checkpoint_running), 0)
-        self.assertEqual(self.get_stat(stat.conn.txn_checkpoint_prep_running), 0)
-        self.assertLess(self.get_stat(stat.conn.txn_checkpoint_prep_min),
-            self.get_stat(stat.conn.txn_checkpoint_time_min))
-        self.assertLess(self.get_stat(stat.conn.txn_checkpoint_prep_max),
-            self.get_stat(stat.conn.txn_checkpoint_time_max))
-        self.assertLess(self.get_stat(stat.conn.txn_checkpoint_prep_recent),
-            self.get_stat(stat.conn.txn_checkpoint_time_recent))
-        self.assertLess(self.get_stat(stat.conn.txn_checkpoint_prep_total),
-            self.get_stat(stat.conn.txn_checkpoint_time_total))
+        # Set them into a variable so that we can print them all out. We've had a failure
+        # on Windows that is very difficult to reproduce so collect what info we can.
+        num_ckpt = self.get_stat(stat.conn.txn_checkpoint)
+        self.pr('txn_checkpoint, number of checkpoints ' + str(num_ckpt))
+        running = self.get_stat(stat.conn.txn_checkpoint_running)
+        self.pr('txn_checkpoint_running ' + str(running))
+        prep_running = self.get_stat(stat.conn.txn_checkpoint_prep_running)
+        self.pr('txn_checkpoint_prep_running ' + str(prep_running))
+
+        prep_min = self.get_stat(stat.conn.txn_checkpoint_prep_min)
+        self.pr('txn_checkpoint_prep_min ' + str(prep_min))
+        time_min = self.get_stat(stat.conn.txn_checkpoint_time_min)
+        self.pr('txn_checkpoint_time_min ' + str(time_min))
+
+        prep_max = self.get_stat(stat.conn.txn_checkpoint_prep_max)
+        self.pr('txn_checkpoint_prep_max ' + str(prep_max))
+        time_max = self.get_stat(stat.conn.txn_checkpoint_time_max)
+        self.pr('txn_checkpoint_time_max ' + str(time_max))
+
+        prep_recent = self.get_stat(stat.conn.txn_checkpoint_prep_recent)
+        self.pr('txn_checkpoint_prep_recent ' + str(prep_recent))
+        time_recent = self.get_stat(stat.conn.txn_checkpoint_time_recent)
+        self.pr('txn_checkpoint_time_recent ' + str(time_recent))
+
+        prep_total = self.get_stat(stat.conn.txn_checkpoint_prep_total)
+        self.pr('txn_checkpoint_prep_total ' + str(prep_total))
+        time_total = self.get_stat(stat.conn.txn_checkpoint_time_total)
+        self.pr('txn_checkpoint_time_total ' + str(time_total))
+
+        self.assertEqual(num_ckpt, 2)
+        self.assertEqual(running, 0)
+        self.assertEqual(prep_running, 0)
+        self.assertLess(prep_min, time_min)
+        self.assertLess(prep_max, time_max)
+        self.assertLess(prep_recent, time_recent)
+        self.assertLess(prep_total, time_total)
 
 if __name__ == '__main__':
     wttest.run()