Bug #11761530: 54035: MSYQLD_SAFE BEHAVIOUR CONSUMES TOO MUCH CPU IF IT IS UNABLE TO RESTART MY

Added a trottling mechanism to mysqld_safe.
On every 5 sub-second restarts it will wait for a second before attempting again.
Note that this requires 'sleep' and 'date' command line utilities.
The code would adjust and will do nothing if any of them is not present.

1 files changed, 38 insertions, 0 deletions

diff --git a/scripts/mysqld_safe.sh b/scripts/mysqld_safe.sh
index 0d2045a65a6..37e0e351ccd 100644
--- a/scripts/mysqld_safe.sh
+++ b/scripts/mysqld_safe.sh
@@ -740,17 +740,55 @@ cmd="$cmd $args"
 test -n "$NOHUP_NICENESS" && cmd="$cmd < /dev/null"
 
 log_notice "Starting $MYSQLD daemon with databases from $DATADIR"
+
+# variable to track the current number of "fast" (a.k.a. subsecond) restarts
+fast_restart=0
+# maximum number of restarts before trottling kicks in
+max_fast_restarts=5
+# flag whether a usable sleep command exists
+have_sleep=1
+
 while true
 do
   rm -f $safe_mysql_unix_port "$pid_file"	# Some extra safety
 
+  start_time=`date +%M%S`
+
   eval_log_error "$cmd"
 
+  end_time=`date +%M%S`
+
   if test ! -f "$pid_file"		# This is removed if normal shutdown
   then
     break
   fi
 
+
+  # sanity check if time reading is sane and there's sleep
+  if test $end_time -gt 0 -a $have_sleep -gt 0
+  then
+    # throttle down the fast restarts
+    if test $end_time -eq $start_time
+    then
+      fast_restart=`expr $fast_restart + 1`
+      if test $fast_restart -ge $max_fast_restarts
+      then
+        log_notice "The server is respawning too fast. Sleeping for 1 second."
+        sleep 1
+        sleep_state=$?
+        if test $sleep_state -gt 0
+        then
+          log_notice "The server is respawning too fast and no working sleep command. Turning off trottling."
+          have_sleep=0
+        fi
+
+        fast_restart=0
+      fi
+    else
+      fast_restart=0
+    fi
+  fi
+
   if @TARGET_LINUX@ && test $KILL_MYSQLD -eq 1
   then
     # Test if one process was hanging.