MDEV-25880 part 2: Improving reliability of the SST scripts

Additional improvements aimed at improving operational
reliability of the SST scripts:

1) Script need to give rsync and stunnel a short time to
   terminate after "kill -9" before the first PID check
   using ps utility;
2) The temporary file used to create the binlog index could
   sometimes remain in the data directory if tar failed and
   then may be reused without being cleaned up (the next
   time when SST was run) - now it's fixed;
3) The temporary file used to build the binlog index is now
   created using mktemp and, if this variable is present in
   the configuration file, in tmpdir;
4) Checking the secret tag in SST via rsync is made faster
   and does not require creating a temporary file, which
   could remain in the data directory in case of failure;
5) Added "-F" option to grep to check the tag when using
   mariabackup/xtrabackup-v2 - to avoid possible collisions
   in case of special characters in the tag value (unlikely
   scenario, but the new check is more reliable).

4 files changed, 49 insertions, 34 deletions

diff --git a/scripts/wsrep_sst_common.sh b/scripts/wsrep_sst_common.sh
index c2f31b2818d..4dedecb439f 100644
--- a/scripts/wsrep_sst_common.sh
+++ b/scripts/wsrep_sst_common.sh
@@ -1223,7 +1223,7 @@ check_pid()
 #
 cleanup_pid()
 {
-    local pid="$1"
+    local pid=$1
     local pid_file="${2:-}"
     local config="${3:-}"
 
@@ -1241,8 +1241,9 @@ cleanup_pid()
                            round=8
                            force=1
                            kill -9 $pid >/dev/null 2>&1
+                           sleep 0.5
                        else
-                           return 1;
+                           return 1
                        fi
                    fi
                 done
@@ -1254,7 +1255,7 @@ cleanup_pid()
     fi
 
     [ -n "$pid_file" ] && [ -f "$pid_file" ] && rm -f "$pid_file"
-    [ -n "$config" ]   && [ -f "$config"   ] && rm -f "$config"
+    [ -n "$config" ]   && [ -f "$config" ]   && rm -f "$config"
 
     return 0
 }
diff --git a/scripts/wsrep_sst_mariabackup.sh b/scripts/wsrep_sst_mariabackup.sh
index 7f97d9e8dea..339a8fcf4a5 100644
--- a/scripts/wsrep_sst_mariabackup.sh
+++ b/scripts/wsrep_sst_mariabackup.sh
@@ -741,15 +741,15 @@ recv_joiner()
         fi
 
         # check donor supplied secret
-        SECRET=$(grep -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
+        SECRET=$(grep -F -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
         if [ "$SECRET" != "$MY_SECRET" ]; then
             wsrep_log_error "Donor does not know my secret!"
             wsrep_log_info "Donor:'$SECRET', my:'$MY_SECRET'"
             exit 32
         fi
 
-        # remove secret from magic file
-        grep -v -- "$SECRET_TAG " "$MAGIC_FILE" > "$MAGIC_FILE.new"
+        # remove secret from the magic file
+        grep -v -F -- "$SECRET_TAG " "$MAGIC_FILE" > "$MAGIC_FILE.new"
         mv "$MAGIC_FILE.new" "$MAGIC_FILE"
     fi
 }
diff --git a/scripts/wsrep_sst_rsync.sh b/scripts/wsrep_sst_rsync.sh
index a602af79af0..fc9f5017937 100644
--- a/scripts/wsrep_sst_rsync.sh
+++ b/scripts/wsrep_sst_rsync.sh
@@ -429,7 +429,7 @@ EOF
             exit 255 # unknown error
         fi
 
-        # second, we transfer InnoDB log files
+        # second, we transfer InnoDB and Aria log files
         rsync ${STUNNEL:+--rsh="$STUNNEL"} \
               --owner --group --perms --links --specials \
               --ignore-times --inplace --dirs --delete --quiet \
@@ -504,20 +504,20 @@ then
 
     SST_PID="$WSREP_SST_OPT_DATA/wsrep_rsync_sst.pid"
 
-    # give some time for lingering stunnel from previous SST to complete
+    # give some time for previous SST to complete:
     check_round=0
     while check_pid "$SST_PID" 0
     do
-        wsrep_log_info "previous SST not completed, waiting for it to exit"
+        wsrep_log_info "previous SST is not completed, waiting for it to exit"
         check_round=$(( check_round + 1 ))
         if [ $check_round -eq 10 ]; then
-            wsrep_log_error "SST script already running."
+            wsrep_log_error "previous SST script still running."
             exit 114 # EALREADY
         fi
         sleep 1
     done
 
-    # give some time for lingering stunnel from previous SST to complete
+    # give some time for stunnel from the previous SST to complete:
     check_round=0
     while check_pid "$STUNNEL_PID" 1
     do
@@ -534,7 +534,7 @@ then
     RSYNC_PID="$WSREP_SST_OPT_DATA/$MODULE.pid"
     RSYNC_CONF="$WSREP_SST_OPT_DATA/$MODULE.conf"
 
-    # give some time for lingering rsync from previous SST to complete
+    # give some time for rsync from the previous SST to complete:
     check_round=0
     while check_pid "$RSYNC_PID" 1
     do
@@ -711,35 +711,49 @@ EOF
 
             # Clean up old binlog files first
             rm -f "$BINLOG_FILENAME".[0-9]*
-            [ -f "$binlog_index" ] && rm "$binlog_index"
+            [ -f "$binlog_index" ] && rm -f "$binlog_index"
+
+            # Create a temporary file:
+            tmpdir=$(parse_cnf '--mysqld|sst' 'tmpdir')
+            if [ -z "$tmpdir" ]; then
+               tmpfile="$(mktemp)"
+            else
+               tmpfile=$(mktemp "--tmpdir=$tmpdir")
+            fi
 
             wsrep_log_info "Extracting binlog files:"
-            tar -xvf "$BINLOG_TAR_FILE" >> _binlog_tmp_files_$!
+            if ! tar -xvf "$BINLOG_TAR_FILE" > "$tmpfile"; then
+                wsrep_log_error "Error unpacking tar file with binlog files"
+                rm -f "$tmpfile"
+                exit 32
+            fi
+
+            # Rebuild binlog index:
             while read bin_file; do
                 echo "$BINLOG_DIRNAME/$bin_file" >> "$binlog_index"
-            done < _binlog_tmp_files_$!
-            rm -f _binlog_tmp_files_$!
+            done < "$tmpfile"
+            rm -f "$tmpfile"
 
             cd "$OLD_PWD"
         fi
     fi
 
-    if [ -r "$MAGIC_FILE" ]
-    then
-        # check donor supplied secret
-        SECRET=$(grep -F -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
-        if [ "$SECRET" != "$MY_SECRET" ]; then
-            wsrep_log_error "Donor does not know my secret!"
-            wsrep_log_info "Donor:'$SECRET', my:'$MY_SECRET'"
-            exit 32
+    if [ -r "$MAGIC_FILE" ]; then
+        if [ -n "$MY_SECRET" ]; then
+            # check donor supplied secret
+            SECRET=$(grep -F -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
+            if [ "$SECRET" != "$MY_SECRET" ]; then
+                wsrep_log_error "Donor does not know my secret!"
+                wsrep_log_info "Donor:'$SECRET', my:'$MY_SECRET'"
+                exit 32
+            fi
+            # remove secret from the magic file, and output
+            # the UUID:seqno & wsrep_gtid_domain_id:
+            grep -v -F -- "$SECRET_TAG " "$MAGIC_FILE"
+        else
+            # Output the UUID:seqno and wsrep_gtid_domain_id:
+            cat "$MAGIC_FILE"
         fi
-
-        # remove secret from magic file
-        grep -v -F -- "$SECRET_TAG " "$MAGIC_FILE" > "$MAGIC_FILE.new"
-
-        mv "$MAGIC_FILE.new" "$MAGIC_FILE"
-        # UUID:seqno & wsrep_gtid_domain_id is received here.
-        cat "$MAGIC_FILE" # Output : UUID:seqno wsrep_gtid_domain_id
     else
         # this message should cause joiner to abort
         echo "rsync process ended without creating '$MAGIC_FILE'"
diff --git a/scripts/wsrep_sst_xtrabackup-v2.sh b/scripts/wsrep_sst_xtrabackup-v2.sh
index 24bff12219d..d76dc346a82 100644
--- a/scripts/wsrep_sst_xtrabackup-v2.sh
+++ b/scripts/wsrep_sst_xtrabackup-v2.sh
@@ -750,15 +750,15 @@ recv_joiner()
         fi
 
         # check donor supplied secret
-        SECRET=$(grep -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
+        SECRET=$(grep -F -- "$SECRET_TAG " "$MAGIC_FILE" 2>/dev/null | cut -d ' ' -f 2)
         if [ "$SECRET" != "$MY_SECRET" ]; then
             wsrep_log_error "Donor does not know my secret!"
             wsrep_log_info "Donor:'$SECRET', my:'$MY_SECRET'"
             exit 32
         fi
 
-        # remove secret from magic file
-        grep -v -- "$SECRET_TAG " "$MAGIC_FILE" > "$MAGIC_FILE.new"
+        # remove secret from the magic file
+        grep -v -F -- "$SECRET_TAG " "$MAGIC_FILE" > "$MAGIC_FILE.new"
         mv "$MAGIC_FILE.new" "$MAGIC_FILE"
     fi
 }