PT-2289 - Allow pt-stalk do disable ps-lock-transactions data collection via parameter

- Implemented feature - Started working on the test
2025-09-10 13:11:32 +00:00 · 2025-09-01 18:17:04 +03:00
parent 05b7efcd27
commit 68a4540a3d
3 changed files with 189 additions and 66 deletions
--- a/bin/pt-stalk
+++ b/bin/pt-stalk
@@ -971,8 +971,10 @@ collect_mysql_data_one() {
      fi
   fi

+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "mysqladmin" ]; then
      $CMD_MYSQLADMIN $EXT_ARGV ext -i$OPT_SLEEP_COLLECT -c$cnt >>"$d/$p-mysqladmin" &
      mysqladmin_pid=$!
+   fi

   ps_instrumentation_enabled=$($CMD_MYSQL $EXT_ARGV -e 'SELECT ENABLED FROM performance_schema.setup_instruments WHERE NAME = "transaction";' \
                                      | sed "2q;d" | sed 'y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/')
@@ -1014,18 +1016,25 @@ collect_system_data() {

 collect_mysql_data_loop() {

-
+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "processlist" ]; then
      (echo $ts; $CMD_MYSQL $EXT_ARGV -e "SHOW FULL PROCESSLIST\G") \
      >> "$d/$p-processlist" &
+   fi
+
   (echo $ts; $CMD_MYSQL $EXT_ARGV -e "SELECT * FROM performance_schema.threads\G") \
   >> "$d/$p-threads" &

   if [ "$have_lock_waits_table" ]; then
+      if ! [ "${OPT_SKIP_COLLECTION}" =~ "lock-waits" ]; then
         (echo $ts; lock_waits "$d/lock_waits.running")   >>"$d/$p-lock-waits" &
+      fi
+      if ! [ "${OPT_SKIP_COLLECTION}" =~ "transactions" ]; then
         (echo $ts; transactions) >>"$d/$p-transactions" &
      fi
+   fi

-   if [ "${mysql_version}" '>' "5.6" ] && [ $ps_instrumentation_enabled == "yes" ]; then
+   if [ "${mysql_version}" '>' "5.6" ] && [ $ps_instrumentation_enabled == "yes" ] \
+      && ! [ "${OPT_SKIP_COLLECTION}" =~ "ps-locks-transactions" ]; then
      ps_locks_transactions "$d/$p-ps-locks-transactions"
   fi

@@ -1252,6 +1261,7 @@ innodb_status() {

   local innostat=""

+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "innodbstatus" ]; then
      $CMD_MYSQL $EXT_ARGV -e "SHOW /*!40100 ENGINE*/ INNODB STATUS\G" \
         >> "$d/$p-innodbstatus$n"
      grep "END OF INNODB" "$d/$p-innodbstatus$n" >/dev/null || {
@@ -1266,6 +1276,7 @@ innodb_status() {
            done
         fi
      }
+   fi
 }

 rocksdb_status() {
@@ -1274,7 +1285,7 @@ rocksdb_status() {
    has_rocksdb=`$CMD_MYSQL $EXT_ARGV -e "SHOW ENGINES" | grep -i 'rocksdb'`
    exit_code=$?

-    if [ $exit_code -eq 0 ]; then
+    if [ $exit_code -eq 0 ] && ! [ "${OPT_SKIP_COLLECTION}" =~ "rocksdbstatus" ]; then
        $CMD_MYSQL $EXT_ARGV -e "SHOW ENGINE ROCKSDB STATUS\G" \
                   >> "$d/$p-rocksdbstatus$n" || rm -f "$d/$p-rocksdbstatus$n"
    fi
@@ -1356,6 +1367,7 @@ collect_mysql_variables() {
   echo -e "\n$sql\n" >> $outfile
   $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile

+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "thread-variables" ]; then
      sql="select * from performance_schema.variables_by_thread order by thread_id, variable_name;"
      echo -e "\n$sql\n" >> $outfile
      $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile
@@ -1367,6 +1379,7 @@ collect_mysql_variables() {
      sql="select * from performance_schema.status_by_thread order by thread_id, variable_name; "
      echo -e "\n$sql\n" >> $outfile
      $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile
+   fi

 }

@@ -1727,6 +1740,18 @@ main() {
      log 'Both options --system-only and --mysql-only specified, collecting only disk-space, hostname, output, and trigger metrics';
   fi

+   if [ "$OPT_SKIP_COLLECTION" ]; then
+      local supported_skips="ps-locks-transactions,thread-variables,innodbstatus,lock-waits,mysqladmin,processlist,rocksdbstatus,transactions"
+      IFS=',' read -ra skips <<< "$OPT_SKIP_COLLECTION"
+      for skip in "${skips[@]}"; do
+         echo "$supported_skips" | grep -q "$skip"
+         if [ $? -ne 0 ]; then
+            log "Invalid --skip-collection value: $skip, exiting."
+            exit 1
+         fi
+      done
+   fi
+
   # Note: $$ is the parent's PID, but we're a child proc.
   # Bash 4 has $BASHPID but we can't rely on that.  Consequently,
   # we don't know our own PID.  See the usage of $! below.
@@ -2391,6 +2416,12 @@ How long to sleep between collection loop cycles.  This is useful with
 C<--no-stalk> to do long collections.  For example, to collect data every
 minute for an hour, specify: C<--no-stalk --run-time 3600 --sleep-collect 60>.

+=item --skip-collection
+
+type: array
+
+A comma-separated list of collection types to skip.  Valid values are: C<"ps-locks-transactions,thread-variables,innodbstatus,lock-waits,mysqladmin,processlist,rocksdbstatus,transactions">.
+
 =item --socket

 short form: -S; type: string
--- a/lib/bash/collect.sh
+++ b/lib/bash/collect.sh
@@ -243,8 +243,10 @@ collect_mysql_data_one() {
   # get and keep a connection to the database; in troubled times
   # the database tends to exceed max_connections, so reconnecting
   # in the loop tends not to work very well.
+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "mysqladmin" ]; then
      $CMD_MYSQLADMIN $EXT_ARGV ext -i$OPT_SLEEP_COLLECT -c$cnt >>"$d/$p-mysqladmin" &
      mysqladmin_pid=$!
+   fi

   ps_instrumentation_enabled=$($CMD_MYSQL $EXT_ARGV -e 'SELECT ENABLED FROM performance_schema.setup_instruments WHERE NAME = "transaction";' \
                                      | sed "2q;d" | sed 'y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/')
@@ -289,18 +291,25 @@ collect_mysql_data_loop() {

   # SHOW FULL PROCESSLIST duplicates information in performance_schema.threads we collecting now
   # Keeping it for backward compatibility and may remove in the future
-
+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "processlist" ]; then
      (echo $ts; $CMD_MYSQL $EXT_ARGV -e "SHOW FULL PROCESSLIST\G") \
      >> "$d/$p-processlist" &
+   fi
+
   (echo $ts; $CMD_MYSQL $EXT_ARGV -e "SELECT * FROM performance_schema.threads\G") \
   >> "$d/$p-threads" &

   if [ "$have_lock_waits_table" ]; then
+      if ! [ "${OPT_SKIP_COLLECTION}" =~ "lock-waits" ]; then
         (echo $ts; lock_waits "$d/lock_waits.running")   >>"$d/$p-lock-waits" &
+      fi
+      if ! [ "${OPT_SKIP_COLLECTION}" =~ "transactions" ]; then
         (echo $ts; transactions) >>"$d/$p-transactions" &
      fi
+   fi

-   if [ "${mysql_version}" '>' "5.6" ] && [ $ps_instrumentation_enabled == "yes" ]; then
+   if [ "${mysql_version}" '>' "5.6" ] && [ $ps_instrumentation_enabled == "yes" ] \
+      && ! [ "${OPT_SKIP_COLLECTION}" =~ "ps-locks-transactions" ]; then
      ps_locks_transactions "$d/$p-ps-locks-transactions"
   fi

@@ -538,6 +547,7 @@ innodb_status() {

   local innostat=""

+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "innodbstatus" ]; then
      $CMD_MYSQL $EXT_ARGV -e "SHOW /*!40100 ENGINE*/ INNODB STATUS\G" \
         >> "$d/$p-innodbstatus$n"
      grep "END OF INNODB" "$d/$p-innodbstatus$n" >/dev/null || {
@@ -552,6 +562,7 @@ innodb_status() {
            done
         fi
      }
+   fi
 }

 rocksdb_status() {
@@ -560,7 +571,7 @@ rocksdb_status() {
    has_rocksdb=`$CMD_MYSQL $EXT_ARGV -e "SHOW ENGINES" | grep -i 'rocksdb'`
    exit_code=$?

-    if [ $exit_code -eq 0 ]; then
+    if [ $exit_code -eq 0 ] && ! [ "${OPT_SKIP_COLLECTION}" =~ "rocksdbstatus" ]; then
        $CMD_MYSQL $EXT_ARGV -e "SHOW ENGINE ROCKSDB STATUS\G" \
                   >> "$d/$p-rocksdbstatus$n" || rm -f "$d/$p-rocksdbstatus$n"
    fi
@@ -646,6 +657,7 @@ collect_mysql_variables() {
   echo -e "\n$sql\n" >> $outfile
   $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile

+   if ! [ "${OPT_SKIP_COLLECTION}" =~ "thread-variables" ]; then
      sql="select * from performance_schema.variables_by_thread order by thread_id, variable_name;"
      echo -e "\n$sql\n" >> $outfile
      $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile
@@ -657,6 +669,7 @@ collect_mysql_variables() {
      sql="select * from performance_schema.status_by_thread order by thread_id, variable_name; "
      echo -e "\n$sql\n" >> $outfile
      $CMD_MYSQL $EXT_ARGV -e "$sql" >> $outfile
+   fi

 }

--- a/t/pt-stalk/pt-2289.t
+++ b/t/pt-stalk/pt-2289.t
@@ -0,0 +1,79 @@
+#!/usr/bin/env perl
+
+BEGIN {
+   die "The PERCONA_TOOLKIT_BRANCH environment variable is not set.\n"
+      unless $ENV{PERCONA_TOOLKIT_BRANCH} && -d $ENV{PERCONA_TOOLKIT_BRANCH};
+   unshift @INC, "$ENV{PERCONA_TOOLKIT_BRANCH}/lib";
+};
+
+use strict;
+use warnings FATAL => 'all';
+use threads;
+use English qw(-no_match_vars);
+use Test::More;
+use Time::HiRes qw(sleep);
+
+use PerconaTest;
+use DSNParser;
+use Sandbox;
+require VersionParser;
+
+my $dp = new DSNParser(opts=>$dsn_opts);
+my $sb = new Sandbox(basedir => '/tmp', DSNParser => $dp);
+my $dbh = $sb->get_dbh_for('source');
+
+if ( !$dbh ) {
+   plan skip_all => 'Cannot connect to sandbox source';
+}
+
+my $cnf      = "/tmp/12345/my.sandbox.cnf";
+my $pid_file = "/tmp/pt-stalk.pid.$PID";
+my $log_file = "/tmp/pt-stalk.log.$PID";
+my $dest     = "/tmp/pt-stalk.collect.$PID";
+my $int_file = "/tmp/pt-stalk-after-interval-sleep";
+my $pid;
+my $output;
+my $retval;
+
+sub cleanup {
+   diag(`rm $pid_file $log_file $int_file 2>/dev/null`);
+   diag(`rm -rf $dest 2>/dev/null`);
+}
+
+# ###########################################################################
+# Test that it collects all data when no --skip-collection is given.
+# ###########################################################################
+
+cleanup();
+
+$retval = system("$trunk/bin/pt-stalk --no-stalk --pid $pid_file --log $log_file --dest $dest --iterations 1 -- --defaults-file=$cnf >$log_file 2>&1");
+
+PerconaTest::wait_until(sub { !-f $pid_file });
+
+is(
+   $retval >> 8,
+   0,
+   "Parent exit 0"
+);
+
+ok(
+   -d $dest,
+   "Creates --dest (collect) dir"
+);
+
+# ps-locks-transactions,thread-variables,innodbstatus,lock-waits,mysqladmin,processlist,rocksdbstatus,transactions
+ok(
+    glob("$dest/*-ps-locks-transactions"),
+    "Collects *-ps-locks-transactions"
+) or diag(`ls $dest`);
+
+
+# #############################################################################
+# Done.
+# #############################################################################
+
+cleanup();
+diag(`rm -rf $dest 2>/dev/null`);
+$sb->wipe_clean($dbh);
+ok($sb->ok(), "Sandbox servers") or BAIL_OUT(__FILE__ . " broke the sandbox");
+done_testing;