pt-heartbeat update uses strings instead of ints

[monty solomon]

Character strings are used as the values for int fields in the UPDATE statement in addition to the value in the WHERE clause.

UPDATE `heartbeat`.`heartbeat` SET ts='2013-10-05T16:51:48.001010', file='bin.000027', position='520884073', relay_master_log_file=NULL, exec_master_log_pos=NULL WHERE server_id='3095733595'\G

CREATE TABLE `heartbeat` (
  `ts` varchar(26) NOT NULL,
  `server_id` int(10) unsigned NOT NULL,
  `file` varchar(255) DEFAULT NULL,
  `position` bigint(20) unsigned DEFAULT NULL,
  `relay_master_log_file` varchar(255) DEFAULT NULL,
  `exec_master_log_pos` bigint(20) unsigned DEFAULT NULL,
  PRIMARY KEY (`server_id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1

[monty solomon (monty+launchpad)]

The pt-heartbeat UPDATE appears in the slow query log

# Query 13: 0.00 QPS, 0.00x concurrency, ID 0xC2C26DA3FC5D8DB3 at byte 68615454
# This item is included in the report because it matches --outliers.
# Scores: V/M = 0.08
# Time range: 2013-10-02 07:34:59 to 2013-10-07 04:54:27
# Attribute pct total min max avg 95% stddev median
# ============ === ======= ======= ======= ======= ======= ======= =======
# Count 0 73
# Exec time 0 167s 2s 4s 2s 3s 416ms 2s
# Lock time 0 6ms 66us 143us 83us 108us 17us 73us
# Rows sent 0 0 0 0 0 0 0 0
# Rows examine 0 73 1 1 1 1 0 1
# Rows affecte 0 73 1 1 1 1 0 1
# Rows read 0 73 1 1 1 1 0 1
# Bytes sent 0 3.71k 52 52 52 52 0 52
# Merge passes 0 0 0 0 0 0 0 0
# Tmp tables 0 0 0 0 0 0 0 0
# Tmp disk tbl 0 0 0 0 0 0 0 0
# Tmp tbl size 0 0 0 0 0 0 0 0
# Query size 0 13.54k 188 191 189.99 183.58 0 183.58
# InnoDB:
# IO r bytes 0 0 0 0 0 0 0 0
# IO r ops 0 0 0 0 0 0 0 0
# IO r wait 0 0 0 0 0 0 0 0
# pages distin 0 403 2 108 5.52 11.95 13.77 1.96
# queue wait 0 0 0 0 0 0 0 0
# rec lock wai 0 0 0 0 0 0 0 0
# String:
# Databases heartbeat
# Hosts localhost
# InnoDB trxID ADAB0869 (1/1%), AE2CEFBD (1/1%)... 71 more
# Last errno 0
# Users
# Query_time distribution
# 1us
# 10us
# 100us
# 1ms
# 10ms
# 100ms
# 1s ################################################################
# 10s+
# Tables
# SHOW TABLE STATUS FROM `heartbeat` LIKE 'heartbeat'\G
# SHOW CREATE TABLE `heartbeat`.`heartbeat`\G
UPDATE `heartbeat`.`heartbeat` SET ts='2013-10-05T16:51:48.001010', file='bin.000027', position='520884073', relay_master_log_file=NULL, exec_master_log_pos=NULL WHERE server_id='3095733595'\G
# Converted for EXPLAIN
# EXPLAIN /*!50100 PARTITIONS*/
select ts='2013-10-05T16:51:48.001010', file='bin.000027', position='520884073', relay_master_log_file=NULL, exec_master_log_pos=NULL from `heartbeat`.`heartbeat` where server_id='3095733595'\G

[monty solomon (monty+launchpad)]

What are the plans to fix this bug? It is open over two years.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PT-1158