0

We face the table level lock issues, almost every day on our test servers. 

TRANSACTION 0, not started
mysql tables in use 97, locked 97
MySQL thread id 429, OS thread handle 0x2aff6ff59700, query id 24900 ec2-*-*-*-*.compute-1.amazonaws.com *.*.*.* sminq cleaning up
---TRANSACTION 10631403, not started
MySQL thread id 321, OS thread handle 0x2aff7b359700, query id 24901 115.112.140.139 sminq init
show engine innodb status
---TRANSACTION 10632661, not started
MySQL thread id 13, OS thread handle 0x2aff4e39a700, query id 24817 localhost 127.0.0.1 rdsadmin cleaning up
---TRANSACTION 10632664, not started
MySQL thread id 6, OS thread handle 0x2aff396c5700, query id 24873 ec2-*-*-*-*.ap-southeast-1.compute.amazonaws.com *.*.*.* sminq cleaning up
---TRANSACTION 10632655, not started
MySQL thread id 7, OS thread handle 0x2aff39706700, query id 24783 ec2-*-*-*-*.ap-southeast-1.compute.amazonaws.com *.*.*.* sminq cleaning up
---TRANSACTION 10632652, not started
MySQL thread id 3, OS thread handle 0x2aff37d28700, query id 24745 ec2-*-*-*-*.ap-southeast-1.compute.amazonaws.com *.*.*.* sminq cleaning up
---TRANSACTION 10627075, not started
MySQL thread id 1, OS thread handle 0x2aff37ca6700, query id 0 Waiting for background binlog tasks
---TRANSACTION 10632663, ACTIVE 7 sec
mysql tables in use 1, locked 1
MySQL thread id 431, OS thread handle 0x2aff37daa700, query id 24863  172.31.3.120 sminq Waiting for table level lock
insert into `sminq`.`Queue_token` (`token_queue_id`,    `total_process_time`, `token_user`, `created_on`, `join_date`, `join_time`, `app_type`, `token_user_group`, `uuid`) values (13, 10, 87, '2016-07-21 04:47:04.157000', '2016-07-21 10:17:04', '10:10:00', 1, NULL, 'D<??BY??7?gk?Uo')
Trx #rec lock waits 0 #table lock waits 0
Trx total rec lock wait time 0 SEC
Trx total table lock wait time 0 SEC
---TRANSACTION 10632646, ACTIVE 45 sec

These keep occurring for Inserts only, we have never faced the issue with Updates or deletes.

I have an Isolation level READ-COMMITTED, along with innodb_autoinc_lock_mode = 2

mysqltuner output for the same server

[--] Up for: 2h 11m 55s (25K q [3.230 qps], 478 conn, TX: 3M, RX: 1M)
[--] Reads / Writes: 82% / 18%
[--] Binary logging is enabled (GTID MODE: OFF)
[--] Total buffers: 1.5G global + 17.0M per thread (100 max threads)
[!!] Maximum reached memory usage: 3.0G (152.35% of installed RAM)
[!!] Maximum possible memory usage: 3.1G (156.50% of installed RAM)
[OK] Slow queries: 0% (0/25K)
[!!] Highest connection usage: 95%  (95/100)
[OK] Aborted connections: 0.00%  (0/478)
[!!] Query cache is disabled
[OK] Sorts requiring temporary tables: 0% (0 temp sorts / 1K sorts)
[OK] Temporary tables created on disk: 24% (424 on disk / 1K total)
[OK] Thread cache hit rate: 80% (95 created / 478 connections)
[OK] Table cache hit rate: 129% (291 open / 224 opened)
[OK] Open file limit used: 0% (64/65K)
[OK] Table locks acquired immediately: 99% (6K immediate / 6K locks)
[OK] Binlog cache memory access: 100.00% ( 1618 Memory / 1618 Total)

-------- MyISAM Metrics -----------------------------------------------------
[!!] Key buffer used: 18.5% (1M used / 8M cache)
[OK] Key buffer size / total MyISAM indexes: 8.0M/2.4M
[!!] Read Key buffer hit rate: 82.2% (90 cached / 16 reads)

-------- InnoDB Metrics -----------------------------------------------------
[--] InnoDB is enabled.
[OK] InnoDB buffer pool / data size: 1.3G/29.0M
[!!] InnoDB buffer pool instances: 8
[!!] InnoDB Used buffer: 1.32% (1139 used/ 86584 total)
[OK] InnoDB Read buffer efficiency: 99.86% (713109 hits/ 714137 total)
[!!] InnoDB Write buffer efficiency: 0.00% (0 hits/ 1 total)
[OK] InnoDB log waits: 0.00% (0 waits / 4915 writes)

Since this is a test server we are running the tests on a t2.small

user160108
  • 930
  • 3
  • 8
  • 38

2 Answers2

1

1.3G buffer_pool in 2GB of RAM? This probably leads to lots of swapping, which is really bad for performance.

For 2GB of RAM and only 29M of data, let's set innodb_buffer_pool_size = 100M. That should enough for now, and safe for later (as the data grows).

(The recommendation of 70% or 80% only applies for machines with at least 4GB of RAM.)

Fix that. If the problem continues, then update the question with newer values, plus SHOW CREATE TABLE for the table(s) involved.

Rick James
  • 135,179
  • 13
  • 127
  • 222
  • Will update these values and test further, our production is on a m3.medium which has around 3.75 GB memory. Our data on this db is 665MB, so with your recommendations our production innodb_buffer_pool_size = 2G should be fine. – user160108 Jul 22 '16 at 08:55
  • No. Let it to 1400M. (2G _might_ be ok, but let's play it safe, especially since you won't immediately be using much more than 665MB.) – Rick James Jul 22 '16 at 15:59
  • Have made the changes and things seem to be working smoothly, will continue to monitor. Thanks for the tips. – user160108 Jul 25 '16 at 06:45
0

An AUTO-INC lock is a special table-level lock taken by transactions inserting into tables with AUTO_INCREMENT columns. In the simplest case, if one transaction is inserting values into the table, any other transactions must wait to do their own inserts into that table, so that rows inserted by the first transaction receive consecutive primary key values.

The innodb_autoinc_lock_mode configuration option controls the algorithm used for auto-increment locking. It allows you to choose how to trade off between predictable sequences of auto-increment values and maximum concurrency for insert operations.

The permissible values are 0, 1, or 2, for “traditional”, “consecutive”, or “interleaved” lock mode

innodb_autoinc_lock_mode = 2 (“interleaved” lock mode)

In this lock mode, no “INSERT-like” statements use the table-level AUTO-INC lock, and multiple statements can execute at the same time. This is the fastest and most scalable lock mode, but it is not safe when using statement-based replication or recovery scenarios when SQL statements are replayed from the binary log.

In this lock mode, auto-increment values are guaranteed to be unique and monotonically increasing across all concurrently executing “INSERT-like” statements. However, because multiple statements can be generating numbers at the same time (that is, allocation of numbers is interleaved across statements), the values generated for the rows inserted by any given statement may not be consecutive.

If the only statements executing are “simple inserts” where the number of rows to be inserted is known ahead of time, there will be no gaps in the numbers generated for a single statement, except for “mixed-mode inserts”. However, when “bulk inserts” are executed, there may be gaps in the auto-increment values assigned by any given statement.

Refer the source Locking and modify the configuration based on your requirement.

Thennarasan
  • 698
  • 6
  • 11