Aurora for MySQL is coming

I am excited about Aurora for MySQL. While there aren't many details, I have two conclusions from the information that is available. First, many talented people did great work on this. Second, many customers want the features it provides and some of these features are otherwise unavailable unless you are web-scale and can afford a team of automation experts. This is a big deal and good for the MySQL community. I am not just writing this to boost my priority on the Aurora preview signup list.

I don't think it matters whether Aurora has better performance. The big story is much better availability and manageability without having to hire an expert MySQL operations team. The limiting factor might be cost but that depends on whether you come from the land where everything is free or from a commercial DBMS. And even in the land of free, the cost for an operations team is large.

Soapbox

Before describing what I learned about it I have a short editorial.

• Is Amazon the reason we need AGPL? It isn't clear to me that they improve upstream MySQL. They do benefit the community by making it easier to run MySQL in the cloud. But I don't see them in the community. Which means I also don't see their developers in the community and who wants to disappear while working on mostly open-source?
• Their marketing leads with up to 5X faster than MySQL which is translated by the tech press into 5X faster than MySQL. When I signed up for an Aurora preview the response email from Amazon also used the 5X faster than MySQL claim. I prefer up to 5X faster than MySQL.
• In the video from AWS re:Invent 2014 James Hamilton makes some interesting claims.
• I am not sure he heard about InnoDB based on this statement -- Just doing a state of the art storage engine would have been worth doing. Take Jim Gray's black book on transactions, implement it, I would be happy. Someone can tell him to be happy.
• In big print -- 3X write performance, 5X read performance. In small print -- sysbench. It will take time to determine whether similar performance & availability can be had elsewhere at a similar cost.
• In big print -- 400X less lag, 2 seconds vs 5 milliseconds. I doubt this. The slide also stated -- synchronous multi-DC replication. I don't see how that is possible except within a datacenter given network latency. Other content from Amazon claims this is async replication. But then Mr. Hamilton stated that everything is there, transactions are not lost were two data centers to instantly disappear. Again, that requires sync replication. This is confusing and I am surprised his claims don't match the other documentation from Amazon.
• It can fix torn pages. InnoDB does this for writes in progress during a crash. I assume Aurora can also do this for pages written a long time ago and that would be a big deal.

Information

What is Aurora? I don't know and we might never find out. I assume it is a completely new storage engine rather than a new IO layer under InnoDB. I encourage people to read the details page, FAQ, slides and pricing guide. The response from Clustrix is also useful. At least one answer on Quora is excellent and this overview of Amazon datacenters will convince you that they network needed to make this work with low latency.

The big deal is that the database is replicated 6X using 3 availability zones (AZs). I assume this means it uses 2 storage clusters per AZ. Documentation from Amazon states this is done using async replication and (committed) writes are available within a few milliseconds or 10s of milliseconds. I assume this is only for Aurora read replicas in the same AZ as the master. Network latency will make that impossible for some remote AZs.  In the presentation at AWS re:Invent there are claims that the replication is synchronous. That is confusing.

For now I will assume that Aurora does async replication to the two remote AZs and the non-primary storage cluster in the primary AZ -- which means that 5 of the 6 copies are updated via async replication. If this is true, then an outage at the primary storage cluster can mean that recent commits are lost. It would be great if someone were to make this clear. My preferred solution would be sync replication within the primary AZ to maintain copies in 2 storage clusters, and then async replication to the 4 copies in the 2 remote AZs. We will soon have multiple solutions in the MySQL community that can do sync replication within a datacenter and async replication elsewhere -- Galera, upstream and lossless semisync. But Aurora is much easier to deploy than the alternatives. My standard question is what commit rate can be sustained when all updates are to the same row? Sync replication with cross-country network round trips makes that slow.

The presentation also claimed this was mostly drop-in compatible. I am interested in compatibility with InnoDB. 

• What are the semantics for cursor isolation? PostgreSQL, InnoDB and Oracle all do snapshot isolation with different semantics for writes. PostgreSQL has excellent documentation to describe the behavior.
• Does Aurora support clustered indexes? 
• What is the max size of an index key?
• How are large columns supported? Is data always inline?
• How is multi-versioning implemented? InnoDB usually does updates in place so there isn't much work for purge to do except for deletes and updates to secondary index columns. 
• Does this use pessimistic or optimistic concurrency control?
• Does this support partitioning?
• What is the block size for reads?
• Does this use compression?

Features

The brief description of features is very interesting. I will summarize that here. The features sound great and I expect them to get a lot of customers.

• Unplanned failover to an Aurora read replica takes a few minutes. Unplanned failover to a new instance can take up to 15 minutes. This happens for customers who aren't spending money on Aurora read replicas. This is another feature that will make Aurora very popular. While it will be nice to make failover faster, the big deal is that they provide this. The usual MySQL deployment required some do-it-yourself effort to get something similar.
• Storage uses SSD. I assume this is based on EBS. They do background scrubbing to detect and correct corrupt pages. Storage grows in 10G increments. You don't have to provision for a fixed amount of GB or TB, they will grow as needed up to 64T (or 64T/table). I am not sure I would want 64T in one instance, but it can make it easy to archive data in place. Also automatic growth to such large database sizes will make it much easier for deployments to avoid sharding especially when an instance has many TB of cold data.
• There are interesting features for point-in-time recovery, incremental backup and snapshots. This is integrated with S3. Incremental backups make it affordable to archive data in place as you don't do full backups for data that doesn't change. But I don't understand all of their backup options. 
• Database is replicated 2 times within 3 AZs so there are 6 copies. Up to 2 copies can be lost and writes are still possible. Up to 3 copies can be lost and reads are still possible. I assume that by copies can be lost they mean storage clusters can be lost. Automatic recovery here is another big deal.
• The buffer pool survives mysqld process restart. I wonder if that is only true for planned restart. Regardless, this is a very useful feature for IO-bound workloads when it is important to have a warm cache. Percona used to have a patch for this with InnoDB.
• Replication is supported in two ways -- via Aurora (at storage level) and MySQL (binlog). My bet is that Aurora replication will be most popular but some people will use MySQL replication to replicas with locally attached storage to save money. They claim much less lag with Aurora and I agree there will be less but I am not sure there will be 400X less. However, they avoid some overhead on commit by not using the binlog and they avoid a lot of complexity by not requiring MySQL replication on the master or slave. I assume that Aurora replication ships deltas rather than page images to be efficient on the network.

Cost

Cost comparisons will be interesting. I am very uncertain about my estimates here as I don't have much experience with prices on AWS or for normal sized datacenter deployments limited to a few servers. I estimate a 3 year cost of $568,814 for a high-end Aurora deployment: largest servers, 1 master, 1 replica, backup and storage. It will be interesting to compare this to non-AWS hardware because you need to account for features that Aurora provides and for the extra availability of the Aurora storage solution. I used 3TB and 10k IOPs from storage because that can be provided easily by a high-end PCIe flash device, but that also assumes a very busy server.

• Reserved db.r3.8xlarge (32 vCPUs, 244GiB RAM) for 2 servers is $115,682 over 3 years
• 50TB of backup at 10 cents/GB/month is $53,100 over 3 years
• 3TB of storage and 10,000 IOPs per server is $400,032 for 2 servers over 3 years. But I am not sure if the storage cost includes IO done to all AZs to maintain the 6 database copies. In that case, the storage cost might be higher.

Updates with secondary index maintenance: 5.7 vs previous releases

My previous results for an update-only workload to a cached and IO-bound database used a table that did not require secondary index maintenance from the update. Prior to that I ran tests using mysqlslap and might have found a regression in InnoDB when the update requires secondary index maintenance (bug 74235). The mysqlslap test did all updates to the same row. The test I describe here chooses the row to update at random and this workload does not reproduce the worst case from bug 74235.

The results are that newer releases tend to do worse at 1 thread, 5.7.5 does best at 32 threads and 5.7.5 does slightly worse than 5.6.21 at less than 32 threads.  We need to fix the regressions in 5.7.5 at less-than-high concurrency workloads.

setup

The test here was the same as described in my previous posts with one exception -- the test tables have a secondary index on (k,c) and note that the column c is set to 0 at test start. The updates are of the form: UPDATE $table SET c = c + 1 WHERE id = $X. The test tables look like:

CREATE TABLE sbtest1 (

  id int(10) unsigned NOT NULL auto_increment,

  k int(10) unsigned NOT NULL default '0',

  c char(120) NOT NULL default '',

  pad char(60) NOT NULL default '',

  PRIMARY KEY  (id),

  KEY k (k,c)

) ENGINE=InnoDB AUTO_INCREMENT=8000001 DEFAULT CHARSET=latin1

There were 8 tables in the database with 8M rows per table. Tests were run for 1 thread with 1 table and 8, 16, 32 threads with 8 tables. The test at each concurrency level was run for 10 minutes. Tests were run for a 1G (IO-bound) and 32G (cached) InnoDB buffer pool. For the cached test the tables were read into the buffer pool at test start. The updates/second rate from these tests had more variance than previous tests. I ran tests many times and report rates here that were typical. For the 1-thread tests that use 1 table, the database (1 table) is 2G. For the 8, 16 and 32 thread tests that use 8 tables the database (8 tables) is 16G.

results at 1 thread

The rates for a cached database and 1 thread match most previous results -- newer releases are much slower than older releases and MySQL 5.7.5 is about 1.5X worse than MySQL 5.0.85. The results for the IO-bound setup don't follow that trend. Newer releases do better. I regret that I did not have the time to explain the difference between 5.6.21 and 5.7.5.

results at many threads

MySQL 5.7.5 and 5.6.21 are much better than older releases. 5.6.21 does better than 5.7.5 at 8 and 16 threads while 5.7.5 is the best at 32 threads.

configuration

Sysbench IO-bound updates: MySQL 5.7 vs previous releases

I repeated the sysbench update-only test after reducing the InnoDB buffer pool to 1G. The test database is 16G so the test should be IO-bound. MySQL 5.7.5 is 10% worse than 5.0.85 at 1 thread and much better at 8+ threads.

The previous blog post has more details on the setup. The configuration is not crash safe as the doublewrite buffer and binlog are disabled and innodb_flush_log_at_trx_commit=2. This test continued to use direct IO for InnoDB but this test requires a high rate of reads from storage. The previous test did many writes but no reads. A fast flash device is used for storage.

results at 1-thread

Only 1 table is used so there is a 2G database and 1G buffer pool. This test is less IO-bound than the many threads tests reported in the next section. Two graphs are displayed below. The first has the updates/second rates and the second has the rates normalized to the rate from 5.0.85. MySQL 5.7.5 is about 10% worse than 5.0.85 and all servers do between 3000 and 4000 updates/second.

results at many threads

These tests uniformly distribute clients across the 8 tables. MySQL 5.7.5 is the best at 16 and 32 threads. MySQL 5.0.85 and 5.1.63 saturate at 8 threads.

Sysbench cached updates: MySQL 5.7 vs previous releases

I used MySQL 5.7.5 to repeat tests I did for MySQL 5.7.2 and 5.6.10 with the sysbench client and a cached database to compare performance for low and high concurrency. My configuration was optimized for throughput rather than crash safety. The performance summary:

• MySQL 5.7.5 is 1.47X slower than 5.0.85 at 1 thread
• MySQL 5.7.5 is a bit slower than 5.6.21 at 8 and 16 threads
• MySQL 5.7.5 is faster than others at 32 threads

setup

I used my sysbench 0.4 fork and set the value of the c column to 0 in the sbtest tables as that column is incremented during the test. The test used 8 tables with 8M rows per table (about 16G total) and the InnoDB buffer pool was 32G. One table was used for the 1 thread case. For the other tests the clients were uniformly distributed across the 8 tables. All data was in the buffer pool at test start. The test was run for 10 minutes at each concurrency level (1, 8, 16, 32 threads). Throughput appeared to be stable throughout the test.

Tests were run for MySQL versions 5.0.85, 5.1.63, 5.6.21 and 5.7.5. Tests were repeated for 4k and 8k innodb_page_size with MySQL 5.7.5 (see -16k, -8k and -4k below). Tests were also repeated with changes to innodb_flush_neighbors (=0), innodb_write_io_threads (=8), innodb_page_cleaners (=4) and innodb_purge_threads (=4). Results for those changes are not provided here as there wasn't an impact on QPS. The change to flush_neighbors reduced the InnoDB page write rate, which is a good thing.

results

The regression from 5.0.85 to recent versions is visible in the single-thread results below.

MySQL 5.7.5 is the champ at 32 threads but is slightly worse than 5.6.21 at 8 and 16 threads. Recent versions are much better than old versions.

configuration

I used this my.cnf for MySQL 5.7.5 and then something similar for other versions:

sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES 

table-definition-cache=1000

table-open-cache=2000

table-open-cache-instances=8

max_connections=20000

key_buffer_size=200M

metadata_locks_hash_instances=256 

query_cache_size=0

query_cache_type=0

server_id=9

performance_schema=0

binlog_format=row

skip_log_bin

innodb_buffer_pool_instances=8

innodb_io_capacity=32000

innodb_lru_scan_depth=4000

innodb_checksum_algorithm=CRC32

innodb_thread_concurrency=0

innodb_buffer_pool_size=32G

innodb_log_file_size=1900M

innodb_flush_log_at_trx_commit=2

innodb_doublewrite=0

innodb_flush_method=O_DIRECT

innodb_thread_concurrency=0

innodb_max_dirty_pages_pct=80

innodb_file_format=barracuda

innodb_file_per_table

This was the command line for one of the tests:

./sysbench --batch --batch-delay=10 --test=oltp --mysql-host=127.0.0.1 --mysql-db=test --oltp-table-size=8000000 --max-time=600 --max-requests=0 --mysql-table-engine=innodb --db-ps-mode=disable --mysql-engine-trx=yes --oltp-table-name=sbtest1 --oltp-skip-trx --oltp-test-mode=incupdate --oltp-dist-type=uniform --oltp-range-size=10 --oltp-connect-delay=0 --percentile=99 --num-threads=1 --seed-rng=1413171854 run

Page read performance: MySQL 5.7 vs previous releases

For this test I used sysbench to determine the peak rate for disk reads per second from InnoDB with fast storage. Fast storage in this case is the OS filesystem cache. I want to know the peak for MySQL 5.7.5 and whether there is a regression from 5.6. My summary is:

• 5.7.5 is 12% to 15% worse than 5.6.21 at 1 thread. I hope we can fix this via bug 74342.
• 5.7.5 is 2% to 4% worse than 5.6.21 at 8, 16 and 32 threads
• 5.6/5.7 are much better than 5.0 especially at high concurrency
• Page size (16k, 8k, 4k) has a big impact on the peak only when compression is used. 

setup

The test server has 40 hyperthread cores and is shared by the sysbench client and mysqld server.

I used my sysbench 0.4 fork configured to fetch 1 row by primary key via SELECT and HANDLER statements. The InnoDB buffer cache was 1G with innodb_flush_method="" (buffered IO). There were 8 tables in the test database with 16M rows per table.  The test database was either 32G uncompressed or 16G compressed so tests with compression had a better InnoDB buffer pool hit rate. Tests were run with 1, 8, 16 and 32 threads. The 1 thread test was limited to 1 table (2G or 4G of data) so this benefits from a better InnoDB buffer pool hit rate. The 8, 16 and 32 thread tests uniformly distribute clients across 8 tables.

Tests were done for MySQL 5.0.85, 5.6.21 and 5.0.85. All versions were tested for 16k pages without compression (see -16k-c0 below). Some MySQL 5.7.5 tests were repeated for 8k and 4k pages (see -4k- and -8k- below). Some MySQL 5.7.5 tests were repeated for 2X compression via zlib (see -c1- below).

results at 1 thread

The results at 1 thread show that peak QPS for 5.7.5 is 12% to 15% worse than for 5.6.21. I hope this can be fixed. I assume the problem is a longer code path from new features in 5.7.

results at 8 threads

The difference between 5.6.21 and 5.7.5 is smaller than at 1 thread. I assume this is the benefit from less mutex contention in 5.7.

results at 16 threads

The difference between 5.6.21 and 5.7.5 is smaller than at 1 thread. I assume this is the benefit from less mutex contention in 5.7.

results at 32 threads

The difference between 5.6.21 and 5.7.5 is smaller than at 1 thread. I assume this is the benefit from less mutex contention in 5.7.

configuration

I used this my.cnf for 5.7.5 and modified it for 5.6:

sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES 

table-definition-cache=1000

table-open-cache=2000

table-open-cache-instances=8

max_connections=20000

key_buffer_size=200M

metadata_locks_hash_instances=256 

query_cache_size=0

query_cache_type=0

server_id=9

performance_schema=0

binlog_format=row

skip_log_bin

innodb_buffer_pool_instances=8

innodb_io_capacity=32000

innodb_lru_scan_depth=4000

innodb_checksum_algorithm=CRC32

innodb_thread_concurrency=0

innodb_buffer_pool_size=1G

innodb_log_file_size=1900M

innodb_flush_log_at_trx_commit=2

innodb_doublewrite=0

# innodb_flush_method=O_DIRECT

innodb_thread_concurrency=0

innodb_max_dirty_pages_pct=80

innodb_file_format=barracuda

innodb_file_per_table

datadir=...

This is an example command line for a test that uses SELECT and 1 thread per table. My testing script starts 8 sysbench clients, one per table. The value for --seed-rng different for every sysbench client (function of current time and table number 1 to 8) to access different keys each run:

./sysbench --batch --batch-delay=10 --test=oltp --mysql-host=127.0.0.1 --mysql-db=test --oltp-table-size=16000000 --max-time=120 --max-requests=0 --mysql-table-engine=innodb --db-ps-mode=disable --mysql-engine-trx=yes --oltp-table-name=sbtest1 --oltp-read-only --oltp-skip-trx --oltp-test-mode=simple --oltp-point-select-all-cols --oltp-dist-type=uniform --oltp-range-size=10 --oltp-connect-delay=0 --percentile=99 --num-threads=1 --seed-rng=1413134256 run

Low-concurrency performance for updates with InnoDB: MySQL 5.7 vs previous releases

First I tested updates with the Heap engine as a base case and now I report on InnoDB to find regressions that are specific to it. The workload is described in the previous blog post. Response time is measured at 1, 4 and 32 threads for two types of updates. The unindexed update increments an unindexed column. The indexed update updates a column that is in a secondary index. Tests are run with the binlog enabled and disabled. Tests are done for MySQL versions 5.0.85, 5.1.63, 5.5.40, 5.6.21 and 5.7.5. In some cases tests are also done with the adaptive hash index disabled. The test database has 64,000 rows but all updates are to one row (lots of contention). The client is mysqlslap running on the same host as mysqld. The test server has 40 hyperthread cores.

For the single-thread update unindexed workload MySQL 5.7.5 is about 1.6X slower than 5.0.85. This is similar to what I reported in previous posts. For the single-thread updated indexed workload MySQL 5.7.5 is about 4X slower than 5.0.85 and much slower than 5.6.21. This is a regression for which I created bug 74325. This and a few other bugs might be motivation to get similar tests into the upstream release process.

The binaries below are tagged with -noahi, -nobl and -bl. The -noahi binary has the binlog and adaptive hash index disabled. The -nobl binary has the binlog disabled. The -bl binary has the binlog enabled.

update indexed, binlog off

Result summary:

• at 1 thread: 5.7.5 is 3.8X slower, 5.6.21 is 1.43X slower
• at 4 threads: 5.7.5 is 3.88X slower, 5.6.21 is slightly faster
• at 32 threads: 5.7.5 is 1.96X slower, 5.6.21 is faster

update indexed, binlog on

Result summary:

• at 1 thread: 5.7.5 is 4X slower, 5.6.21 is 1.37X slower
• at 4 threads: 5.7.5 is 2.1X slower, 5.6.21 is 1.09X slower
• at 32 threads: 5.7.5 is 2.26X slower, 5.6.21 is faster

update unindexed, binlog off

Result summary:

• at 1 thread: 5.7.5 is 1.65X slower, 5.6.21 is 1.44X slower
• at 4 threads: 5.7.5 is 1.33X slower, 5.6.21 is faster. Note that 5.7.5 is the only post-5.0.85 release still slower than 5.0.85 at 4 threads.
• at 32 threads: 5.7.5 is faster, 5.6.21 is faster

update unindexed, binlog on

Result summary:

• at 1 thread: 5.7.5 is 1.62X slower, 5.6.21 is 1.43X slower
• at 4 threads: 5.7.5 is 1.39X slower, 5.6.21 is 1.18X slower
• at 32 threads: 5.7.5 is faster, 5.6.21 is faster

Low-concurrency performance for updates and the Heap engine: MySQL 5.7 vs previous releases

The low-concurrency performance regression evaluation continues after evaluating range queries. This post covers updates with the Heap engine. The Heap engine is used as the base case and the next post has results for InnoDB. The database has 64,000 rows but all updates were for the same row. Tests were done for updates to an indexed and then an unindexed column at 1, 4 and 32 threads. Tests were also repeated with the binlog disabled and then enabled. There are regressions at 1, 4 and 32 threads. The summary is:

• Response time at 1 thread for 5.7.5 is between 1.55X and 1.67X worse than 5.0.85 
• Response time at 32 threads for 5.7.5 is between 1.19X and 1.49X worse than 5.0.85 
• In all cases it is worse in 5.7 than in 5.6
• Regressions at 4 and 32 threads are much worse when the binlog is enabled

All of the graphs graphs use normalized response time. For each tested version that is the time at X threads divided by the time at X threads for MySQL 5.0.85 where X is 1, 4 and 32.

update unindexed, binlog off

This section has results for updates to an unindexed column with the binlog disabled. Newer releases are faster at 4 and 32 threads. There is a gradual regression at 1 thread and the response time for MySQL 5.7.5 is 1.55X worse than for 5.0.85.

update unindexed, binlog on

This section has results for updates to an unindexed column with the binlog enabled. Newer releases are faster at 4 and 32 threads. There is a gradual regression at 1 thread and the response time for MySQL 5.7.5 is 1.67X worse than for 5.0.85. The single-thread regression here is worse than with the binlog disabled, so there is some regression from the binlog code.

update indexed, binlog off

This section has results for updates to an indexed column with the binlog disabled. Newer releases are faster at 4 and 32 threads. There is a gradual regression at 1 thread and the response time for MySQL 5.7.5 is 1.56X worse than for 5.0.85.

update indexed, binlog on

This section has results for updates to an indexed column with the binlog enabled. Newer releases are faster at 4 and 32 threads. There is a gradual regression at 1 thread and the response time for MySQL 5.7.5 is 1.67X worse than for 5.0.85. The single-thread regression here is worse than with the binlog disabled, so there is some regression from the binlog code.

command lines

Tests used mysqlslap as described in previous posts. The command lines for updates to the non-indexed and indexed columns are:
mysqlslap --concurrency=$t -h127.0.0.1 --create-schema=test1 --number-of-queries=$(( $t * 100000 )) --create=${ddl} --query="update foo set l = l + 2 where i=100"

mysqlslap --concurrency=$t -h127.0.0.1 --create-schema=test1 --number-of-queries=$(( $t * 100000 )) --create=${ddl} --query="update foo set k = k + 2 where i=100"