Let me say NO, but nothing really bad just annoying and thanks to Oli (always great) less annoying then how it could be.
Anyhow, I had to put all up in a RED HAT box on EC2.
So I start to take a look here an there to see what is the best way to do it, I also take the configurator from Severalnines, but for the moment I am not going to use it,
and in any case I never use it for the my.cnf parameters.
But what was the scope of all this?
I have a customer that could be interested in using a scaling MySQL solution, for his business.
Requirements as usual are 100% HA (OK we know this is not possible), scaling in writes as much as possible, geographic distribution.
Data set will be starting from 40G (peanuts) to 6 Tb (well not so peanuts) in 2 years.
More the scalability need to be "out" and "in" so architecture need to be able to scale as we like in relation to the "seasons".
What from there?
Well first of all define the generic architecture that should serve better, then find few good candidates that could fit in from the technological part
do some POCs and take numbers...
So here we are POCs and Numbers but first of all Installation.
As said I need to do that using REDHAT 6 EL so no way to use the pre-compile version, but taking a try will not harm.
Time to test and see how it works.
What I want to have as start:
3 nodes as start
data replicate from to each node
do stress test
download and installation
Excluding the demo package (that I don't want to use)
Any how I have found the site confuse, what I have to take? the package with ONLY the wsrep patch, or also the galera provider, all none both?
I had to download and extract to see that the galera libraries where not in the wsrep version, noting bad having it clear and on the same page could help.
Then I took the file from several nine using the configurator ...
DO I REALLY need it ??
mmm let us see
You will deploy on ec2, using redhat, and x86_64 architecture. You can also go into s9s-galera-xyz/scripts/install/.s9s/config
to make adjustments (but be careful) - if you get it wrong, refresh this page and start over with fresh deployment package.
Upload your keypair to /root/.ssh/Availng.pem on the cluster-control server
tar xvfz s9s-galera-1.0.0.tar.gz
./deploy.sh 2>&1 |tee cc.log
A detailed log is also written to $HOME/s9s_deploy.log
As said I decide not to use it for the moment.
My directory layout will be:
Next I had to do the usual thing:
setup MySQL user and group
I am in rush (never good), so I did start and configure MySQL, but then when I run it:
111118 17:34:55 [Note] Event Scheduler: Loaded 0 events
111118 17:34:55 [Note] WSREP: wsrep_load(): loading provider library '/usr/local/galera/lib64/galera/libgalera_smm.so'
111118 17:34:55 [ERROR] WSREP: wsrep_load(): dlopen(): libssl.so.6: cannot open shared object file: No such file or directory
111118 17:34:55 [ERROR] WSREP: wsrep_load(/usr/local/galera/lib64/galera/libgalera_smm.so) failed..
111118 17:34:55 [Note] WSREP: wsrep_load(): loading provider library 'none'
111118 17:34:55 [ERROR] Aborting
111118 17:34:55 [Note] WSREP: Service disconnected.
111118 17:34:56 [Note] WSREP: Some threads may fail to exit.
111118 17:34:56 InnoDB: Starting shutdown...
111118 17:34:57 InnoDB: Shutdown completed; log sequence number 1595682
111118 17:34:57 [Note] /usr/local/mysql/bin/mysqld: Shutdown complete
Ok Ok I know they said that so no way I need to recompile and install.
As Oli wrote issue is set up all before so let me see for RED HAT:
Take MySQL as for Oli instructions http://www.fromdual.com/building-galera-replication-from-scratch
and patch: wget http://launchpad.net/codership-mysql/5.5/5.5.15-21.2/+download/mysql-5.5.15-wsrep_21.2.patch
apply patch and download the galera replicator
BEFORE COMPILE install
yum -y install openssl
yum -y install boost-devel
yum -y install openssl-devel.x86_64
yum -y install check-devel.x86_64
Then run :
Have a coffee or two ...
cp libgalera_smm.so /opt/mysql/lib/plugin/
cp garb/garbd /opt/mysql/bin/
WOW I GOT IT
11122 18:19:53 [Note] WSREP: Flow-control interval: [8, 16]
As state in the internal text file a note about the LIMITATIONS
1) Currently replication works only with InnoDB storage engine. Any writes to
tables of other types, including system (mysql.*) tables are not replicated.
However, DDL statements are replicated in statement level, and changes
to mysql.* tables will get replicated that way.
So, you can safely issue: CREATE USER...,
but issuing: INSERT INTO mysql.user..., will not be replicated.
2) DELETE operation is unsupported on tables without primary keys.
Rows in tables without primary keys may appear in different order on
different nodes. As a result SELECT...LIMIT... may return slightly
3) Unsupported queries:
* LOCK/UNLOCK tables is not supported in multimaster configuration.
* lock functions (GET_LOCK(), RELEASE_LOCK()... )
4) Query log cannot be directed to table. If you enable query logging,
you must forward the log to a file:
log_output = FILE
Use general_log and general_log_file to choose query logging and the
log file name
5) Maximum allowed transaction size is defined by wsrep_max_ws_rows and
wsrep_max_ws_size. Anything bigger (e.g. huge LOAD DATA) will be rejected.
6) Due to cluster level optimistic concurrency control, transaction issuing
COMMIT may still be aborted at that stage. There can be two transactions
writing to same rows and committing in separate cluster nodes, and only one
of the them can successfully commit. The failing one will be aborted.
For cluster level aborts, MySQL/galera cluster gives back deadlock error
code (Error: 1213 SQLSTATE: 40001 (ER_LOCK_DEADLOCK)).
7) XA transactions can not be supported due to possible rollback on commit.
8) Currently supported only in master-slave mode without parallel applying
(will be fixed in the nearest release):
* cascading foreign key constraints,
* secondary unique indexes and
* primary varchar indexes with non-binary collation.
Creating the full environment
Now that I have all up and working, I can finally use my current EC2 as template and add boxes to my cluster.
From here the real Fun will start ...
-> Expanding the cluster
-> Starting my dataset
-> Pushing on throttle and take numbers