We have a moderately large production MySQL database. Periodically, we will run commands, usually via a rails migration, that while running, bog down the database. As a specific example, we might add an index to a large table.
Is there any method that can reduce the priority MySQL gives to a specific task. A sort of "nice" within MySQL itself? I found this, which is what inspired the question:
PostgreSQL tips and tricks
Since adding an index causes the work to be done within the DB and MySQL process, lowering the priority of the Rails migration process seems like it won't help. Are there other ways we can lower the priority?
We use multiple, replicated database servers to make changes like this.
In our case, db1 is the master, replicated to db2. (db1->db2).
Start by making the change to db2. If things lock, replication will stall, but that's OK.
Move your traffic to db2. Any remnant traffic going to db1 will replicate over, so you won't lose anything.
Once there's no traffic on db1, rebuild it as a slave of db2 (db2->db1).
That's the general idea and you get very little downtime and you don't have to pull an all-nighter! We actually have three servers, so it's a little more complicated, but not much.
Good luck.
Unfortunately, there is no simple way to do this: commands that alter the database structure don't have a priority option.
If your tables are MyISAM, you can try this:
mysqlhotcopy to make a backup of the table
import that backup it into a different database server (one that's not under load)
make the changes there
make a mysqlhotcopy backup of the altered table
import it into the live server
Note that this may or may not be faster than adding the index on the live server, depending on the time it takes you to transfer the table back and forth.
Related
I'm currently thinking about the following problem:
A customer has set up a simple master/slave replication between two mariaDB systems. For unknown reasons they have set the flag "Replicate_Wild_Ignore_Table" to skip "logdb.%". Obviously, they decided to skip the skipping of that database and want the logdb to be included in the replication again.
I'm curious now, is it possible to somehow remove that flag and have the database in question be replicated as the rest or is there no way to circumvent the "stop slave, dump master, import dump, recreate replication based on current logpos, start slave" procedure?
You can't assume that the master still has all relevant binlogs that once contained updates to the logdb.% tables. That is, even if you could re-apply those updates, do you have enough history to account for all changes to the tables?
Another risk is if you use statement-based replication, if there were ever statements that referenced both a table in logdb.% and a table in another database, the replication filter has skipped that statement. So for example:
INSERT INTO mydb.mytable SELECT * FROM logdb.othertable;
Therefore even the tables that are not in logdb.% might be compromised. The point is you don't know for sure.
The bottom line is that you should definitely reinitialize the replica now by taking a current backup of the master, and avoid using replication filters in the future.
If you use InnoDB tables, you might consider using Percona XtraBackup to make the process easier. See https://www.percona.com/doc/percona-xtrabackup/2.3/howtos/setting_up_replication.html
I need to synch development mysql db with the production one.
Production db gets updated by user clicks and other data generated via web.
Development db gets updated with processing data.
What's the best practice to accomplish this?
I found some diff tools (eg. mySQL diff), but they don't manage updated records.
I also found some application solution: http://www.isocra.com/2004/10/dumptosql/
but I'm not sure it's a good practice as in this case I need to retest my code each time I add new innodb related tables.
Any ideas?
Take a look at mysqldump. It may serve you well enough for this.
Assuming your tables are all indexed with some sort of unique key you could do a dump and have it leave out the 'drop/create table' bits. Have it run as 'insert ignore' and you'll get the new data without effecting the existing data.
Another option would be to use the query part of mysqldump to dump only the new records from the production side. Again - have mysqldump leave off the 'drop/create' bits.
Each night we run mysqldump from production and dump it to another schema to use as a test or development server, on the same server. As the databases grow it is taking much of the night to finish. I'm trying to think of a way to speed this up. I thought since I am running binlog I thought maybe once a week do the Mysqldump and then the rest of the week just up date the other schema from the binlogs. Of course all this needs to be scripted if it is possible.
Also, is it faster to a mysqldump and import from one schema to another or is it faster to use an update from one schema to another to update all of the tables?
Is it possible to have a master/master on the same instance of mysql?
Did you look into MySQL replication? That's exactly what it does. Reads binlog from the master and executes it on the slave.
There are some limitations, for instance it assumes that you don't do writes to your slaves for data integrity sake.
Master-master replication is technically possible if you partition your data so same pieces of it not changed on different servers. Not sure you can do this about the same instance.
I've got a database which I intend to replicate for backup reasons (performance is not a problem at the moment).
We've set up the replication correctly and tested it and all was fine.
Then we realized that it replicates all the writes to the temporary tables, which in effect meant that replication of one day's worth of data took almost two hours for the idle slave.
The reason for that is that we recompute some of the data in our db via cronjob every 15 mins to ensure it's in sync (it takes ~3 minutes in total, so it is unacceptable to do those operations during a web request; instead we just store the modifications without attempting to recompute anything while in the web request, and then do all of the work in bulk). In order to process that data efficiently, we use temporary tables (as there's lots of interdependencies).
Now, the first problem is that temporary tables do not persist if we restart the slave while it's in the middle of processing transactions that use that temp table. That can be avoided by not using temporary tables, although this has its own issues.
The more serious problem is that the slave could easily catch up in less than half an hour if it wasn't for all that recomputation (which it does one after the other, so there's no benefit of rebuilding the data every 15 mins... and you can literally see it stuck at, say 1115, only to quickly catch up and got stuck at 1130 etc).
One solution we came up with is to move all that recomputation out of the replicated db, so that the slave doesn't replicate it. But it has disadvantages in that we'd have to prune the tables it eventually updates, making our slave in effect "castrated", ie. we'd have to recompute everything on it before we could actually use it.
Did anyone have a similar problem and/or how would you solve it? Am I missing something obvious?
I've come up with the solution. It makes use of replicate-do-db mentioned by Nick. Writing it down here in case somebody had a similar problem.
The problem with just using replicate-(wild-)do* options in this case (like I said, we use temp tables to repopulate a central table) is that either you ignore temp tables and repopulate the central one with no data (which causes further problems as all the queries relying on the central table being up-to-date will produce different results) or you ignore the central table, which has a similar problem. Not to mention, you have to restart mysql after adding any of those options to my.cnf. We wanted something that would cover all those cases (and future ones) without the need for any further restart.
So, what we decided to do is to split the database into the "real" and a "workarea" databases. Only the "real" database is replicated (I guess you could decide on a convention of table names to be used for replicate-wild-do-table syntax).
All the temporary table work is happening in "workarea" db, and to avoid the dependency problem mentioned above, we won't populate the central table (which sits in "real" db) by INSERT ... SELECT or RENAME TABLE, but rather query the tmp tables to generate a sort of a diff on the live table (ie. generate INSERT statements for new rows, DELETE for the old ones and update where necessary).
This way the only queries that are replicated are exactly the updates that are required, nothing else, ie. some (most?) of the recomputation queries hapenning every fifteen minutes might not even make its way to slave, and the ones that do will be minimal and not computationally expensive at all, just simple INSERTs and DELETEs.
In MySQL, as of 5.0 I believe, you can do table wildcards to replicate specific tables. There are a number of command-line options that can be set but you can also do this via your MySQL config file.
[mysqld]
replicate-do-db = db1
replicate-do-table = db2.mytbl2
replicate-wild-do-table= database_name.%
replicate-wild-do-table= another_db.%
The idea being that you tell it to not replicate any tables other than the ones you specify.
We have a live MySQL database that is 99% INSERTs, around 100 per second. We want to archive the data each day so that we can run queries on it without affecting the main, live database. In addition, once the archive is completed, we want to clear the live database.
What is the best way to do this without (if possible) locking INSERTs? We use INSERT DELAYED for the queries.
http://www.maatkit.org/ has mk-archiver
archives or purges rows from a table to another table and/or a file. It is designed to efficiently “nibble” data in very small chunks without interfering with critical online transaction processing (OLTP) queries. It accomplishes this with a non-backtracking query plan that keeps its place in the table from query to query, so each subsequent query does very little work to find more archivable rows.
Another alternative is to simply create a new database table each day. MyIsam does have some advantages for this, since INSERTs to the end of the table don't generally block anyway, and there is a merge table type to being them all back together. A number of websites log the httpd traffic to tables like that.
With Mysql 5.1, there are also partition tables that can do much the same.
I use mysql partition tables and I've achieve wonderful results in all aspects.
Sounds like replication is the best solution for this. After the initial sync the slave gets updates via the Binary Log, thus not affecting the master DB at all.
More on replication.
MK-ARCHIVER is a elegant tool to archive MYSQL data.
http://www.maatkit.org/doc/mk-archiver.html
MySQL replication would work perfectly for this.
Master -> the live server.
Slave -> a different server on the same network.
Could you keep two mirrored databases around? Write to one, keep the second as an archive. Switch every, say, 24 hours (or however long you deem appropriate). Into the database that was the archive, insert all of todays activity. Then the two databases should match. Use this as the new live db. Take the archived database and do whatever you want to it. You can backup/extract/read all you want now that its not being actively written to.
Its kind of like having mirrored raid where you can take one drive offline for backup, resync it, then take the other drive out for backup.