3 ms·

I wonder what performance of Bucardo + Postgres is like vs. Galera + MariaDB. I'm mainly curious if switching from sync to async would be worth the performance

I wonder what performance of Bucardo + Postgres is like vs. Galera + MariaDB. I'm mainly curious if switching from sync to async would be worth the performance benefits. Then again, I suppose since Bucardo can handle MySQL/MariaDB it'd be fairer to compare Galera vs. Bucardo on MySQL.

dscrd 12y ago
Does Galera still suffer from the "gap locks" and related deadlocks? I was quite horrified to find this out. I would hope that that's not a problem with Bucardo or in fact any serious database system.
- opendais 12y ago
  You mean InnoDB's gap locks? http://www.mysqlperformanceblog.com/2012/03/27/innodbs-gap-locks/ http://www.mysqlperformanceblog.com/2012/03/27/innodbs-gap-l... http://pooteeweet.org/blog/745 http://pooteeweet.org/blog/745 If you actually have that problem in production, it is because you are spreading the writes out to all of your Galera servers and simultaneously doing conflicting updates/transactions. I'm uncertain what exactly you expect any database to do in this situation? Silently discard one?
  - dscrd 12y ago
    Spreading writes to all servers: yes. Conflicting updates: no; this problem happens when doing simple inserts. See http://www.toofishes.net/blog/mysql-deadlocking-simple-inserts/ http://www.toofishes.net/blog/mysql-deadlocking-simple-inser... -- that's an old blog post but this still happens with a reasonably late mysql and mariadb. I have no reason to believe that it's not a fundamental design failure. It's of course not a Galera issue per se, but rather a more serious InnoDB issue, but it causes problems very easily when using Galera. Perhaps there's a workaround that I don't know of yet? I'd be happy to hear of it. Perhaps it's a fundamental RDBMS problem that I don't understand yet? I'd be happy to be educated if so. As to your last question, I'd expect the database system to accept all such inserts to all instances since they don't conflict with each other in any way.

dscrd 12y ago

Does Galera still suffer from the "gap locks" and related deadlocks? I was quite horrified to find this out. I would hope that that's not a problem with Bucardo or in fact any serious database system.

opendais 12y ago
You mean InnoDB's gap locks? http://www.mysqlperformanceblog.com/2012/03/27/innodbs-gap-locks/ http://www.mysqlperformanceblog.com/2012/03/27/innodbs-gap-l... http://pooteeweet.org/blog/745 http://pooteeweet.org/blog/745 If you actually have that problem in production, it is because you are spreading the writes out to all of your Galera servers and simultaneously doing conflicting updates/transactions. I'm uncertain what exactly you expect any database to do in this situation? Silently discard one?
- dscrd 12y ago
  Spreading writes to all servers: yes. Conflicting updates: no; this problem happens when doing simple inserts. See http://www.toofishes.net/blog/mysql-deadlocking-simple-inserts/ http://www.toofishes.net/blog/mysql-deadlocking-simple-inser... -- that's an old blog post but this still happens with a reasonably late mysql and mariadb. I have no reason to believe that it's not a fundamental design failure. It's of course not a Galera issue per se, but rather a more serious InnoDB issue, but it causes problems very easily when using Galera. Perhaps there's a workaround that I don't know of yet? I'd be happy to hear of it. Perhaps it's a fundamental RDBMS problem that I don't understand yet? I'd be happy to be educated if so. As to your last question, I'd expect the database system to accept all such inserts to all instances since they don't conflict with each other in any way.