Re: fairly current mysql v postgresql comparison need for

From: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
To: Mike Mascari <mascarm(at)mascari(dot)com>
Cc: Arjen van der Meijden <acm(at)tweakers(dot)net>, "'scott(dot)marlowe'" <scott(dot)marlowe(at)ihs(dot)com>, "'Randal L(dot) Schwartz'" <merlyn(at)stonehenge(dot)com>, pgsql-general(at)postgresql(dot)org
Subject: Re: fairly current mysql v postgresql comparison need for
Date: 2003-03-24 20:58:25
Message-ID: 200303242058.h2OKwP818902@candle.pha.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general


Supposedly MySQL did add MVCC in some fashion, so I assume that's how
they would do it.

---------------------------------------------------------------------------

Mike Mascari wrote:
> Bruce Momjian wrote:
> > Our MVCC allows it because there are multiple copies of the row --- pre-dump
> > and post-dump.
>
> Exactly. However, according to the mySQL docs I see the
> mysqldump option:
>
> "--single-transaction
>
> This option issues a BEGIN SQL command before dumping data from
> server. It is mostly useful with InnoDB tables and
> READ_COMMITTED transaction isolation level, as in this mode it
> will dump the consistent state of the database at the time then
> BEGIN was issued without blocking any applications. When using
> this option you should keep in mind that only transactional
> tables will be dumped in a consistent state, e.g., any MyISAM or
> HEAP tables dumped while using this option may still change
> state. The --single-transaction option was added in version 4.0.2."
>
> They also have mysqlhotcopy which:
>
> "mysqlhotcopy is a Perl script that uses LOCK TABLES, FLUSH
> TABLES and cp or scp to quickly make a backup of a database.
> It's the fastest way to make a backup of the database or single
> tables, but it can only be run on the same machine where the
> database directories are."
>
> Again, without MVCC, I fail to see how InnoDB or BDB can perform
> a hot-backup without blocking UPDATES/DELETES until the dump is
> complete. Of course mysqlhotcopy just locks the whole database
> until the backup is complete.
>
> scott(dot)marlowe(at)ihs(dot)com also wrote:
>
> > As long as there aren't any transactions pending while
> > you run it. It does have issues then. The
> > hotbackuptool that MySQL comes with is explicitly
> > documented by the mysql folks as NOT working for
> > innodb tables.
>
> Apparently --single-transaction was added in 4.0.2. It still
> doesn't resolve the underlying concurrency issues associated
> with a non-MVCC database, however.
>
> medi(dot)montaseri(at)intransa(dot)com also wrote:
>
> > Not to diverge from the main issue, but was this a
> > referene to a conversation between Bill Gate and
> > Steve Jobs in "Silicon Valley Pirates" movie?
>
> Not intentionally ;-)
>
> Mike Mascari
> mascarm(at)mascari(dot)com
>
>

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Robert Treat 2003-03-24 21:06:46 Re: Point in time recovery?
Previous Message Arjen van der Meijden 2003-03-24 20:58:13 Re: fairly current mysql v postgresql comparison need for