Обсуждение: Why does everyone think MySQL is easier?

Just a comment on the comparative state of MySQL rather than anything of
significance.

Why do people say MySQL is easier to install and manage than PostgreSQL? I
can't say I've had much difficulty getting PostgreSQL installed and working
from source a yet I've just installed a binary MySQL tarball and it doesn't
start.

Okay, the key is to work on the principle that the instructions are a guide and
not an exact list of what should be done. But then I haven't tried using it
yet.


--
Nigel J. Andrews

"Nigel J. Andrews" <nandrews@investsystems.co.uk> writes:
> Why do people say MySQL is easier to install and manage than PostgreSQL?

I think that "conventional wisdom" is a reflection of where things were
three or four years ago, not where they are today.  I can't speak to
MySQL's improvement, but we've sure made some huge strides since then.

            regards, tom lane

My experience is that the majority of the things considered
easier to do in MySQL are just done differently in Postgres
and hence makes Postgres look slightly harder to manage to
those used to MySQL (and vice versa, but I suspect there
are more people with MySQL experience trying out Postgres
than the other way around).


A few things comes to mind making MySQL somewhat easier though:

- it is extremely rare for an upgrade to require dump/restore

- you don't have to worry about running vacuum analyze regularly


The common dump/restore requirement with Postgres makes it
harder to manage an installation (especially with lots of
databases) since a software upgrade must be synchronized
with a database reload for all databases (*).

The need for vacuum makes Postgres work less well straight
out of the box, if you don't know about it your performance
will slowly go down the drain. (I don't remember how the
vacuum improvements in 7.3 change this, will a 7.3
installation work reasonably well without vacuum being run
at all e.g. in a scenario with lots of updates? (**))


      _
Mats Lofkvist
mal@algonet.se


(*) A temporary fix to this could be to change the
    installation to install under a versioned directory
    (a la gcc) just making the lastest version the default
    and maybe then change postmaster to start the correct
    version of the server after checking the database?
    Even if the latter part is left out, just having the
    old versions still available by default would improve
    things imho. And of course postmaster should refuse
    to start with a version mismatch (at least unless
    some override flag is used), but maybe this is
    already the case?

(**) If not, the auto-vacuum I have seen discussed
    will be even more welcome :-)

On 28 Nov 2002 at 15:35, Mats Lofkvist wrote:
> My experience is that the majority of the things considered
> easier to do in MySQL are just done differently in Postgres
> and hence makes Postgres look slightly harder to manage to
> those used to MySQL (and vice versa, but I suspect there
> are more people with MySQL experience trying out Postgres
> than the other way around).

I agree with that.. While I can start/stop/use postgresql in sleep, I can not
say for sure how to start mysql from command-line..

I just found it astonishigly difficult to start and manage.. Only way I could
get it working was to link /var/lib/mysql to a spacious location and use mysql
service.

And I am not satisfied with startup scripts provided with distros. They are
usually dumber than expected( Long time back, I wrote to mandrake to have
configurable database location after release of Mandrake 8.0.. No luck so
far..)

> The need for vacuum makes Postgres work less well straight
> out of the box, if you don't know about it your performance
> will slowly go down the drain. (I don't remember how the
> vacuum improvements in 7.3 change this, will a 7.3
> installation work reasonably well without vacuum being run
> at all e.g. in a scenario with lots of updates? (**))

In other words, 7.3 has good performance as it is. So if you vacuum, you will
see a really blazing performance..


Bye
 Shridhar

--
Absentee, n.:    A person with an income who has had the forethought to remove
himself from the sphere of exaction.        -- Ambrose Bierce, "The Devil's
Dictionary"

On Thu, 2002-11-28 at 10:10, Shridhar Daithankar wrote:
> On 28 Nov 2002 at 15:35, Mats Lofkvist wrote:
> > The need for vacuum makes Postgres work less well straight
> > out of the box, if you don't know about it your performance
> > will slowly go down the drain. (I don't remember how the
> > vacuum improvements in 7.3 change this, will a 7.3
> > installation work reasonably well without vacuum being run
> > at all e.g. in a scenario with lots of updates? (**))
>
> In other words, 7.3 has good performance as it is. So if you vacuum, you will
> see a really blazing performance..

Or more accurately, in 7.3 the performance of indexscans should degrade
less if there are a lot of expired tuples in the table (thanks to Tom
Lane's work). It won't effect seqscans, though.

Cheers,

Neil

Am Donnerstag, 28. November 2002 01:38 schrieb Tom Lane:
> "Nigel J. Andrews" <nandrews@investsystems.co.uk> writes:
> > Why do people say MySQL is easier to install and manage than PostgreSQL?
>
> I think that "conventional wisdom" is a reflection of where things were
> three or four years ago, not where they are today.  I can't speak to
> MySQL's improvement, but we've sure made some huge strides since then.

Just two days ago I had an argument with an MySQL user. He complained that PG
has "those weird \ commands" where MySQL uses a notation matching the normal
SQL look-and-feel.

With kind regards / mit freundlichem Gruß
    Holger Klawitter
--
Holger Klawitter                          http://www.klawitter.de
lists@klawitter.de