Re: Integrity on large sites

Поиск
Список
Период
Сортировка
От Dave Page
Тема Re: Integrity on large sites
Дата
Msg-id 4655D407.1020108@postgresql.org
обсуждение исходный текст
Ответ на Re: Integrity on large sites  ("Joshua D. Drake" <jd@commandprompt.com>)
Список pgsql-general
Joshua D. Drake wrote:
> Dave Page wrote:
>
>>> problem with your setup. Granted, MySQL is a pretty bad database, but
>>> it's not *that* bad -- your example implies that heavily MyISAM-based
>>> (you don't say whether this is MyISAM or InnoDB) sites such as
>>> Slashdot and Flickr should be falling over every hour.
>>
>> I'm not going to comment on who's fault it is, but the OP quoted 100
>> updates and 600 selects per *second*. I can't imagine Flickr or Slashdot
>> (which is heavily csched for reading) are under anything like that sort
>> of constant load.
>
> Uhmmm.... I would not be surprised at *all* at slashdot or flickr doing
> that type of velocity. We have customers right now that under peak are
> doing 10 times that and yes these are customers that have similar types
> of websites.

Well taking the /. case, it's well known that they generate static pages
periodically in much the same way as we do for our website and serve
those, rather than hitting the comments database for every web hit.

As for the updates, I imagine they are next to non-existent on /. but if
you look at the comment posts instead which are probably their most
frequent type of non-select query, at 100 inserts per second, that
equates to 8,640,000 comments added per day. Assuming they publish, say
20 stories per day, that averages at 432,000 comments, per story, per day.

I don't recall the last time I saw that sort of response to a /. story...

Regards, Dave.

В списке pgsql-general по дате отправления:

Предыдущее
От: Ron Johnson
Дата:
Сообщение: Re: why postgresql over other RDBMS
Следующее
От: Scott Marlowe
Дата:
Сообщение: Re: Timestamp with time zone: why not?