From: | Chris Bitmead <chris(dot)bitmead(at)bigfoot(dot)com> |
---|---|
To: | |
Cc: | pgsql-sql(at)postgreSQL(dot)org |
Subject: | Re: [SQL] Large objects - bug? caveat? feature? |
Date: | 1999-04-24 02:58:53 |
Message-ID: | 3721336D.10BB8F3C@bigfoot.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-sql |
Justin Long wrote:
> 2,000 byte article. In other words, it doesn't shrink the file > to the edited size.
Does the interface have any equivilent to the UNIX O_TRUNC?
> Secondly, I notice that in my data/base/... area that whenever I create an
> object it creates a single file on the disk. Does that mean that if I have
> 100,000 articles in my knowledge base, that it is possible that I will have
> 100,000 files individual 8-to-10k files on my hard drive?
It's worse than that I think. I believe you get _two_ files for each
large object. Large objects really suck badly.
> Does Linux suffer
> degradation in performance when having that many files in a
> directory?
Absolutely does suffer. Even worse, your regular database tables are in
the same directory, so they'll suffer too. Another problem, is pgdump
doesn't dump large objects so you have to figure out some other backup
strategy.
--
Chris Bitmead
http://www.bigfoot.com/~chris.bitmead
mailto:chris(dot)bitmead(at)bigfoot(dot)com
From | Date | Subject | |
---|---|---|---|
Next Message | Chris Bitmead | 1999-04-24 12:05:44 | Percentages? |
Previous Message | Bruce Momjian | 1999-04-23 21:29:24 | Re: [SQL] what is "cost"? |