Re: 500 times slower - Mailing list pgsql-odbc

From Karol Szkudlarek
Subject Re: 500 times slower
Date
Msg-id 420A18D3.8090107@mikronika.com.pl
Whole thread Raw
In response to Re: 500 times slower  ("Dave Page" <dpage@vale-housing.co.uk>)
List pgsql-odbc
Dave Page wrote:
>
>
> Err, no neither am I. Why do you think it's got something to do with
> Nagle/delayed ACKs?
>
> The only thing that instantly rings bells for me is that the max size of
> a text field is 8190 bytes at present (which really should be increased,
> if not removed altogether), which won't fit in the default buffer. But
> then, I wouldn't expect to see the performance drop you describe with a
> 4096 byte buffer, only one much smaller.
>
> Anyone else got any ideas?
>
> Regards, Dave
>

Hi Dave!

In my test case I read LOB of 10 megabytes size.
So then in the function (with 4kB SOCK_BUFFER_SIZE):
    SOCK_get_next_byte(SocketClass *self)
below line (socket.c):

self->buffer_filled_in = recv(self->socket, (char *) self->buffer_in,
self->buffer_size, 0);

runs in the following schema: 0msec, 200msec, 0msec, 200msec... etc.
So above times suggest delayed ACK timer (200msec).

Regards,
Karol


pgsql-odbc by date:

Previous
From: "Dave Page"
Date:
Subject: Re: 500 times slower
Next
From: "Dave Page"
Date:
Subject: Re: 500 times slower