Re: very slow largeobject transfers through JDBC

From: Dave Cramer <pg(at)fastcrypt(dot)com>
To: m(at)matevarga(dot)net
Cc: dmitigr(at)gmail(dot)com, "pgsql-generallists(dot)postgresql(dot)org" <pgsql-general(at)lists(dot)postgresql(dot)org>
Subject: Re: very slow largeobject transfers through JDBC
Date: 2018-09-01 22:55:33
Message-ID: CADK3HHJ_=qTU4=5RiMGgf0QGxPFD4Z+MKYqzD3p8G99vTS4pbg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Fri, 31 Aug 2018 at 10:15, Mate Varga <m(at)matevarga(dot)net> wrote:

> I see -- we could try that, though we're mostly using an ORM (Hibernate)
> to do this. Thanks!
>
> On Fri, Aug 31, 2018 at 3:57 PM Dmitry Igrishin <dmitigr(at)gmail(dot)com> wrote:
>
>> пт, 31 авг. 2018 г. в 16:35, Mate Varga <m(at)matevarga(dot)net>:
>> >
>> > Hi,
>> >
>> > we're fetching binary data from pg_largeobject table. The data is not
>> very large, but we ended up storing it there. If I'm copying the data to a
>> file from the psql console, then it takes X time (e.g. a second), fetching
>> it through the JDBC driver takes at least 10x more. We don't see this
>> difference between JDBC and 'native' performance for anything except
>> largeobjects (and bytea columns, for the record).
>> >
>> > Does anyone have any advice about whether this can be tuned or what the
>> cause is?
>> I don't know what a reason of that, but I think it's reasonable and
>> quite simple to call lo_import()/lo_export() via JNI.
>>
>
Can't imagine that's any faster. The driver simply implements the protocol

Do you have any code to share ? Any other information ?

Is the JDBC connection significantly further away network wise ?

Dave Cramer

davec(at)postgresintl(dot)com
www.postgresintl.com

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Dave Peticolas 2018-09-01 23:45:50 Re: locate DB corruption
Previous Message Guyren Howe 2018-09-01 17:33:16 Re: Rules and decision logic triggered by / acting on JSONB nodes in Postgres