Re: psycopg2 hang with multithread frequent queries

From: Daniele Varrazzo <daniele(dot)varrazzo(at)gmail(dot)com>
To: "P(dot) Christeas" <xrg(at)linux(dot)gr>
Cc: psycopg(at)postgresql(dot)org
Subject: Re: psycopg2 hang with multithread frequent queries
Date: 2012-09-18 00:34:57
Message-ID: CA+mi_8a8Ge3B5F-f5ah0y-WCacRB7vyj__k+C6CRRRgFTgDzKg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: psycopg

On Tue, Sep 18, 2012 at 12:39 AM, P. Christeas <xrg(at)linux(dot)gr> wrote:
> On Tuesday 18 September 2012, David Roid wrote:
>> Hi Daniele, Federico,
>>
>> Just want to know, if I keep signal handler but switch from thread to
>> subprocess, i.e. put the signal handler code with database queries into
>> another process, is that safe?
>>
>
> IMHO, doing anything more than setting some "flag" variables and waking up
> regular threads is bad practice for a signal handler..

Yeah, in a process handler you are not even supposed to call malloc.
This on top of the fact that in python interaction between threads and
signals is unpredictable.

If you want more refined synchronization across agents you may take a
look at gevent/eventlet. But without knowing your problem is hard to
guess a solution.

-- Daniele

In response to

Responses

Browse psycopg by date

  From Date Subject
Next Message David Roid 2012-09-18 01:59:54 Re: psycopg2 hang with multithread frequent queries
Previous Message P. Christeas 2012-09-17 23:39:32 Re: psycopg2 hang with multithread frequent queries