Received: from guerilla.foo.bar (qmailr@hennen32s.iserlohn.netsurf.de [194.195.194.226]) by hub.org (8.9.2/8.9.1) with SMTP id JAA14891 for ; Sat, 13 Feb 1999 09:54:15 -0500 (EST) (envelope-from sas@schell.de) Received: (qmail 1341 invoked by uid 1000); 13 Feb 1999 14:55:35 -0000 Date: Sat, 13 Feb 1999 15:55:35 +0100 From: Sascha Schumann To: pierre@desertmoon.com Cc: pgsql-sql@postgreSQL.org Subject: Re: [SQL] connection timeout? Message-ID: <19990213155535.A1310@schell.de> References: <19990213010342.19457.qmail@desertmoon.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.1i In-Reply-To: <19990213010342.19457.qmail@desertmoon.com>; from pierre@desertmoon.com on Fri, Feb 12, 1999 at 06:03:42PM -0700 X-Operating-System: Linux guerilla 2.2.1 On Fri, Feb 12, 1999 at 06:03:42PM -0700, pierre@desertmoon.com wrote: > All, > > > Is there a connection timeout on the client code? That is if I were > to leave the client running but idle for a period of time I'd like > the client to time out and quit....is there anything like that > in PostgreSQL, can it be added? > > -=pierre I've employed the following method in one program (it's a CGI which may stay longer around without using the database connection): ----- PGconn *conn; void sig_alrm(int nr) { signal(SIGALRM, SIG_DEF); if(conn) PQfinish(conn); conn = NULL; } void use_pgsql(void) { if(!conn) { conn = PQconnectdb(...); signal(SIGALRM, sig_alrm); } alarm(60); } ----- Before you use any database action now, you call use_pgsql() which will start the counter and set up the connection. -- Regards, Sascha Schumann | Consultant | finger sas@schell.de | for PGP public key