Re: Problem with "Persistent Connections"

Hi,

I'm still having problems with this:

Netware 6.5 cpr

OHD-WEB:m my*                                                             
    
MYSQLD.NLM                                                                
    
  Loaded from [SYS:/MYSQL/BIN/] on Jun 19, 2006  11:02:23 am              
    
  (Address Space = OS)                                                    
    
  MySQL Database Server                                                   
    
  Version 4.01.19 May 1, 2006                                             
    
  (c) 2003-2005 Novell, Inc. Portions (c) 2003 MySQL AB. All Rights
Reserved.  
MYSQLD_SAFE.NLM                                                           
    
  Loaded from [SYS:MYSQLBIN] on Jun 19, 2006  11:02:14 am                 
 
  (Address Space = OS)                                                    
    
  MySQL Database Server Monitor                                           
    
  Version 4.01.19 May 1, 2006                                             
    
  (c) 2003-2005 Novell, Inc. Portions (c) 2003 MySQL AB. All Rights
Reserved.  

2   krs   ohd-web.thunnissenonderhoud.nl:1059   Geen   Sleep   201  

My.cnf:
[mysqld]
default-character-set=latin1
language=english
bind-address=0.0.0.0
port=3306
datadir=web:/mysql/data
skip-locking
skip-innodb
wait_timeout=180
interactive_timeout=180
net_read_timeout=180


The php code is using pconnect.

Any other clues?

> > The problem is seen on NetWare too. We will fix it and put in the
build 
> > as soon as possible. What version of MySQL you are using?
> > Thanks,
> > Ram
> > > I searched the MySQL site, and I can see a mention of it with
FreeBSD:
> > > Known problem when using LinuxThreads on FreeBSD is:
> > >       The connection times (wait_timeout, interactive_timeout and

> > > net_read_timeout) values are not honored. The symptom is that
persistent 
> > > connections can hang for a very long time without getting closed
down 
> > > and that a 'kill' for a thread will not take affect until the
thread 
> > > does it a new command
> > >       This is probably a signal handling problem in the thread
library 
> > > where the signal does not break a pending read. This is supposed
to be 
> > > fixed in FreeBSD 5.0
> > > Is this also a problem on Netware?