LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

[lvs-users] Checkcommand defunct processes

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: [lvs-users] Checkcommand defunct processes
From: Bruno Corsi dos Santos <corsibp@xxxxxxxxxxxx>
Date: Thu, 24 Apr 2008 07:22:49 -0700 (PDT)
Hi there,

I'm using the ldirector with heartbeat daemon on a CentoOS 5 (x86_64) and I 
realized that when I use the "checkcommand" option to run a script or 
executable to check wheather a service is alive or not, the process exits 
properly but it keeps in "defunct" state. There are up and down services that 
are being tested, so, I also realized that it happens just when the 
"checkcommand" process takes more than "checktimeout" seconds to complete. 

I decided to check the availability of my UDP service writting a code myself to 
check whether the server is up or down, so, I had to use the "checkcommand" 
option.

I also saw that sometimes the restart of the heartbeat hangs. But it doesn't 
happen so often...

Does anyone have already run into such trouble? I got thousands of "defunct" 
processes after some while, so, I have to restart the ldirectord to get rid of 
such "defunct" processes. I'm running the version script from the 
heartbeat-ldirectord-2.1.3-3.el5.centos rpm package from CentOS repository. A 
temporary solution I got was to increase the "checktimeout" providing that the 
ldirectord timeout would never happen. 

Any help would be welcome,

Thanks in advance,
Bruno





      Abra sua conta no Yahoo! Mail, o único sem limite de espaço para 
armazenamento!
http://br.mail.yahoo.com/


<Prev in Thread] Current Thread [Next in Thread>