[rhn-users] rhnsd daemon update

Todd Warner taw at redhat.com
Fri Jan 21 01:31:28 UTC 2005


On Mon, 17 Jan 2005, Jesse Becker wrote:

> On Sat, Jan 15, 2005 at 09:50:45PM +0800, Tsai Li Ming wrote:
> > From the rhn website, "The Red Hat Network Daemon (rhnsd) periodically 
> > connects to Red Hat Network to check for updates and notifications."
> > 
> > Does the rhnsd daemon perform system updates using up2date 
> > automatically, similar to yumd? If not, can up2date be configured to 
> 
> The rhnsd program will check for updates approximately once every two
> hours.  However, I have heard reports, and directly seen, that rhnsd
> will occasionally get stuck, and do nothing at all until it is restarted.  
> 
> > update the system automatically without having to use a cron job?
> 
> This is the purpose of rhnsd.  However, I have found it more reliable to
> simply run /usr/sbin/rhn_check from cron. 
> 
> Digging through the rhnsd source code, it appears to do three things:  
> 	1) Call rhn_check every 120 minutes.
> 	2) Jitter the 120 min delay time slightly so not all hosts slam
> 	   the RHN servers all at once.
> 	3) provide some very limited logging.
> 
> I've had the rhnsd program freeze on me enough times that I no longer use it.
> Instead, I have the following in cron, and it seems to work well:
> 	
> 	0 */2 * * * perl -e 'sleep rand(3600)'; /usr/sbin/rhn_check	

This is the first report of rhnsd freezing on someone that I have heard.
Do you have any more information to provide? If it happens again, can
you strace it for me?

And thank you for doing the randomization in your perl script. :)

-- 
____________
 /odd Warner                                    <taw@{redhat,pobox}.com>
             Geek Herder - QA/Sust-Eng/Rel-Eng - Red Hat Network
---------------------gpg info in the message headers--------------------
"Four legs good; two legs bad."




More information about the rhn-users mailing list