I found out that our firewalls have session timeouts, thus the jabber connections were timing out. (This also causes the strange behavior of a clients 'netstat -an |grep 5222' command saying "ESTABLISHED" even though there is no connection on the spacewalk server.)
After doing the following, I am having no issues with timeouts... ---------------------- sed -i 's/<interval>.*/<interval>120<\/interval>/' /etc/jabberd/*.xml* sed -i 's/<keepalive>.*/<keepalive>120<\/keepalive>/' /etc/jabberd/*.xml* sed -i 's/<idle>.*/<idle>600<\/idle>/' /etc/jabberd/*.xml* |grep '<idle>' rhn-satellite restart Restarted osad on all clients (This step may not be necessary.) ---------------------- The above settings are all explained in the .xml configs under /etc/jabberd/ . Essentially, it will send a "whitespace" character as a keepalive. (Every 2 minutes in my case) I verified this in tcpdump and jabber does only send 1 small packet every 2 minutes. Of course depending on your firewall timeout configs, you may need to tweak this a bit. Hope this helps someone else. -Josh On Fri, 2009-11-27 at 14:52 -0500, Mullis, Josh (CCI-Atlanta) wrote: > > Thanks guys. > > Appreciate the info. > > > I think this issue has to do with the clients being across multiple > routers and possible network blips. > > I have no issues with clients on the same subnet. > > > > > > ______________________________________________________________________ > > From: spacewalk-list-boun...@redhat.com > To: spacewalk-list@redhat.com > Sent: Thu Nov 26 08:13:13 2009 > Subject: Re: [Spacewalk-list] question about jabber server and osad > clients > > > I don't know if it's a know bug, but I do know that this behavior was > seen in both RHN4.x and RHN5.x . > > There are 2 ways you can deal with this: > > - cron job that restarts the osad daemon at X interval. > - cron job that checks if the osad daemon is connected to it's server > and restarts it when not.... > > I don't have the script for the second solution, but I remember it to > be quiet easy (40 lines or so). One thing to know is that you need to > check if the osad daemon is enabled or not, otherwise you might end up > in starting a lot of osad daemons where you don't want them. > > regards > > Jouk > > On Thu, Nov 26, 2009 at 1:57 PM, Colin Coe <colin....@gmail.com> > wrote: > > I saw this behavior in Satellite v5.1 at a previous client > site... My > workaround was restarting osad on the clients daily. > > CC > > > On Thu, Nov 26, 2009 at 8:49 PM, Michiel van Es > <michi...@info.nl> wrote: > > Anyone? > > Is this a known bug or others then Josh and myself are > experiencing? > > Is there a setting we should change in > the /etc/sysconfig/rhn/osad file? > > > > Kind regards, > > > > Michiel > > > > > > -------- Original Message -------- > > Subject: Re: [Spacewalk-list] question about jabber server > and osad clients > > From: Michiel van Es <michi...@info.nl> > > To: josh.mul...@cox.com <josh.mul...@cox.com>, > spacewalk-list@redhat.com > > <spacewalk-list@redhat.com> > > Date: 11/25/2009 03:49 PM > > > >> > >> Same here... especially the netstat -an | grep 5222 on the > client showing > >> the ESTABLISHED is strange.. > >> > >> > >> -------- Original Message -------- > >> Subject: [Spacewalk-list] question about jabber server and > osad clients > >> From: Mullis, Josh (CCI - Atlanta) <josh.mul...@cox.com> > >> To: spacewalk-list@redhat.com <spacewalk-list@redhat.com> > >> Date: 11/25/2009 03:46 PM > >> > >>> To build on this... > >>> > >>> I am having the same issue. > >>> More on the side of the client not responding to actions > from the > >>> spacewalk server after a random period of time. > >>> > >>> (NTP is in sync on both) > >>> > >>> > >>> > >>> Set debug to 10 on client yesterday and get nothing in the > logs when it > >>> stops responding. > >>> (Nothing in "osa-dispatcher" log on server either) > >>> > >>> > >>> > >>> Also, I noticed that on the client side that netstat shows > an > >>> "ESTABLISHED" connection to the spacewalk server. > >>> > >>> But netstat on the spacewalk server does not show a > connection from the > >>> client. > >>> > >>> > >>> > >>> > >>> > >>> > >>> > >>> > >>> On Wed, 2009-11-25 at 08:17 -0500, Michiel van Es wrote: > >>>> > >>>> Hi, > >>>> > >>>> We seem to habe a small problem with our Jabber server on > spacewalk and > >>>> the osad client on the spacewalk nodes. > >>>> > >>>> Before we reboot the spacewalk server we see a 104 > connections connected > >>>> with local port 5222 (jabber). > >>>> Whenever we reboot the Spacewalk server the connections > are limited to 3 > >>>> (local connections). > >>>> It seems that the client connections aren't reconnecting > if the > >>>> connection is lost. > >>>> > >>>> Is there a setting in /etc/sysconfig/osad or such on the > client side > >>>> that we have to alter to force clients/nodes to reconnect > if the > >>>> spacewalk server is down? > >>>> > >>>> Kind regards, > >>>> > >>>> Michiel > >>>> > >>>> _______________________________________________ > >>>> Spacewalk-list mailing list > >>>> Spacewalk-list@redhat.com > <mailto:Spacewalk-list@redhat.com> > >>>> https://www.redhat.com/mailman/listinfo/spacewalk-list > >> > >> _______________________________________________ > >> Spacewalk-list mailing list > >> Spacewalk-list@redhat.com > >> https://www.redhat.com/mailman/listinfo/spacewalk-list > > > > _______________________________________________ > > Spacewalk-list mailing list > > Spacewalk-list@redhat.com > > https://www.redhat.com/mailman/listinfo/spacewalk-list > > > > > > > -- > RHCE#805007969328369 > > > _______________________________________________ > Spacewalk-list mailing list > Spacewalk-list@redhat.com > https://www.redhat.com/mailman/listinfo/spacewalk-list > > >
_______________________________________________ Spacewalk-list mailing list Spacewalk-list@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-list