[lvs-users] Defunt Process

Henrique Fernandes sf.rique at gmail.com
Mon Aug 9 21:56:15 BST 2010


But the points is to use the checktimeout toghetter with my script.

My script does not consider timeout, or it is to long, i raise the timeout
and the problem stop, but with this solution, someother tie it will have the
checktime higher than checktimeout, then it will became defunct again.

But the point is that i wnat to use the timeout of ldirectord, you are
saying i can't use togetter with an script ?

So i am gonna probally set time out per virtual section and set this
especific to 0, an dlet my script hanlde timeout.

There is no other solution ? Like ldiretord kill the process properlly so
they don't became defunct?

Thanks!

[]'sf.rique


On Mon, Aug 9, 2010 at 5:50 PM, Simon Horman <horms at verge.net.au> wrote:

> On Mon, Aug 09, 2010 at 05:32:11PM -0300, Henrique Fernandes wrote:
> > My test is a python script that sends an email!
> >
> > If it sends the email ok! with exit with 0 anything else exits with 1
> >
> > The test is working ok!
> >
> > The only problem is when thet test takes more than the checktimeout, in
> that
> > case, my script becomes an defaunt proccess.
> >
> > If you need anything else let me know!
>
> Hi Henrique,
>
> I'm a little confused. The purpose of the checktimeout is to time-out in
> such situations. So I suggest increasing its value to something more
> appropriate to your script. You can set it on a per-virtual or global
> basis. A value of 0 should disable the timeout (in this case at least).
> Can you see if this resolves the problem you are seeing?
>
> >
> > checktimeout=4
> > checkinterval=2
> > autoreload=yes
> > logfile="/var/log/ldirectord.log"
> > #logfile="local2"
> > quiescent=yes
> > fork = yes
> >
> > virtual=X.X.X.X:25
> >         netmask=255.255.255.128
> >         real=10.0.1.10:25 masq
> >         real=X.X.X.X masq
> >         real=10.0.1.11:25 masq
> >         service=smtp
> >         protocol=tcp
> > #       checktype=connect
> >         checktype=external
> >         checkcommand="/usr/local/bin/checa-email.py"
>
> e.g.
>        checktimeout=30
>
> >
> > virtual=X.X.X.X:143
> >         netmask=255.255.255.128
> >         real=10.0.1.10:143 masq
> >         real=10.0.1.11:143 masq
> >         real=X.X.X.X:143 masq
> >         service=imap
> >         protocol=tcp
> >         checktype=negotiate
> >         login="xxxxxx"
> >         passwd="xxxxxx"
> >
> > []'sf.rique
> >
> >
> > On Mon, Aug 9, 2010 at 5:25 PM, Simon Horman <horms at verge.net.au> wrote:
> >
> > > On Mon, Aug 09, 2010 at 05:17:56PM -0300, Henrique Fernandes wrote:
> > > > I have a problem with heartbeat and ldirectord
> > > >
> > > > I have made a script that checks if some services are function. But
> the
> > > > problem is when this check takes longes that ldiretord timeout
> > > >
> > > > If that happes, the system creates an defunct proccess, and this list
> > > keeps
> > > > growing.
> > > >
> > > > does any one have an idea ?
> > > >
> > > > I am current using heartbeat and ldiretorc from centos repository!
> > > >
> > > > centos5
> > > > Linux Director v1.186-ha-2.1.3
> > >
> > > Hi Henrique,
> > >
> > > could you give a little bit more details about what
> > > your test is and what your ldirectord.cf looks like?
> > >
> > >
> > _______________________________________________
> > Please read the documentation before posting - it's available at:
> > http://www.linuxvirtualserver.org/
> >
> > LinuxVirtualServer.org mailing list - lvs-users at LinuxVirtualServer.org
> > Send requests to lvs-users-request at LinuxVirtualServer.org
> > or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>



More information about the lvs-users mailing list