Hi Pavlos!

Am Dienstag, 30. November 2010 22:28 schrieb Pavlos Parissis:
> Hi Nikita,
>
> On 30 November 2010 08:42, Nikita Michalko <[email protected]> 
wrote:
> > Hi Pavlos,
> >
> > Am Dienstag, 30. November 2010 05:59 schrieb Pavlos Parissis:
> >> On 29 November 2010 23:43, Lars Ellenberg <[email protected]> 
wrote:
> >> > On Mon, Nov 29, 2010 at 10:24:17PM +0800, Mia Lueng wrote:
> >> >> Hi:
> >> >> I have configured a cluster with two nodes.    Lan setting is
> >> >> A
> >> >> eth0:  192.168.10.110
> >> >> eth1: 172.16.0.1
> >> >>
> >> >> B
> >> >> eth0:192.168.10.111
> >> >> eth1: 172.16.0.2
> >> >>
> >> >> I have configured a resource ip_0  192.168.10.100  on eth0. But when
> >> >> I unplug the eth0 link on A,  the resource can not be taken over to B
> >> >> and no any log output.
> >> >>
> >> >> I've checked the /usr/lib/ocf/resource.d/heartbeat/IPaddr2 script and
> >> >> found there are  no codes for nic link status checking.
> >> >>
> >> >> How can i monitor the nic link status to protect the virtual ip
> >> >> address? Thanks.
> >> >
> >> > http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Expl
> >> >ain ed/ch09s03s03.html
> >>
> >> I am sorry this is not the expected behaviour, at least to me. I
> >> expect from the IPaddr2 to report a failure in a case the interface is
> >> not available. What is the point to maintain an IPaddr2 resource on
> >> interface which is not up? Furthermore, using the ping resource adds
> >> more complexity and basically utilizes Layer 3 protocol in order to
> >> monitor a layer 2 device ( the NIC). Using ip link show on the device
> >> should a very easy way to check link status on the NIC, ip tool
> >> supports also LOWER_UP.
> >
> >  - what about configure monitor operation of IP in cib.xml - sth. like
> > this: <resources>
> >      <primitive id="IPaddr_194_37_40_42" class="ocf" provider="heartbeat"
> > type="IPaddr">
> >         <meta_attributes id="primitive-IPaddr_194_37_40_42meta"/>
> >          <operations>
> >            <op name="monitor" interval="60s" id="IPaddr_194_37_40_42_mon"
> > timeout="60s"/>
> >          </operations>
> >
> > - it works for me very well ;-)
>
> Mia has reported on this thread that having monitor enabled is not
> enough for reporting problems on the link.
> What do you mean works for you? Have you tried to pull out the network
> cable from an interface on which you have an IPaddr2 resource running?

- I use the "IPaddr" RA, not IPaddr2. And of course did I test it thoroughly - 
succesfully ;-)
 You can also use "ifconfig ethx down " to simulate cable pull out  ...

> Does that action cause a failure on the IPaddr2 resource?

-  on IPaddr YES!
>
> Cheers,
> Pavlos
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems

-- 
AIP - Dr. Nikita Michalko | Main office at: | Working hours: 8:00-16:30 
Tel:   +43 1 408 35 57-14 | A -1160 Wien    | except on Fri: 7:30-14:30
Fax: +43 1 408 35 57-26   | Grundsteing. 40 | [email protected]
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to