2008/8/28 Gianluca Cecchi <[EMAIL PROTECTED]>:
> Hello,
> I have 2.1.4 installed on RH 5.2
> My ha.cf contains
> auto_failback off
>
> and my haresources contains
> nfsnode2 drbddisk::drbd-resource-0 \
>        Filesystem::/dev/drbd0::/drbd0::ext3 \
>        Filesystem::/drbd0/nfsapps::/nfsapps::ext3::bind \
>        IPaddr::10.1.1.103/24/eth0 \
>        killnfsd \
>        nfslock \
>        nfs
>
> without crm=on and running in 1.x style, the file is parsed ok.
>
> Converting to 2.x style I note these behaviours:
>
> 1) The IPaddr line is not correctly translated. I get nic and netmask values
> inverted.
> The same if in 1.x style I put IPaddr::10.1.1.103/eth0/24
>
>         <primitive class="ocf" id="IPaddr_10_1_1_103" provider="heartbeat"
> type="IPaddr">
>           <operations>
>             <op id="IPaddr_10_1_1_103_mon" interval="5s" name="monitor"
> timeout="5s"/>
>           </operations>
>           <instance_attributes id="IPaddr_10_1_1_103_inst_attr">
>             <attributes>
>               <nvpair id="IPaddr_10_1_1_103_attr_0" name="ip" value="
> 10.1.1.103"/>
>               <nvpair id="IPaddr_10_1_1_103_attr_1" name="nic" value="24"/>
>               <nvpair id="IPaddr_10_1_1_103_attr_2" name="cidr_netmask"
> value="eth0"/>
>             </attributes>
>           </instance_attributes>
>         </primitive>
>

This seems to be a bug, 2.1.4 is not recent, you can get newer
packages here (I think you should remove your current package instead
of upgrading):
http://download.opensuse.org/repositories/server:/ha-clustering/RHEL_5/

>
> I have to correct the file and then start heatbeat
> BTW, what is the difference between IPaddr and IPaddr2?

IPaddr --> uses ifconfig
IPaddr2 --> uses iproute2 (only supports Linux)

Probably other relevant differences that i'm not aware...

>
> 2) It seems that the translated 2.x configuration has auto failback on.
> Infact I have nfsnode2 that is DC and is running resources.
> I shutdown it and nfsnode1 carries on the services.
> But when I start again nfsnode2 it takes the resources again.
> How to prevent this?

I think that this configuration attribute does that:

name="default-resource-stickiness" value="INFINITY"

> I see the rsc_location tab set up by the converison script
>
>     <constraints>
>       <rsc_location id="rsc_location_nfsdida1" rsc="nfsdida1">
>         <rule id="prefered_location_nfsdida1" score="100">
>           <expression attribute="#uname"
> id="prefered_location_nfsdida1_expr" operation="eq" value="nfsnode2"/>
>         </rule>
>       </rsc_location>
>     </constraints>
>
> Any hints and/or pointers on this and on constraints set up in general?
>
> Thanks in advance,
> Gianluca

Regards,
Ciro
_______________________________________________
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