[ale] RedHat HA Assistance
scott mcbrien
smcbrien at gmail.com
Wed Sep 12 13:30:32 EDT 2012
That would be my suggestion. Mark it as the lest preferred in your
failover domain, so that it won't get assigned the service in a
failover. It could even be a VM ;-) But yes, I'd recommend that over
a qdisk. In my experience, qdisks are flaky.
-Scott
On Wed, Sep 12, 2012 at 1:27 PM, Sam Davis <aracthabar at gmail.com> wrote:
> They do reboot each other when the cluster is up and running. Would a very
> small third node with an extremely low priorty be preferential over a qdisk?
>
> Sam
>
>
> On Wed, Sep 12, 2012 at 1:23 PM, Scott McBrien <smcbrien at gmail.com> wrote:
>>
>> It's because your cluster is split-brained. Best option would be to add a
>> third node. If you can't do that, you could add a qdisk *wretches*.
>> Seriously, third node.
>>
>> Also, I the machines aren't constantly rebooting each other, I'd guess
>> your fencing isn't working as well.
>>
>> -Scott
>>
>> On Sep 12, 2012, at 1:15 PM, Sam Davis <aracthabar at gmail.com> wrote:
>>
>> > Hello All,
>> >
>> > I'm wondering if anyone had expirence with RedHat's HA solution.
>> > We're trying to setup a 2 node failover cluster for a Drupal install using
>> > multipath disks attached via HBAs. It seems to work fine at the start, but
>> > after a while neither node can take ownership of the clustered volume group.
>> > At this point, even trying to stop the clvmd service just hangs. Any hints
>> > or pointers would be appreciated. Below is the cluster.conf
>> >
>> >
>> > Thank you,
>> > Sam Davis
>> >
>> >
>> >
>> > <?xml version="1.0"?>
>> > <cluster config_version="25" name="moodle_cluster">
>> > <clusternodes>
>> > <clusternode name="mario.REDACTED.REDACTED" nodeid="1">
>> > <fence>
>> > <method name="RSA">
>> > <device name="Mario_RSA"/>
>> > </method>
>> > </fence>
>> > </clusternode>
>> > <clusternode name="luigi.REDACTED.REDACTED" nodeid="2"/>
>> > </clusternodes>
>> > <fencedevices>
>> > <fencedevice agent="fence_ipmilan" auth="password"
>> > ipaddr="REDACTED" login="admin" name="Mario_RSA" passwd="REDACTED"/>
>> > <fencedevice agent="fence_ipmilan" auth="password"
>> > ipaddr="REDACTED" login="admin" name="Luigi_RSA" passwd="REDACTED"/>
>> > </fencedevices>
>> > <rm>
>> > <failoverdomains>
>> > <failoverdomain name="Mario_Preferred"
>> > nofailback="1" ordered="1">
>> > <failoverdomainnode
>> > name="mario.REDACTED.REDACTED" priority="1"/>
>> > <failoverdomainnode
>> > name="luigi.REDACTED.REDACTED"/>
>> > </failoverdomain>
>> > </failoverdomains>
>> > <resources>
>> > <ip address="REDACTED" sleeptime="10"/>
>> > <clusterfs device="/dev/vg_moodle/moodle_lv"
>> > force_unmount="on" fsid="27336" fstype="gfs2" mountpoint="/srv"
>> > name="moodle_gfs" options="rw,noatime,_netdev"/>
>> > <script file="/etc/init.d/httpd" name="Httpd"/>
>> > </resources>
>> > <service domain="Mario_Preferred" name="moodle_service"
>> > recovery="relocate">
>> > <clusterfs ref="moodle_gfs"/>
>> > <ip ref="REDACTED"/>
>> > <script ref="Httpd"/>
>> > </service>
>> > </rm>
>> > <cman expected_votes="1" two_node="1"/>
>> > </cluster>
>> >
>> > _______________________________________________
>> > Ale mailing list
>> > Ale at ale.org
>> > http://mail.ale.org/mailman/listinfo/ale
>> > See JOBS, ANNOUNCE and SCHOOLS lists at
>> > http://mail.ale.org/mailman/listinfo
>>
>> _______________________________________________
>> Ale mailing list
>> Ale at ale.org
>> http://mail.ale.org/mailman/listinfo/ale
>> See JOBS, ANNOUNCE and SCHOOLS lists at
>> http://mail.ale.org/mailman/listinfo
>
>
>
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo
>
More information about the Ale
mailing list