[Linux-HA] Set-up help

[Brian R. Hellman]

>
> Some questions:
>
> 1.  Must I have a second physical interface on each box?
>   
No you can do it with one interface
> 2.  Must I create and bring up an aliased IP on node 1?
>   
Heartbeat will create the IP address  your haresources file should look
something like:

alpha IPaddr::192.168.5.2/24/eth0

> 3.  Can the SHA1 password be a plain text word, or must I generate one?
>   
Can be plain text or generated, your choice
> 4.  Should this work if I simply stop httpd on danvers (node 1)?  It doesn't.
>   
heartbeat manages httpd so stopping heartbeat or powering the node down
is the way to test, if you use v2 style you can setup monitoring to
restart a stopped service if you wish.
> 5.  Must httpd be running on both boxes?  It seems to get shut down on danvers 
> (node 1) when I start heartbeat on that machine.
>   
httpd will be started and stopped by heartbeat as necessary, so no it
does not need to be running on both.
> 6.  Do I need to install an other software?
>   
I think you are fine with what you have.
There is a simple heartbeat setup with IP only on section 2 of this
article.  It deals with DRBD, but the heartbeat section can still be
applied to what you are playing around with.

http://www.howtoforge.com/drbd-8.3-third-node-replication-with-debian-etch

Best regards,
Brian
> I'll probably think of more questions right after I hit the send key.
>   


Dimitri Yioulos wrote:
> Hello to all.
>
> I'm a Linux-HA newbie.  I've read several how-to's, etc., but am struggling 
> with getting HA to work for me.  I would most appreciate the community's help 
> both in understanding how HA works, and in getting my set-up operational.  
> Apologies for the length of my post.
>
> As a test/learning exercise, I've installed heartbeat-2.1.3-3, 
> heartbeat-pils-2.1.3-3, heartbeat-stonith-2.1.3-3, and heartbeat-gui-2.1.3-3 
> via RPM on a CentOS 4 and a CentOS 5 box.  Here is other relavant information 
> about hte two boxes:
>
> CentOS 4 - hostname=danvers (confirmed with uname -n), interface=eth0 
> 192.168.100.12, eth0:0 192.168.103.6(IP aliased)
> CentOS 5 - hostname=pembroke, interface=eth1 192.168.100.6
>
> Neither machine has a second physical interface (well, node 2 does, but it's 
> not operational), both are on the local area network 192.168.100.0/22.  Both 
> are running, e.g. apache.  All HA configuration-related files are the same on 
> both machines.
>
> Here are the configuration-related files:
>
> ha.cf:
>
> use_logd yes
> keepalive 500ms
> deadtime 2
> warntime 1
> initdead 8
> udpport 694
> bcast eth1
> node danvers
> node pembroke
> crm on
>
>
> haresources
>
> danvers IPaddr::192.168.103.6::httpd
>
>
> authkeys
>
> auth 1
> 1 sha1 mypassword
>
>
> output of crm_mon -i1:
>
> Defaulting to one-shot mode
> You need to have curses available at compile time to enable console mode
>
>
> ============
> Last updated: Tue Mar 10 13:00:58 2009
> Current DC: pembroke (675a48e3-3c9e-49ee-ba3c-2bf6dd107be3)
> 2 Nodes configured.
> 1 Resources configured.
> ============
>
> Node: pembroke (675a48e3-3c9e-49ee-ba3c-2bf6dd107be3): online
> Node: danvers (aa78e7bf-db5c-49a3-a201-875e0747f149): online
>
> resource_       (lsb:httpd):    Started pembroke  (note: I used heartbeat_gui 
> to create the resource, but mya have done that wrong.)
>
>
> I think I'm trying to failover apache.
>
>
> Once again, you're help is most appreciated.
>
> Dimitri
>
>