|
Shaun McCullagh wrote:
>
> Hi,
>
> Thanks to Horms, LVS/DR now works fine on our testbed setup.
>
> I've been testing the failover system. Again this appears to work fine, but
> there are two oddities:
>
> 1) I keep getting this in ha-debug
>
> >heartbeat: 2000/10/17_12:10:09 debug: master_status_process: node
> >>[fazz.information-innovation.com] failed authentication
>
> 2) pppd keeps reporting this in /var/log/messages on both LVS machines at 1
> second intervals
>
> >Oct 17 13:13:40 fizz pppd[12809]: pppd 2.3.11 started by root, uid 0
> >Oct 17 13:13:40 fizz pppd[12809]: Device ttyS0 is locked by pid 12258
> >Oct 17 13:13:40 fizz pppd[12809]: Exit.
I think this due to the fact that you have "serial /dev/ttyS0" _and_
"ppp-udp dev/ttyS0 10.0.0.1" in your ha.cf file. I assume that heartbeat
opens ttyS0 for a generic serial heartbeat and then trys to open that
device again for a ppp-heartbeat. Solution: decide wether you want to
use generic serial or ppp heartbeat or use both but on different serial
ports.
[snip]
> ======================= haresources ===================================
>
> vs1.information-innovation.com IPaddr::192.168.31.157/24
> ldirectord::ldirectord.cf
>
> ========================================================================
In the haresources file should be a line starting with the hostname
(same as uname -n) of the primary HA-node. In your setup this would be
fizz.information-innovation.com and not vs1... because vs1 is the name
of the service IP that you share between your HA nodes. So your
haresources file should look like this:
fizz.information-innovation.com 192.168.31.157 ldirectord
There is no need to specify the netmask if it is the usual one for your
net class.
And to the following ha.cf:
Only include the nodes wich are part of the ha cluster (fizz and fazz)
in your ha.cf. Do not include any other (real-)server!
By the way:
ha.cf, haresources and authkeys must be identical on both nodes! (just
copy them over to be sure)
> =============================== ha.cf (fizz)
> ==================================
>
> debugfile /var/log/ha-debug
> logfile /var/log/ha-log
> keepalive 1
> deadtime 3
> serial /dev/ttyS0
--- and here I assume the ppp problem ----
- change it ttyS1 or comment it out -
> ppp-udp /dev/ttyS0 10.0.0.1
> baud 9600
> udp eth0
> nice_failback on
> node fizz.information-innovation.com
> node fazz.information-innovation.com
--- the following are wrong ---------------------
> node mayday.information-innovation.com
> node bacteria.information-innovation.com
> node napoleon.information-innovation.com
> node vs1.information-innovation.com
Greetings,
Juri
--
juri.haberland@xxxxxxxxxxxxx
system engineer innominate AG
clustering & security networking people
phone: +49-30-308806-45 fax: -77 http://innominate.de
|
