[Linux-HA] opinions on simple 2 node heartbeat setup

[George H]

I have a similar setup but with less network interfaces, mainly eth1
for heartbeat + all other network stuff, eth0 dedicated for DRBD.

If I pull the network plug on one of the nodes the other takes over,
but I notice that heartbeat runs the drbddisk script before DRBD it
self declares the remote host dead and the resource never gets taken
over. I tried messing with the timeouts from DRBDs side but the most I
can get it is to wait 6 seconds before cutting it off.  Heartbeat
fails over very quickly, I had to put in a 10 second delay script
before it runs the drbddisk (that way DRBD can properly wait 6 seconds
to go stand alone). FYI I am talking about after the other node gets
STONITHed, drbd needs to wait like 6 seconds before getting the broken
pipe error, and only then will running the drbddisk script work.

Do you get the same problem or solved it in a different way?


On 2/7/07, Ralf Gross <Ralf-Lists at ralfgross.de> wrote:
> Hi,
>
> I've a simple setup here. Two server, one is primay with samba/CVS/SVN/Apache,
> the other one is secondary and mainly our backup2tape server. Because the
> secondary server is idle most of the time, I set it up as secondary for
> drbd and heartbeat.
>
> Debian Etch, heartbeat-2 2.0.7-2, drbd 0.7.21
>
> Main purpose is that we are able to switch the samba shares to the secondary
> server if something bad happends with the primary.
>
> Both server are equipped with 4 GbE NICs.
>
> * eth0: client vlan
> * eth1: backup vlan
> * eth2, eth3: heartbeat vlan (2 x NICs, linux xor bonding, bond0)
>
> The heartbeat lan is used for drbd and hearbeat.
>
> x.60.9.1 is the switch ip. Both server are located in the same vlans, but in
> different buildings.
>
> The config below is working, but I would like to get some feedback if there are
> any pitfalls that I may have overlooked.
>
> ha.cf
>
> debugfile /var/log/ha-debug
> logfile /var/log/ha-log
> deadtime 30
> warntime 10
> initdead 120
> bcast   bond0
> auto_failback off
> node    VU0EM003
> node    VU0EM005
> ping    x.60.9.1
> respawn hacluster /usr/lib/heartbeat/ipfail
>
> haresources
>
> VU0EM003 IPaddr::x.60.9.211/24/eth0 \
>   drbddisk::r0 \
>   drbddisk::r1 \
>   drbddisk::r2 \
>   drbddisk::r3 \
>   LVM::volg1   \
>   Filesystem::/dev/volg1/server::/server/::xfs::noatime \
>   Filesystem::/dev/volg1/public::/public/::xfs::noatime,quota,grpquota \
>   Filesystem::/dev/volg1/home::/home/::xfs::noatime,quota,grpquota \
>   samba
>
>
>
> Ralf
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>


-- 
"Nothing is impossible for the person that doesn't have to do it"
"The probability of anything happening is in inverse ratio to its desirability"
"If I were a roman statue, I'd be made alabastard"
--
George H
george.dma at gmail.com