On 4 Jan 2007, at 18:45, James Wade wrote:

I’ve looked through hobbitclient.cfg and it has the following:

BBDISP="192.168.0.218" # IP address of the Hobbit server

BBDISPLAYS="" # IP of multiple Hobbit servers.
BBDISP must be "0.0.0.0".

So, Do I set BBDISPLAYS to:

BBDISPLAYS=”192.168.0.219” (For my other Hobbit Server) Set:

BBDISP="0.0.0.0"

BBDISPLAYS="ip.address.for.svr1 ip.address.for.svr2"

Note space not commas, and quotes around both. The 0.0.0.0 for BBDISP
is the marker for multiple servers, and BBDISPLAYS contains the IP's
for the servers. This is the same for all clients (including the
~hobbit/client/etc/hobbit on the servers themselves!).

Will the clients then send data to both servers? How do you account
for paging? Yes, it will. Do you set one machine as the primary and turn-off the other machine?

If one machine fails, can you enable the other to start paging. Paging, as in notifications? I currently have a different paging rule
set on the 'backup' box, which just alerts on hobbit server things
being down, ie:

HOST=myprimaryhobbithost SERVICE=bbd,bbgen,bbtest COLOR=red SCRIPT /home/hobbit/bin/sms_notify ADMINS

(Which sends an SMS)

Easiest is to keep the primary hobbit-alerts.cfg handy on the
secondary box. There might be a better, or more intelligent way. One
idea I've considered to have some check in the notification script
that periodically checks, and then just drops notifications if it CAN
fetch status from the primary host, but sends them if it CAN'T, ie.
if the primary host is down, the external 'SCRIPT' alert rules kick
in. See bb(1) manpage, specifically wrt. 'query'. But I really
haven't given this much thought, and I wouldn't be surprised if
Henrik or somebody else has a better solution.

Any pointers would be really appreciated.

Hope that helps,

Regards,

Richard

-- Richard Leyton - richard at leyton.org http://www.leyton.org