On Jan 23, 2008, at 4:52 PM, Andraz Tori wrote:
Thanks for your fast answer, and a few more questions:
it seems that clones need to know their ips at configuration time.
However
since I plan using ec2-alike service, ips begin to get known at the
time
when node is actually started, can i notify linuxha of the exact ip
at some
later point in time (when specific script starts to wake the node up)?
if you dont want the cluster to manage the IPs as resource(s), then
you dont need to tell us at all
it was just another other option
And can i change clone_max in real time?
sure
bye
andraz
On Jan 23, 2008 12:03 PM, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
On Jan 22, 2008, at 10:03 AM, Andraz Tori wrote:
I am looking into Linux-HA to see if I could implement a system I
need. I
hope this is the right mailing list to ask if LinuxHA is the right
solution
for my challenge.
As a computing infrastructure I have some kind of cloud service
(imagine
EC2), what I want to do is the following:
There are 5 different types of backend services (resources) in the
system
and one frontend. However I need multiple resources of the same type
running
at any time.
http://linux-ha.org/v2/Concepts/Clones
So I need LinuxHA to take care of having 3 resources of the
same type available at any point in time. Hopefully I can send
change
messages so I can increase or decrease the number of resources of
the same
type that need to be available at any point in time (maybe from 3 to
5 for a
few days and then scale down to two) .
Sure, just change the value of clone_max
Then I have my own frontend deamon which users all those 5 backend
resource
types. I want this daemon to query LinuxHA every 10 seconds (via
some API)
and get information about where the resources are available (list ip
numbers
where resources are located for each resource type) so application
gets the
necessary information to do the load balancing.
IIRC, you could also do the load balancing with a cloned IPaddr2
resource.
Or colocate a number of known IPs with the application clone.
The last thing i need is a failover for the frontend application,
but I can
easily imagine how to do that with LinuxHA... However I need some
clues
about the proper way to do what I described earlier.
Thank you very much for your help
bye
Andraz Tori
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
