I have the following type of scenario. I have established a resource
group which does a number of infrastructure type things: sets IP
addresses, sets correct password files, sets correct system links and
mount points, etc. I now have a bunch of primitives that I want to
colocate with this resource group, but I do not want them part of the
group because I do not want the failure/reset of one to cause the others
to reset.
For example, I'm running a FTP server, a HTTP server, a SMTP server, and
an OpenDAP backend server on the server that runs the resource group. Now
if I simply add them to the existing resource group in the order they're
listed above, then if the HTTP server gets reset during a monitor action,
it will cause the SMTP server and OpenDAP backend to shutdown and then
restart them after it restarts the HTTP server -- even if those 2 are not
actually dependent on the HTTP server. This is how resource groups work,
I get that.
What I want is a configuration where each of those four servers is
colocated with the resource group "in parallel" (for lack of a more
accurate term) such that a reset of one of the four servers follwing a
failed monitor action does not trigger a reset of any other resource. So
would the following configuration work or would it create a 'resource set'
(extraneous information such as params stripped for ease of reading):
primitive ip ocf:heartbeat:IPaddr2 params ip="1.2.3.4"
primitive job ocf:pps:jobfile params role="test" job="first"
primitive pwd ocf:pps:pwdfile params role="test"
primitive ftpd ocf:pps:proftpd
primitive httpd ocf:heartbeat:apache
primitive smtpd ocf:heartbeat:postfix
primitive bes ocf:pps:besServer
group infra_group ip job pwd
colocation inf_ftpd inf: ftpd infra_group
colocation inf_http inf: httpd infra_group
colocation inf_mail inf: smtpd infra_group
colocation inf_http inf: httpd infra_group
colocation inf_mail inf: smtpd infra_group
colocation inf_odap inf: bes infra_group
Before commenting on if this will work as I want as stated above, did I
get the colocation order correct in the statements? In other words, the
infrastructure group must be up and running prior to starting any of the
other servers (ftpd|httpd|smtpd|bes), so are those lines correct? Do I
need to reverse it, i.e.:
colocation inf_ftpd inf: infra_group ftpd
or do I need to use an 'order' statement instead, i.e.:
order ftp_infra mandatory: infra_group:start ftpd
So once the issue of whether my basic colocation (or order) statements are
correct is resolved, the next issue is will this do what I want?
For example, assume the following failure scenarios (i.e. the monitor
returns not_running and attempts to restart the named primitive):
'ftpd' fails a monitor check
Is it the only resource that is restarted?
Or do all the subsequent colocation resources do it too?
'job' fails a monitor check
So the 'pwd' resource gets stopped, then 'job' gets
stopped and started, then 'pwd' gets started due to the
resource group.
But do all (ftpd|httpd|smtpd|bes) of the servers with
their individual colocation statements also get reset?
Thanks very much for your help.
Tony
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
