AFAIK, by default VCS doesn't understand the underlying applications... so
I'd think it would be safe to do (thus one cluster node could be running
different applications from another... doing the application level clustering
requires end user work/config).
"Chris Cox" <ccox_nop...@airmail.net> wrote in message
news:MfSdnSrnqI76BLHV...@posted.internetamerica...
It is certainly *possible* to add a new Type and then resource to VCS
while online. However, it's not necessarily trivial. Are you just
adding a new instance (resource) of an existing agent, or are you adding
an entirely new agent (Type) to the cluster?
In many cases it's easier to shut down the cluster (not the running
applications), then restart with a modified main.cf.
But you wouldn't want to do either one on a production cluster without
testing or knowing what you were doing. Understand how it works, test
it, then put it in place.
--
Darren
"Darren Dunham" <ddu...@taos.com> wrote in message
news:p53Xj.1994$r82...@nlpi069.nbdc.sbc.com...
If you say "VCS server" and "agent" in the same post, I'm going to
assume you're talking about a "VCS agent".
If you install random software on a VCS server that has no interaction
with the cluster (like sendmail or SSH), then VCS doesn't care.
--
Darren