I am running some vms with vmware workstation and while I am getting my job done, I know there are some things I am just screwing up.
I have two "setups" - one is a hadoop cluster, 3 VMs working together, which I have set up on CentOS 6.3 images. When I originally set it up, I think I got everything working fine. However, now for the second time, I think my networking settings are screwed up. I set the VMs up with Bridged mode, then went into the individual VMs, tweaked the hostname, /etc/hosts, and a few other files. I saw what ifconfig reported, added those entries as the ones on my vms in /etc/hosts, and things were running smoothly.
However, at one point after a reboot, the IPs changed (seems like the they were reassigned by DHCP, which I totally understand).
What I am guessing now is that I should've configured the VMs to operate in NAT instead of bridged mode, or perhaps added entries on the host computer's (Win 7) /etc/hosts file and/or created static routes on my home network's router.
Is there a best practice for setting up networking in cases like mine. Right now it looks like I will have to grab the new assigned lease from ifconfig, plug that back into the etc/hosts files, and restart the network interfaces on the VMs. Simple enough, but there's extra work on the hadoop cluster itself for either reconfiguring or deleting services and nodes and readding.
Obviously, there must be a better way than my caveman approach!
thanks for any help,
Probably just ignorance on my part...in the VMs themselves, I edit the /etc/host files and make sure they are all the same (3 nodes with same settings)...it's the networking outside of the VMs where I think I've screwed things up...can certainly assign static IPs from the router and have those match the networking on the VMs...I'm just a hack who only gets by with a smattering of sys admin knowledge