grickert wrote:
Thanks David,
Let me be a little more verbose, and hopefully figure this out.
I went into the linode manager for TEST and assigned the Private IP. I had read the Static Networking Guide and as indicated restarted networking, and successfully pinged both IP's from a TEST ssh session.
I then went to a DEV shell, and attempted to ping them with no success to the PrivateIP. I then started searching around and found somewhere that a reboot was required, so I rebooted TEST,
At this point you should have been good on TEST, at least as far as its access to the private network. Prior to that reboot the private address would not have been reachable from any other Linode. It worked locally since the guest OS recognized the internal address as itself, and it never had to leave the local Linode through the host.
Quote:
and am still not able to ping the PrivateIP from DEV.
From your reply, I now believe that DEV will need to be booted. Now before I boot it, I am going to assign and configure a private IP for DEV.
If I'm reading this right, does that mean that through this point you had not yet assigned a private IP on DEV (both in the manager and on the guest itself) and then rebooted? If not, then DEV would certainly not be able to reach TEST over the private network since it simply didn't have an interface on that network. But yes, I would expect that if you assign a private address to DEV in the manager, update it's local guest configuration to match and then reboot, that it should be able to now reach TEST.
Quote:
Now after I boot DEV, should I be able to ping DEV from TEST without rebooting TEST again?
Yes. The reboot would be to ensure DEV had a working private interface, which TEST already has from its own earlier reboot.
Quote:
Any insight into what is actually happening with this process to help me understand/absorb this would be appreciated.
Thanks again for all the help.
Gary
Not sure how best to clarify further, but it's a question of the private address being working both internally on your guest and then externally on the host. So for each Linode, the process is essentially:
- Add a private address in the manager. This updates your configuration but makes no changes to live systems.
- Configure your Linode. This lets your guest OS recognize the private network traffic when it arrives or is generated. (You can also do this step after the reboot, but then will need to refresh the configuration locally such as restarting networking to avoid another reboot).
- Reboot your Linode. When it restarts, the host obeys the manager configuration and understands to map traffic for the assigned private address to your Linode.
It's not a perfect analogy, but you could think of the host as a router on the private network. Without it knowing about your private address (which it picks up during the first reboot after assigning the private address in the manager) you can do stuff locally on your Linode but can't reach the outside world (where outside in this case is just the local data network).
-- David