Re: [Testbed-admins] Errors booting admin mfs

[Barry Trent <btrent@triticom.com>]

Success, at last.

It's ironic that you told Chris Benninger in the other active support 
thread here to check out the "overlapping diagnostics" on my thread, but 
it was actually something you told HIM that led to my solution.

You told him:

> But, I can tell you how tmcc figures out who boss is ... it looks to
> see who its DNS server. If this fails:
>
> pc> /usr/local/etc/emulab/tmcc.bin bossinfo

And I realized that I had presumed that pc1 would reverse DNS itself 
(which I got to work). It's actually a reverse lookup of boss which is 
the key, and I hadn't put the right reverse record into the "head" file 
for my control net reverse zone.

I added reverse lookups for boss and ops and the admin mfs boots up fine 
on pc1 (at least much better than before!). I'll give it a closer look 
on Monday, but I think I'm over another hump. I presume I should add 
reverse records for all the various static names/nodes on the control net?

Thanks for all your help! I'm guessing this won't be my last post here :)

PS - An unrelated question: I changed the config of the sshd on boss to 
one closer to the setup I normally use (only rsa authentication allowed 
for logins). Is that likely to cause me any problems? I know that boss 
ssh's into other testbed nodes, and that seems to work fine. Does anyone 
ssh into boss as part of normal testbed operations?