New instances stuck unable to run puppet (and no sshing in!)
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	yuvipanda
	Jun 10 2015, 12:31 AM

Description

Found on the console of a newly created instance that I couldn't ssh into (ores-redis-01)

[1;31mError: Could not request certificate: Connection refused - connect(2)[0m``` over and over again

Mentioned In: T94866: /etc/ssh/userkeys/ubuntu or /etc/ssh/userkeys/admin notices for every puppet run on labs instances

yuvipanda raised the priority of this task from to Unbreak Now!.

yuvipanda updated the task description. (Show Details)

This is because the instance is trying to hit virt1000 for its initial run:

IP 10.68.18.32.37826 > 208.80.154.18.8140: Flags [S], seq 3234121293, win 29200, options [mss 1460,sackOK,TS val 743043 ecr 0,nop,wscale 7], length 0

IIRC the name a fresh puppet install tries to use is 'puppet' so switching that to the new master should fix it.

hmm, so you can't actually log in before initial puppet run, so that's going to be problematic, I guess. Do we need to build fresh images?

Presumably, once 'puppet' is switched to the new master in DNS the new instances will just work (possibly after a reboot)

I called andrew, and it looks like new images do need to be built. For now turning on puppetmaster on virt1000 as well.

Yup, that works for now! Thanks @Andrew!

New images were built.

Restricted Application added a subscriber: Matanya. · View Herald TranscriptJul 24 2015, 3:59 AM