Feature #10721
closedSoftware deployment of new Sepia community lab (typica machines)
0%
Description
There is a vm setup (pulpito-rdu.front.sepia.ceph.com) for the pulpito setup and the new main tythology node is typica002 (setup similarly to octo).
Nodes that should be added to the machine database:
typica003-> typica192
And typica177-192 should be marked down as those machines are currently being used for training. I will update locks appropriately when stuff is up.
Updated by Zack Cerza about 9 years ago
- Status changed from New to In Progress
- Priority changed from Normal to High
The existing labs are behaving better these days so I'm going to start working on this.
Updated by Zack Cerza about 9 years ago
We're going to need paddles-rdu.front.sepia.ceph.com to point to the same host as pulpito-rdu
Updated by Zack Cerza about 9 years ago
- Status changed from In Progress to 12
DNS is wonky. I'm blocked and don't know how to fix it.
ubuntu@typica002:~$ hostname typica002 ubuntu@typica002:~$ dnsdomainname sepia.ceph.com ubuntu@typica002:~$ host typica002.sepia.ceph.com Host typica002.sepia.ceph.com not found: 3(NXDOMAIN) ubuntu@typica002:~$ host typica002.front.sepia.ceph.com typica002.front.sepia.ceph.com has address 172.20.133.2 ubuntu@typica002:~$ host pulpito-rdu.front.sepia.ceph.com pulpito-rdu.front.sepia.ceph.com has address 172.20.132.3
ubuntu@typica001:~$ hostname typica001 ubuntu@typica001:~$ dnsdomainname front.sepia.ceph.com ubuntu@typica001:~$ host typica001.front.sepia.ceph.com Host typica001.front.sepia.ceph.com not found: 3(NXDOMAIN) ubuntu@typica001:~$ host typica002.front.sepia.ceph.com Host typica002.front.sepia.ceph.com not found: 3(NXDOMAIN) ubuntu@typica001:~$ host pulpito-rdu.front.sepia.ceph.com Host pulpito-rdu.front.sepia.ceph.com not found: 3(NXDOMAIN)
Updated by Dan Mick about 9 years ago
Some investigation: I think typica002's domain is wrong; it should be front.sepia.ceph.com, and I think it originates from the /etc/hosts entry for "uname -h" (typica002)
Lab DNS is PowerDNS on the 'gw1' machine, cached by dnscache01234. typica002 is in that db as typica002.front.sepia.ceph.com.
typica001 seems to have both a redhat.com and sepia nameserver in /etc/resolv.conf; not sure why.
ubuntu@typica001:~$ dig +short typica001.front.sepia.ceph.com
ubuntu@typica001:~$ dig +short @172.20.132.1 typica001.front.sepia.ceph.com
172.20.133.1
Updated by Zack Cerza about 9 years ago
Thanks for that, Dan!
I fixed typica002's domain, and at least temporarily fixed typica001's resolv.conf. Where's this 'gw1' machine? I believe I need access to it.
Updated by Zack Cerza about 9 years ago
- Status changed from 12 to In Progress
With Dan's help I got past the issue I was having and am now running preliminary tests in the lab:
http://pulpito-rdu.front.sepia.ceph.com/teuthology-2015-02-26_16:38:19-smoke-master---basic-typica/
Shouldn't be much longer now.
Updated by Zack Cerza about 9 years ago
64 nodes are unreachable and I don't see any evidence of ipmi being set up :-/
Updated by Zack Cerza almost 9 years ago
- Status changed from In Progress to Resolved