** Description changed: juju2 beta 5 on xenial: When new lxd containers are launched on the local provider a) the PTR record initially points to 'ubuntu.lxd' (not the machine name) b) After launching another container, it points to nothing -- the second container gets 'ubuntu.lxd' c) sometime else, after circumstances I can't explain, containers get the correct juju-UUID hostname in the PTR record. + UPDATE: I think circumstances I can't explain == follow-up DHCPREQ/ACK + Why this is bad: rabbit is extremly sensitive to hostnames and the charm actually relies on the data returned in the reverse lookup. When it changes, rabbit refuses to start and hook errors ensue. See the following reverse lookups from the same lxd provdier serving a few different models and notice the 3 different cases I outline above. - ubuntu@juju-cc280254-dba7-433a-8247-06f7154d0649-machine-16:~$ dig -x 10.73.33.23 ; <<>> DiG 9.9.5-3ubuntu0.8-Ubuntu <<>> -x 10.73.33.23 ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 62399 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0 ;; QUESTION SECTION: ;23.33.73.10.in-addr.arpa. IN PTR ;; Query time: 1 msec ;; SERVER: 10.73.33.1#53(10.73.33.1) ;; WHEN: Tue Apr 26 05:45:33 UTC 2016 ;; MSG SIZE rcvd: 42 ubuntu@juju-cc280254-dba7-433a-8247-06f7154d0649-machine-16:~$ dig -x 10.73.33.51 ; <<>> DiG 9.9.5-3ubuntu0.8-Ubuntu <<>> -x 10.73.33.51 ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 19088 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 1280 ;; QUESTION SECTION: ;51.33.73.10.in-addr.arpa. IN PTR ;; ANSWER SECTION: 51.33.73.10.in-addr.arpa. 0 IN PTR juju-d11a663a-1d44-415c-8c4c-123e92adf5d1-machine-3.lxd. ;; Query time: 0 msec ;; SERVER: 10.73.33.1#53(10.73.33.1) ;; WHEN: Tue Apr 26 05:45:36 UTC 2016 ;; MSG SIZE rcvd: 122 ubuntu@juju-cc280254-dba7-433a-8247-06f7154d0649-machine-16:~$ dig -x 10.73.33.179 ; <<>> DiG 9.9.5-3ubuntu0.8-Ubuntu <<>> -x 10.73.33.179 ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 36202 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 1280 ;; QUESTION SECTION: ;179.33.73.10.in-addr.arpa. IN PTR ;; ANSWER SECTION: 179.33.73.10.in-addr.arpa. 0 IN PTR ubuntu.lxd. ;; Query time: 0 msec ;; SERVER: 10.73.33.1#53(10.73.33.1) ;; WHEN: Tue Apr 26 05:45:41 UTC 2016 ;; MSG SIZE rcvd: 78 - You can reproduce this easily on the local lxd provider by doing juju deploy -n 3 cs:trusty/ubuntu though... I can't easily get c) to come up. I'm not sure how it's triggered. At some point, lxd updates the PTR record correctly, and I'm not sure when yet. Also, try deploying rabbit. sometimes it works, sometimes it doesn't. Also, if it does work, try restarting the daemon process, it will typically break.
-- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1574963 Title: juju2 lxd launch hostname reverse lookup inconsistent To manage notifications about this bug go to: https://bugs.launchpad.net/juju-core/+bug/1574963/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
