| Summary: | When one of the leading zeros of the last IPv6 quartet is omitted the public VIP gets set on all 3 controllers | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat OpenStack | Reporter: | Marius Cornea <mcornea> | ||||
| Component: | rhosp-director | Assignee: | Dan Sneddon <dsneddon> | ||||
| Status: | CLOSED WONTFIX | QA Contact: | yeylon <yeylon> | ||||
| Severity: | high | Docs Contact: | |||||
| Priority: | unspecified | ||||||
| Version: | 7.0 (Kilo) | CC: | athomas, bfournie, dsneddon, jcoufal, mburns, mcornea, morazi, rhel-osp-director-maint, srevivo | ||||
| Target Milestone: | y3 | ||||||
| Target Release: | 7.0 (Kilo) | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Known Issue | |||||
| Doc Text: |
Cause: IPv6 addresses should be expressed either with no leading zeros, or with enough leading zeroes to pad out the full address.
Consequence: If an IPv6 is supplied with only some of the leading zeroes, this will cause the IP address to be interpreted incorrectly.
Workaround (if any): Always use full leading zeroes, or do not use leading zeroes. For instance, the IPv6 address 2001:db8::1 may also be expressed as 2001:db8:0001, but 2001:db8:01 will fail.
Result: You must pay attention that IPv6 addresses are specified in the above way, or deployment may fail as a result.
|
Story Points: | --- | ||||
| Clone Of: | |||||||
| : | 1302081 (view as bug list) | Environment: | |||||
| Last Closed: | 2016-02-07 16:03:35 UTC | Type: | Bug | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Bug Depends On: | |||||||
| Bug Blocks: | 1302081 | ||||||
| Attachments: |
|
||||||
We should add a 7.3 release note for this. Dan, can you provide the release note text here? (In reply to Marius Cornea from comment #0) Marius, It sounds like this does not occur if the leading zero is not omitted? So, if we tell people to use this: [{'start': '2001:db8:fd00:1000:dead:beef:cafe:0001', 'end': '2001:db8:fd00:1000:dead:beef:cafe:0009'}] Instead of this: [{'start': '2001:db8:fd00:1000:dead:beef:cafe:001', 'end': '2001:db8:fd00:1000:dead:beef:cafe:009'}] Then it works? I need to understand what the workaround is to help write the release note. Yes, when providing all the zeros it works as expected: 2001:db8:fd00:1000:dead:beef:cafe:0001 |
Created attachment 1118437 [details] /var/log/messages Description of problem: When one of the leading zeros of the last IPv6 quartet is omitted the public VIP gets set on all 3 controllers. Version-Release number of selected component (if applicable): openstack-tripleo-heat-templates-0.8.6-112.el7ost.noarch How reproducible: 100% Steps to Reproduce: 1. Deploy IPv6 enabled overcloud with the following external allocation pool: ExternalInterfaceDefaultRoute: 2001:db8:fd00:1000:dead:beef:cafe:f00 ExternalNetCidr: '2001:db8:fd00:1000::/64' ExternalAllocationPools: [{'start': '2001:db8:fd00:1000:dead:beef:cafe:001', 'end': '2001:db8:fd00:1000:dead:beef:cafe:009'}] Actual results: stack@instack:~>>> for ctrl in $(nova list | grep controller | awk {'print $12'} | awk -F'=' {'print $2'}); do ssh heat-admin@$ctrl 'cat /etc/hostname; /usr/sbin/ip a s dev vlan10 | grep inet6 | grep 2001';done overcloud-controller-0.localdomain inet6 2001:db8:fd00:1000:dead:beef:cafe:1/64 scope global tentative dadfailed inet6 2001:db8:fd00:1000:dead:beef:cafe:3/64 scope global overcloud-controller-1.localdomain inet6 2001:db8:fd00:1000:dead:beef:cafe:1/64 scope global tentative dadfailed inet6 2001:db8:fd00:1000:dead:beef:cafe:2/64 scope global overcloud-controller-2.localdomain inet6 2001:db8:fd00:1000:dead:beef:cafe:1/64 scope global inet6 2001:db8:fd00:1000:dead:beef:cafe:4/64 scope global ip-2001.db8.fd00.1000.dead.beef.cafe.001 (ocf::heartbeat:IPaddr2): Stopped Expected results: The VIP resource is started and the IP address gets set only on one of the nodes at a time. Additional info: Attaching /var/log/messages.