yahoo-eng-team team mailing list archive

[OpenStack Infra <1625334@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/372794
Committed: https://git.openstack.org/cgit/openstack/neutron/commit/?id=14fee9cfcc816128606fc11d33800af90a0b1859
Submitter: Jenkins
Branch:    master

commit 14fee9cfcc816128606fc11d33800af90a0b1859
Author: Carl Baldwin <carl.baldwin@xxxxxxx>
Date:   Mon Sep 12 15:31:08 2016 -0600

    Try to reuse existing IPs when a port update specifies subnet
    
    If a port update specifies only a subnet_id for a fixed_ip then we
    want to look at existing fixed_ips to see if that subnet_id is already
    there. This avoids allocating a new IP address on the subnet and
    deallocating the old one.
    
    Without some special care, this breaks the code path for prefix
    delegation. One could argue that PD needs reworking. However, as a
    stop-gap measure, we still run the old code path if the address is an
    EUI-64 address. This allows PD to continue to work as it was
    originally written and it doesn't do any harm because allocating
    EUI-64 addresses is repeatable.
    
    This commit removes a test case from the DNS integration tests. The
    test was specifically testing that DNS records we updated in the case
    where a subnet id was passed to re-allocate a fixed_ip. Since the use
    case no longer works, the test doesn't make sense.
    
    This commit also preserves the ability to add an additional IP from a
    subnet for which the port already has IPs.
    
    Change-Id: Iba5d54efa7f99ed82275ffc8e5be975b373c29d3
    Related-Bug: #1622616
    Closes-Bug: #1625334


** Changed in: neutron
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1625334

Title:
  Update port with subnet_id in fixed_ips allocates a new IP when
  existing one could be used.

Status in neutron:
  Fix Released

Bug description:
  This issue has been seen twice causing this related bug [1]. What
  happens is the DHCP agent updates the port using the subnet_id but not
  the actual ip_address that the port already has. So, the server
  allocates a new IP address and throws out the old one.

  What it should do is recognize that there is already an IP address on
  the port that satisfies the request and avoid the churn.

  A previous attempt was made [1] to address this bug but was reverted
  because it had a side effect [3].  Need a fix that addresses this
  issue without the side-effect.

  [1] https://bugs.launchpad.net/neutron/+bug/1622616/
  [2] https://review.openstack.org/#/c/369051/
  [3] https://bugs.launchpad.net/neutron/+bug/1623800

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1625334/+subscriptions