yahoo-eng-team team mailing list archive

[Rodolfo Alonso <2045237@xxxxxxxxxxxxxxxxxx>]

*** This bug is a duplicate of bug 2039550 ***
    https://bugs.launchpad.net/bugs/2039550

Hello:

I can confirm that this issue is the same as reported in [1]. That was
fixed in [2] (up to Wallaby). Please use a version containing this patch
(if that version is still not released, please propose a
"openstack/releases" patch bumping the version).

Regards.

[1]https://bugs.launchpad.net/neutron/+bug/2039550
[2]https://review.opendev.org/q/topic:%22bug/2039550%22


** This bug has been marked a duplicate of bug 2039550
   [IPAM] During port bulk creation, if a "fixed_ip" request is incorrect, the previous IPAM allocations generated are not deleted

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/2045237

Title:
  [bulk] there is no  clean operation if bulk create ports fails

Status in neutron:
  Confirmed

Bug description:
  When we use the bulk api to create ports use one subnet e.g., '935fe38e-743f-45e5-a646-4ebbbf16ade7', and then we found any errors occurred.
  We use sql queries to find diffrent IPs between ipallocations and ipamallocations:

  ip_address in ipallocations but not in ipamallocations:

  MariaDB [neutron]> select count(*) from ipallocations where subnet_id='935fe38e-743f-45e5-a646-4ebbbf16ade7' && ip_address  not in (select ip_address from ipamallocations where ipam_subnet_id in (select id from  ipamsubnets where neutron_subnet_id='935fe38e-743f-45e5-a646-4ebbbf16ade7')) ;
  +----------+
  | count(*) |
  +----------+
  |      873 |
  +----------+
  1 row in set (0.01 sec)

  ip_address in ipamallocations but not in ipallocations:

  MariaDB [neutron]> select count(*)  from ipamallocations where ipam_subnet_id in (select id from  ipamsubnets where neutron_subnet_id='935fe38e-743f-45e5-a646-4ebbbf16ade7') && ip_address not in (select ip_address from ipallocations where subnet_id='935fe38e-743f-45e5-a646-4ebbbf16ade7');
  +----------+
  | count(*) |
  +----------+
  |       63 |
  +----------+
  1 row in set (0.01 sec)

  It seems that there are still resources remaining if ports creation
  fails, and we cannot find the operation to clean up failed ports.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/2045237/+subscriptions