[openstack-dev] bug deputy report (week of Jan 22)

Akihiro Motoki amotoki at gmail.com
Tue Jan 30 13:57:52 UTC 2018


Hi neutrinos,

Sorry for a bit late report of the bug deputy of last week.
I think there are several number of interesting bugs reported.

----

[Needs attention]

https://bugs.launchpad.net/neutron/+bug/1745642 SG hybrid iptables
driver and FWaaS OVS driver create overlapping conntrack zones

MTU topics on ovs-agent
https://bugs.launchpad.net/neutron/+bug/1744101 vxlan interfaces doesn't get MTU
https://bugs.launchpad.net/neutron/+bug/1745150 neutron doesn't set MTU on ovs
I am not sure it depends on deployments or not. We need to evaluate
these carefully, while the priority haven't been decided.

https://bugs.launchpad.net/neutron/+bug/1746000 dnsmasq does not
fallback on SERVFAIL
It is still Undecided status. The solution of dropping strict-order
option sounds reasonable to me, but I haven't understood why
strict-order of dnsmasq is required.

https://bugs.launchpad.net/neutron/+bug/1745468 Conntrack entry
removal can take a long time on large deployments
Brian is working on this, but it is worth attracted for the release.

https://bugs.launchpad.net/neutron/+bug/1745386 Update FloatingIP to
set QoS policy on it fails
It is medium priority, but it is worth fixed as this is one of new
features in queens. The fix is already proposed.

https://bugs.launchpad.net/neutron/+bug/1745443 cannot restrict
/var/lib/neutron permissions
I am not sure we can allow 'dnsmasq' user to access
/var/lib/neutron/dhcp instead of /var/lib/neutron.
More input on SElinux(?) would be appreciated.

[tricky bugs]

https://bugs.launchpad.net/neutron/+bug/1745412
test_l3_agent_scheduler intermittent failures when running locally

[Open question]

Neutron haproxy logs are not being collected
https://bugs.launchpad.net/devstack/+bug/1744359
-> Is there anything remaining? The merged fix looks complete to me.

[Closes Bug]

Many fullstack tests failing because of some error in L3 agent
https://bugs.launchpad.net/neutron/+bug/1745013
-> Nice finding, Slawek. It was eventlet bug.
http://lists.openstack.org/pipermail/openstack-dev/2018-January/126580.html



More information about the OpenStack-dev mailing list