[rdo-users] OVN Errors after Upgrade

[James Hirst]

Hi All,

I'm new to this mailing list and wasn't quite sure where to ask this
question; I recently upgraded to Victoria from Ussuri and am using OVN for
networking. One of my compute nodes suddenly has an issue where any
instances launched on it are not able to access the provider network. I
have confirmed the issue is not the security group configuration, or a
firewall on the compute node.

 I ran a tcpdump on the compute node and am able to see the packets to/from
the test cirros instance I launched, so it is receiving the replies: (I
also cannot ping in the other direction either)

18:46:49.232241 fa:16:3e:3b:3d:7f > fc:ec:da:4a:0b:75, ethertype 802.1Q
(0x8100), length 102: vlan 20, p 0, ethertype IPv4, 172.16.19.36 >
172.16.20.24: ICMP echo request, id 53249, seq 3, length 64
18:46:49.232699 fc:ec:da:4a:0b:75 > fa:16:3e:3b:3d:7f, ethertype 802.1Q
(0x8100), length 102: vlan 20, p 0, ethertype IPv4, 172.16.20.24 >
172.16.19.36: ICMP echo reply, id 53249, seq 3, length 64
18:46:50.232777 fa:16:3e:3b:3d:7f > fc:ec:da:4a:0b:75, ethertype 802.1Q
(0x8100), length 102: vlan 20, p 0, ethertype IPv4, 172.16.19.36 >
172.16.20.24: ICMP echo request, id 53249, seq 4, length 64
18:46:50.233207 fc:ec:da:4a:0b:75 > fa:16:3e:3b:3d:7f, ethertype 802.1Q
(0x8100), length 102: vlan 20, p 0, ethertype IPv4, 172.16.20.24 >
172.16.19.36: ICMP echo reply, id 53249, seq 4, length 64


I checked logs from neutron-server, libvirt, nova-compute, and
ovn-controller; the only interesting logs that I can see are these in
ovn-controller.log:
2020-11-23T17:38:14.169Z|00055|lflow|WARN|Dropped 31 log messages in last
74 seconds (most recently, 37 seconds ago) due to excessive rate
2020-11-23T17:38:14.169Z|00056|lflow|WARN|error parsing actions "ct_commit
{ ct_label.blocked = 1; }; /* drop */": Syntax error at `{' expecting `;'.
2020-11-23T17:39:28.165Z|00057|lflow|WARN|Dropped 141 log messages in last
74 seconds (most recently, 74 seconds ago) due to excessive rate
2020-11-23T17:39:28.165Z|00058|lflow|WARN|error parsing match "reg0[8] == 1
&& (outport == @pg_b98b63c9_cda5_4c9b_8d66_881f71b51644 && ip4 && tcp &&
tcp.dst == 2380)": Syntax error at
`@pg_b98b63c9_cda5_4c9b_8d66_881f71b51644' expecting port group name.
2020-11-23T17:39:28.166Z|00059|lflow|WARN|Dropped 15 log messages in last
74 seconds (most recently, 74 seconds ago) due to excessive rate
2020-11-23T17:39:28.166Z|00060|lflow|WARN|error parsing actions "ct_commit
{ ct_label.blocked = 1; }; /* drop */": Syntax error at `{' expecting `;'.

I'm not exactly sure what these are or where I can find more information on
them, but I can see that they are new since the upgrade. The other compute
nodes in the cloud are fine, its just this one for some reason. I already
tried to rebuild the ovsdb conf.db file on this node, but that didn't seem
to change anything.

Help would very much be appreciated, if I am asking in the wrong place,
please let me know where I should pose this question. Also, is there an RDO
Slack or something similar? I know there is for Kubernetes/Openshift, but I
couldn't find one for RDO.

Thanks in advance,
James H
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rdoproject.org/pipermail/users/attachments/20201123/4ef170de/attachment.html>