*** hujie11 has quit IRC | 01:39 | |
*** liuhaixia has quit IRC | 02:05 | |
openstackgerrit | yuan wei proposed openstack/dragonflow: Impliment security group fullstack and ut cases https://review.openstack.org/299105 | 03:39 |
---|---|---|
oanson | nick-ma: ping | 05:20 |
nick-ma | oanson: pang | 05:26 |
nick-ma | oanson: pong~~~~~~ | 05:26 |
oanson | Hi | 05:28 |
oanson | I wanted to ask about https://review.openstack.org/#/c/298121/ | 05:28 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Lookup router_port's network_id with lswitch's topic https://review.openstack.org/298121 | 05:32 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Do not log error on missing table during pub/sub test https://review.openstack.org/297198 | 05:54 |
*** irenab has joined #openstack-dragonflow | 05:59 | |
*** gampel1 has joined #openstack-dragonflow | 06:00 | |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Add publisher's table https://review.openstack.org/297167 | 06:07 |
gampel | yuli:ping | 06:11 |
gampel | yuli ping | 06:15 |
gampel | yuli ping | 06:17 |
raofei | yuli ping | 06:18 |
gampel | yuli: raofei will help you with the setup | 06:19 |
gampel | yuli: do you have a setup running with DVR enabled ? | 06:19 |
gampel | raofei: are we building the br-ex on the compute node as well in the devstack script | 06:20 |
gampel | yuli: ?? | 06:20 |
raofei | yes. actually, br-ex is created by df-controller currently. | 06:20 |
raofei | devstack script just set the neutron.conf for external network items. | 06:21 |
gampel | we need documentation for the end user how to add the physical nic to br-ex | 06:21 |
gampel | it can be in a different patch on our wiki on how to set up distributed DNAT | 06:22 |
raofei | ok. I think it user use devstack, then I will add a item in local.conf to assign physical nic to connect external network | 06:22 |
raofei | yes. I will do it. | 06:22 |
*** yuli_s has joined #openstack-dragonflow | 06:23 | |
yuli_s | Hello all | 06:23 |
gampel | I am not sure we need the physical nic in the devsatck but we need guide for the admin how to set it up | 06:23 |
raofei | At the beginning, i think br-ex and physical nic is managed by adminstrator | 06:23 |
raofei | OK | 06:23 |
gampel | yuli: raofei will help you with the setup whats the status of the testing ? | 06:24 |
raofei | The behaviour is similar with the original l3 feature, adminstrator manage the br-ex and physical network. | 06:24 |
yuli_s | my computer had powered off | 06:25 |
raofei | @Yuli, I send a message to you this morning | 06:25 |
yuli_s | raofei, gmail ? | 06:25 |
raofei | Before creating fip, you need 1)set the gateway into router. eg, neutron router-gateway-set router-demo ext-net; 2) then create fip. eg, neutron floatingip-create ext-net; | 06:25 |
raofei | just irc channel | 06:26 |
yuli_s | my computer was closed | 06:26 |
yuli_s | please send it to stremovsky@gmail.com | 06:27 |
raofei | when you physical service is recovery, please try to do it. | 06:27 |
raofei | this is same with centralized router. | 06:27 |
yuli_s | raofei, why public network is not created automatically now ? | 06:28 |
* yuli_s running stack | 06:29 | |
raofei | I need to verify it on my setup. it's created on my setup | 06:29 |
yuli_s | I am doing stack every day | 06:30 |
yuli_s | it just happened yesterday for the first time, | 06:31 |
gsagie | raofei: maybe send yuli_s your local.conf | 06:31 |
gsagie | so you can compare, maybe the problem is there | 06:31 |
raofei | Maybe it's because we disabled the l3-agent. but I have to check it again. | 06:31 |
yuli_s | I run stack several times yesterday and it was heppening all the time | 06:31 |
yuli_s | ok, give me your email, | 06:32 |
yuli_s | i will send you the file | 06:32 |
gsagie | also maybe its a different devstack version you are trying with | 06:32 |
gsagie | raofei: you work with master devstack/master neutron? | 06:32 |
raofei | yes. I use the master devstack | 06:32 |
raofei | but I haven't update it for 1 month. | 06:33 |
yuli_s | you better update | 06:34 |
raofei | I think if devstack cannot create external network, it will not impact our test. we can create network manually | 06:34 |
raofei | sure, I will update it and test it today. | 06:34 |
yuli_s | raofei, we have done that yesterday | 06:35 |
raofei | yes, and you found fip cannot be created successfully. | 06:36 |
yuli_s | btw | 06:36 |
yuli_s | what is your email, | 06:36 |
raofei | I checked your router configuration, it's because you didn't set the gateway. | 06:36 |
yuli_s | I will send you my local.conf | 06:36 |
raofei | milo.raofei@huawei.com | 06:37 |
raofei | OK, I will check it on my setup now | 06:37 |
yuli_s | done | 06:38 |
yuli_s | u use ubuntu ? | 06:39 |
raofei | ok, we are same | 06:39 |
yuli_s | if yes, remove the python api code | 06:40 |
yuli_s | before running stack | 06:40 |
yuli_s | in /usr/local/lib/python2.7/dist-packages/python_* | 06:40 |
yuli_s | in /usr/local/lib/python2.7/dist-packages/neutron* | 06:40 |
yuli_s | in /usr/local/lib/python2.7/dist-packages/oslo* | 06:41 |
yuli_s | when running "clean.sh" it is not removed | 06:41 |
raofei | do you mean need to clear the setup? | 06:42 |
raofei | yes, I always clear the setup by clean.sh | 06:42 |
yuli_s | yes, it just do not removes the library code | 06:42 |
yuli_s | as a result, new version might not work, so you remove the libraries code too, | 06:43 |
yuli_s | before running stack.sh | 06:43 |
yuli_s | and you can also run "git pull" in devstack installation dir | 06:44 |
yuli_s | to get the latest version | 06:44 |
yuli_s | ok, I got the stack up & running | 06:45 |
openstackgerrit | yuan wei proposed openstack/dragonflow: Impliment security group fullstack and ut cases https://review.openstack.org/299105 | 06:46 |
gsagie | yuli_s: i dont think its related to the libs | 06:47 |
yuli_s | yes, just he used an old version of code, | 06:47 |
yuli_s | i had to remove old libs a week ago to be able to run stack | 06:48 |
gsagie | raofei: i will try to run your patch as well, can you please send me your local.conf | 06:48 |
gsagie | that you were able to work with | 06:48 |
gsagie | to gal.sagie@gmail.com | 06:48 |
gsagie | and what other actions are you doing beside the stack? do you also connect the physical NIC to br-ex and change routing? | 06:49 |
yuli_s | again, I have a problem like yesterday : "Port to be associated" is empty ! | 06:57 |
yuli_s | ok, found a problem this time ! | 07:02 |
openstackgerrit | Nguyen Hung Phuong proposed openstack/dragonflow: Fix typos in Dragonflow files https://review.openstack.org/299167 | 07:06 |
yuli_s | from vm I can ping private gw | 07:08 |
yuli_s | no ping to 8.8.8.8 | 07:08 |
yuli_s | raofei, any ideas ?\ | 07:14 |
raofei | sorry, just see the message. I will send the email | 07:16 |
raofei | @Eran, yuli I sent a mail to you. | 07:23 |
raofei | @Eran, I create the br-ex and add a physical nic into br-ex(such as eth1). When df-controller get restart, it will create patch ports to connect br-int and br-tun. | 07:25 |
raofei | @Yuli, do you add the private network into router? | 07:26 |
yuli_s | yes | 07:28 |
yuli_s | i am checking now the tcpdump | 07:32 |
yuli_s | i see ping on tap device | 07:32 |
yuli_s | 10:32:52.980181 IP 10.0.0.3 > google-public-dns-a.google.com: ICMP echo request, id 22529, seq 98, length 64 | 07:33 |
yuli_s | nothing on br-ex | 07:34 |
yuli_s | sudo tcpdump -i br-ex ICMP | 07:34 |
yuli_s | sudo tcpdump -i br-ex icmp | 07:34 |
raofei | do you test on single node or mutiple node? | 07:35 |
yuli_s | single box | 07:35 |
raofei | if single node, you need disable q-l3 | 07:35 |
raofei | please dump flow of br-int | 07:36 |
yuli_s | #enable_service q-l3 | 07:36 |
raofei | show me the dump-flows | 07:36 |
yuli_s | sending it to your email | 07:36 |
raofei | where is you external gateway? | 07:37 |
yuli_s | good question | 07:38 |
yuli_s | let me check | 07:38 |
raofei | The problem is df controller cannot learn the arp from gateway. | 07:38 |
yuli_s | yes | 07:39 |
yuli_s | i do not remember that I specified anywhere the external gw | 07:40 |
yuli_s | today | 07:40 |
yuli_s | i have this in local.config: PUBLIC_NETWORK_GATEWAY=10.100.100.8 | 07:40 |
yuli_s | how to make sure it is used ? | 07:41 |
raofei | the external gateway is managed by administrator, you can just configure the 10.100.100.8 on some device of external network | 07:42 |
yuli_s | I have it is place | 07:42 |
yuli_s | br-ex has addr:10.100.100.8 | 07:43 |
openstackgerrit | Merged openstack/dragonflow: OVSDB monitor to filter port update events https://review.openstack.org/297188 | 07:43 |
yuli_s | raofei, what you advise me to do ? | 07:44 |
yuli_s | i do not see any icmp packets on br-ex | 07:45 |
raofei | you need to tcpdump whether br-ex send arp to br-int | 07:45 |
raofei | you just ping fip -I 10.100.100.8 | 07:45 |
raofei | to trigger the arp | 07:46 |
raofei | or you can do as below. tcpdump -i br-ex -n arp | 07:47 |
raofei | and then restart df-controller | 07:47 |
raofei | we need to capture the arp request from controller on br-ex | 07:47 |
yuli_s | 10:47:38.556176 ARP, Request who-has 10.100.100.51 tell 10.100.100.1, length 46 | 07:48 |
yuli_s | 10:47:39.558349 ARP, Request who-has 10.100.100.50 tell 10.100.100.1, length 46 | 07:48 |
yuli_s | 10:47:39.559549 ARP, Request who-has 10.100.100.49 tell 10.100.100.1, length 46 | 07:48 |
yuli_s | do not see arp request from the vm on br-ex | 07:48 |
raofei | what's you vm ip? what's the fip? | 07:48 |
raofei | and show me your ext-subnet | 07:49 |
yuli_s | vm ip: 10.0.0.3 | 07:49 |
yuli_s | fip: 10.100.201.200 | 07:49 |
yuli_s | neutron net-show cb3f53f3-2694-4924-8b31-a7735906610c | 07:50 |
yuli_s | +-----------------+--------------------------------------+ | 07:50 |
yuli_s | | Field | Value | | 07:50 |
yuli_s | +-----------------+--------------------------------------+ | 07:50 |
yuli_s | | admin_state_up | True | | 07:50 |
yuli_s | | created_at | 2016-03-30T06:50:10 | | 07:51 |
raofei | what's the external subnet info? | 07:51 |
yuli_s | | description | | | 07:51 |
yuli_s | | id | cb3f53f3-2694-4924-8b31-a7735906610c | | 07:51 |
yuli_s | | is_default | False | | 07:51 |
yuli_s | | name | pub-external | | 07:51 |
yuli_s | | router:external | True | | 07:51 |
yuli_s | | shared | False | | 07:51 |
yuli_s | | status | ACTIVE | | 07:51 |
yuli_s | | subnets | 17c869eb-909c-464b-9572-e555cb044230 | | 07:51 |
yuli_s | | tags | | | 07:51 |
yuli_s | | tenant_id | 85104976665e47cfaefe8ad50e7de019 | | 07:51 |
yuli_s | | updated_at | 2016-03-30T06:50:10 | | 07:51 |
yuli_s | +-----------------+--------------------------------------+ | 07:51 |
yuli_s | neutron subnet-show 17c869eb-909c-464b-9572-e555cb044230 | 07:51 |
yuli_s | +-------------------+------------------------------------------------------+ | 07:51 |
yuli_s | | Field | Value | | 07:51 |
yuli_s | +-------------------+------------------------------------------------------+ | 07:51 |
yuli_s | | allocation_pools | {"start": "10.100.201.200", "end": "10.100.201.230"} | | 07:51 |
yuli_s | | cidr | 10.100.0.0/16 | | 07:51 |
yuli_s | | created_at | 2016-03-30T06:51:18 | | 07:51 |
yuli_s | | description | | | 07:51 |
yuli_s | | dns_nameservers | | | 07:51 |
yuli_s | | enable_dhcp | True | | 07:51 |
yuli_s | | gateway_ip | 10.100.0.1 | | 07:51 |
yuli_s | | host_routes | | | 07:51 |
yuli_s | | id | 17c869eb-909c-464b-9572-e555cb044230 | | 07:51 |
yuli_s | | ip_version | 4 | | 07:51 |
yuli_s | | ipv6_address_mode | | | 07:51 |
yuli_s | | ipv6_ra_mode | | | 07:52 |
yuli_s | | name | pub-subnet | | 07:52 |
yuli_s | | network_id | cb3f53f3-2694-4924-8b31-a7735906610c | | 07:52 |
yuli_s | | subnetpool_id | | | 07:52 |
yuli_s | | tenant_id | 85104976665e47cfaefe8ad50e7de019 | | 07:52 |
yuli_s | | updated_at | 2016-03-30T06:54:58 | | 07:52 |
yuli_s | +-------------------+------------------------------------------------------+ | 07:52 |
yuli_s | from vm I can not ping 10.100.0.1 | 07:52 |
raofei | the gateway ip is 10.100.0.1, but not 10.100.100.8 | 07:53 |
yuli_s | df-db dump | grep demo-router | 07:53 |
yuli_s | Key = e6c8cdd4-1bdc-4947-ae51-e9f81df26a45 , Value = {"topic": "85104976665e47cfaefe8ad50e7de019", "external_ids": {"neutron:router_name": "demo-router"}, "distributed": true, "name": "e6c8cdd4-1bdc-4947-ae51-e9f81df26a45", "ports": [{"name": "7e730086-e29d-4c97-adc8-dec93cc62821", "lswitch": "ff2cc4d6-e2ea-4ec7-a293-569e245300af", "topic": "85104976665e47cfaefe8ad50e7de019", "mac": "fa:16:3e:63:4f:95", "tunnel_key": 7, " | 07:53 |
yuli_s | lrouter": "e6c8cdd4-1bdc-4947-ae51-e9f81df26a45", "network": "10.0.0.1/24"}]} | 07:53 |
raofei | you need to configure the external gateway ip and it's 10.100.0.1 ! | 07:53 |
yuli_s | sec ;) | 07:53 |
yuli_s | neutron subnet-update 17c869eb-909c-464b-9572-e555cb044230 --gateway_ip 10.100.100.8 | 07:54 |
yuli_s | still no connectivity | 07:58 |
yuli_s | any ideas ? | 08:06 |
raofei | maybe you need to configure the gateway ip into a external device | 08:07 |
raofei | meanwhile you need to capture the arp request from controller on br-ex | 08:08 |
yuli_s | look here | 08:09 |
yuli_s | cookie=0x0, duration=3873.147s, table=20, n_packets=1823, n_bytes=175161, priority=100,ip,metadata=0x1,nw_src=10.0.0.3 actions=goto_table:30 | 08:09 |
yuli_s | i do not have table 30 ! | 08:10 |
yuli_s | EGRESS_NAT_TABLE = 30 | 08:11 |
yuli_s | is missing | 08:12 |
yuli_s | here: sudo ovs-ofctl dump-flows -O Openflow13 br-int | 08:13 |
*** gampel1 has quit IRC | 08:14 | |
raofei | it's because df controller CANNOT learn the gateway arp, so there is a rule missing. | 08:17 |
raofei | That's why I ask you to capture the arp request from controller on br-ex!!! | 08:18 |
raofei | @Yuli, please configure the gateway ip into an external network device. and add a physical nic into br-ex, and this physical nic connect the external network. | 08:27 |
openstackgerrit | hujie proposed openstack/dragonflow: Fix ovsdb monitor connect ovsdb fail bug https://review.openstack.org/299199 | 08:34 |
yuli_s | hm | 08:38 |
openstackgerrit | hujie proposed openstack/dragonflow: Fix ovsdb monitor connect ovsdb fail bug https://review.openstack.org/299199 | 08:39 |
yuli_s | sudo tcpdump -i br-ex arp | 08:39 |
yuli_s | 11:39:08.606370 ARP, Request who-has 10.100.100.8 tell 10.100.100.1, length 46 | 08:40 |
yuli_s | 11:39:08.606402 ARP, Reply 10.100.100.8 is-at 74:d4:35:e0:bb:b6 (oui Unknown), length 28 | 08:40 |
yuli_s | 11:39:13.672559 ARP, Request who-has 10.100.100.1 tell 10.100.100.8, length 28 | 08:40 |
yuli_s | 11:39:13.672802 ARP, Reply 10.100.100.1 is-at 84:c9:b2:62:89:e0 (oui Unknown), length 46 | 08:40 |
yuli_s | 11:39:32.726401 ARP, Request who-has 10.100.100.46 tell 10.100.100.1, length 46 | 08:40 |
yuli_s | 11:39:32.946487 ARP, Request who-has 10.100.100.8 tell 10.100.100.1, length 46 | 08:40 |
yuli_s | 11:39:32.946505 ARP, Reply 10.100.100.8 is-at 74:d4:35:e0:bb:b6 (oui Unknown), length 28 | 08:40 |
yuli_s | 11:39:56.530400 ARP, Request who-has 10.100.100.1 tell 10.100.100.35, length 46 | 08:40 |
yuli_s | 11:39:57.326556 ARP, Request who-has 10.100.100.8 tell 10.100.100.1, length 46 | 08:40 |
yuli_s | 11:39:57.326579 ARP, Reply 10.100.100.8 is-at 74:d4:35:e0:bb:b6 (oui Unknown), length 28 | 08:40 |
yuli_s | 10.100.100.1 is real gw | 08:40 |
yuli_s | changing gw in subnet to 10.100.100.1 | 08:42 |
yuli_s | gives better result | 08:43 |
yuli_s | 11:42:02.209539 ARP, Reply 10.100.0.3 is-at fa:16:3e:c8:da:c5 (oui Unknown), length 28 | 08:43 |
yuli_s | 11:42:02.608061 ARP, Request who-has 10.100.100.35 tell 10.100.100.1, length 46 | 08:43 |
yuli_s | 11:42:03.209436 ARP, Reply 10.100.0.3 is-at fa:16:3e:c8:da:c5 (oui Unknown), length 28 | 08:43 |
yuli_s | 11:42:04.209577 ARP, Reply 10.100.0.3 is-at fa:16:3e:c8:da:c5 (oui Unknown), length 28 | 08:43 |
yuli_s | 11:42:21.536514 ARP, Request who-has 10.100.100.8 tell 10.100.100.1, length 46 | 08:43 |
yuli_s | 11:42:21.536527 ARP, Reply 10.100.100.8 is-at 74:d4:35:e0:bb:b6 (oui Unknown), length 28 | 08:43 |
yuli_s | 11:42:23.756600 ARP, Request who-has 10.100.100.46 tell 10.100.100.1, length 46 | 08:43 |
yuli_s | 11:42:33.896445 ARP, Request who-has 10.100.100.1 tell 10.100.100.8, length 28 | 08:43 |
yuli_s | 11:42:33.897154 ARP, Reply 10.100.100.1 is-at 84:c9:b2:62:89:e0 (oui Unknown), length 46 | 08:43 |
raofei | I'm confused. please show me the latest ext-subnet | 08:43 |
yuli_s | i run this command | 08:44 |
yuli_s | neutron subnet-update 17c869eb-909c-464b-9572-e555cb044230 --gateway_ip 10.100.100.1 | 08:44 |
raofei | just now I send a mail to you | 08:44 |
yuli_s | neutron subnet-show 17c869eb-909c-464b-9572-e555cb044230 | 08:44 |
yuli_s | +-------------------+------------------------------------------------------+ | 08:44 |
yuli_s | | Field | Value | | 08:44 |
yuli_s | +-------------------+------------------------------------------------------+ | 08:44 |
yuli_s | | allocation_pools | {"start": "10.100.201.200", "end": "10.100.201.230"} | | 08:44 |
yuli_s | | cidr | 10.100.0.0/16 | | 08:44 |
yuli_s | | created_at | 2016-03-30T06:51:18 | | 08:44 |
yuli_s | | description | | | 08:44 |
yuli_s | | dns_nameservers | | | 08:44 |
yuli_s | | enable_dhcp | True | | 08:44 |
yuli_s | | gateway_ip | 10.100.100.1 | | 08:44 |
yuli_s | | host_routes | | | 08:44 |
yuli_s | | id | 17c869eb-909c-464b-9572-e555cb044230 | | 08:44 |
yuli_s | | ip_version | 4 | | 08:44 |
yuli_s | | ipv6_address_mode | | | 08:44 |
yuli_s | | ipv6_ra_mode | | | 08:44 |
yuli_s | | name | pub-subnet | | 08:44 |
yuli_s | | network_id | cb3f53f3-2694-4924-8b31-a7735906610c | | 08:44 |
yuli_s | | subnetpool_id | | | 08:44 |
yuli_s | | tenant_id | 85104976665e47cfaefe8ad50e7de019 | | 08:44 |
yuli_s | | updated_at | 2016-03-30T08:42:01 | | 08:45 |
yuli_s | +-------------------+------------------------------------------------------+ | 08:45 |
raofei | Where is the 10.100.100.1 configured? | 08:45 |
yuli_s | it is physical gw, outside | 08:45 |
raofei | ok, now it cannot work still? | 08:46 |
yuli_s | sec. | 08:46 |
yuli_s | table 30 is missing | 08:46 |
yuli_s | route -n | 08:47 |
yuli_s | Kernel IP routing table | 08:47 |
yuli_s | Destination Gateway Genmask Flags Metric Ref Use Iface | 08:47 |
yuli_s | 0.0.0.0 10.100.100.1 0.0.0.0 UG 0 0 0 br-ex | 08:47 |
yuli_s | 10.100.0.0 0.0.0.0 255.255.0.0 U 0 0 0 br-ex | 08:47 |
yuli_s | 192.168.122.0 0.0.0.0 255.255.255.0 U 0 0 0 virbr0 | 08:47 |
raofei | are you enable em1 up | 08:48 |
raofei | ifconfig em1 up | 08:48 |
raofei | do as below | 08:48 |
raofei | 1. tcpdump -i em1 -n arp | 08:48 |
raofei | 2. ping fip -I 10.100.100.1 on the outside gateway device. | 08:49 |
raofei | you need captured the arp request on the em1 firstly | 08:49 |
yuli_s | 11:51:00.544607 ARP, Request who-has 10.100.100.50 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | 11:51:00.546637 ARP, Request who-has 10.100.100.49 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | 11:51:00.548626 ARP, Request who-has 10.100.100.48 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | 11:51:01.550544 ARP, Request who-has 10.100.100.47 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | 11:51:01.552898 ARP, Request who-has 10.100.100.46 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | 11:51:01.555442 ARP, Request who-has 10.100.100.45 tell 10.100.100.1, length 46 | 08:51 |
yuli_s | is there a way for you to connect to my huawei box ? | 08:51 |
yuli_s | like in rdesktop ? | 08:52 |
raofei | it seems your gateway cannot connect to your setup | 08:52 |
raofei | you didn't capture the packet | 08:53 |
yuli_s | i will talk with the IT regarding the rdesktop | 08:55 |
yuli_s | be back in a minute | 08:55 |
yuli_s | i have talked with the IT manager | 08:58 |
openstackgerrit | Merged openstack/dragonflow: Add tests for L3 application https://review.openstack.org/297316 | 08:59 |
yuli_s | i think it should be private | 08:59 |
openstackgerrit | Merged openstack/dragonflow: Fix py3 mock cannot be compared to int https://review.openstack.org/298595 | 09:00 |
openstackgerrit | duankebo proposed openstack/dragonflow: Solve the problem that sometimes network can't be recycled due to some ports of that network are used by some vms https://review.openstack.org/297940 | 09:01 |
openstackgerrit | duankebo proposed openstack/dragonflow: k https://review.openstack.org/299226 | 09:13 |
*** hujie1 has joined #openstack-dragonflow | 09:15 | |
openstackgerrit | duankebo proposed openstack/dragonflow: Please enter the commit message for your changes. Lines starting https://review.openstack.org/296522 | 09:18 |
*** liuhaixia has joined #openstack-dragonflow | 09:23 | |
openstackgerrit | feipeng proposed openstack/dragonflow: Add Redis Devstack https://review.openstack.org/297916 | 09:51 |
yuli_s | gal is here | 10:00 |
yuli_s | i can ping 10.100.100.8 from other host | 10:01 |
yuli_s | ping to 10.100.201.200 from other server is not working | 10:01 |
yuli_s | gal we have a problem running ping from VM to outer world | 10:03 |
yuli_s | only ping to local gw works | 10:03 |
yuli_s | on public network subnet | 10:08 |
yuli_s | 1. i tried to use Ip assigned on br-ex | 10:08 |
yuli_s | 1. i tried to use Ip oh physical gw | 10:08 |
yuli_s | 2. i tried to use Ip of physical gw | 10:09 |
yuli_s | in openflow rules we have table 30 missing (EGRESS_NAT_TABLE = 30) | 10:09 |
yuli_s | <raofei> it's because df controller CANNOT learn the gateway arp, so there is a rule missing.\ | 10:10 |
yuli_s | so, as raofei suggested there is a problem with ARP packets | 10:11 |
yuli_s | somewhere in my lab | 10:11 |
yuli_s | so, we are trying to debug this | 10:12 |
raofei | if your setup is built on vm(with nested vm), then it's vey easy to set the external gateway. | 10:13 |
yuli_s | btw, what should I use as a gateway_ip in pub-subnet ? | 10:14 |
yuli_s | btw, what IP should I use as a gateway_ip in pub-subnet ? | 10:14 |
raofei | gateway_ip | 10.100.100.1 | 10:14 |
raofei | this is the external gateway ip | 10:15 |
yuli_s | ok great, | 10:15 |
raofei | btw, you need disable the q-l3 | 10:15 |
yuli_s | this IP I used for the last few hours | 10:15 |
yuli_s | it is disabled and not running | 10:16 |
gsagie | raofei: the controller learn the physical gateway using learn flows? | 10:17 |
gsagie | i guess you mean the physical gateway MAC | 10:17 |
gsagie | is there anyway that yuli_s can first hard code the MAC and try it out | 10:18 |
yuli_s | raofei, can you point me to Openflow rule where the df controller receives the gateway arp ? | 10:24 |
yuli_s | be back a bit later | 11:09 |
yuli_s | going to eat | 11:09 |
oanson | Hi, I'm dropping in on this :) | 11:34 |
oanson | shlomo@ubuntu:~$ arp -an | 11:34 |
oanson | ? (10.100.200.200) at <incomplete> on br-ex | 11:34 |
oanson | ? (10.100.100.17) at 3c:97:0e:e0:9a:c2 [ether] on br-ex | 11:34 |
oanson | ? (10.100.100.37) at 3c:97:0e:a4:4d:84 [ether] on br-ex | 11:34 |
oanson | ? (10.100.100.1) at 84:c9:b2:62:89:e0 [ether] on br-ex | 11:34 |
oanson | ? (10.100.201.200) at <incomplete> on br-ex | 11:34 |
oanson | shlomo@ubuntu:~$ | 11:34 |
hujie1 | Hi Omer, I have reviewed your code about ovsdb monitor filter br-int bug, I have a question, when ovsdb monitor receive the notify message from ovsdb, how could you make sure that the port info has been stored in the IDL cache? | 11:36 |
*** openstackgerrit has quit IRC | 11:47 | |
*** openstackgerrit has joined #openstack-dragonflow | 11:48 | |
oanson | Hi. It looks like the gw ip is defined badly. It should be 10.100.100.1. It looks like its configured to be 10.100.0.1. yuli_s will look into it in his env. | 12:04 |
yuli_s | ok | 12:11 |
yuli_s | in df db | 12:11 |
yuli_s | Key = cb3f53f3-2694-4924-8b31-a7735906610c , Value = {"topic": "85104976665e47cfaefe8ad50e7de019", "subnets": [{"dhcp_ip": "10.100.0.2", "enable_dhcp": true, "lswitch": "cb3f53f3-2694-4924-8b31-a7735906610c", "dns_nameservers": [], "gateway_ip": "10.100.100.1", "cidr": "10.100.0.0/16", "id": "17c869eb-909c-464b-9572-e555cb044230"}], "external_ids": {"neutron:network_name": "pub-external"}, "name": "cb3f53f3-2694-4924-8b31 | 12:12 |
yuli_s | -a7735906610c"} | 12:12 |
yuli_s | so, in df db we have a correct gateway ip - "gateway_ip": "10.100.100.1" | 12:12 |
yuli_s | in openflow rules we have: | 12:14 |
yuli_s | cookie=0x0, duration=6411.445s, table=15, n_packets=0, n_bytes=0, priority=100,arp,arp_spa=10.100.0.1,arp_tpa=10.100.201.200,arp_op=2 actions=CONTROLLER:65535 | 12:14 |
yuli_s | cookie=0x0, duration=6411.445s, table=15, n_packets=0, n_bytes=0, priority=100,arp,arp_spa=10.100.0.1,arp_tpa=10.100.0.1,arp_op=1 actions=CONTROLLER:65535 | 12:14 |
yuli_s | so, you need to alter these rules on gateway_ip change event | 12:14 |
yuli_s | here is a command I used to change gateway_ip: | 12:16 |
yuli_s | neutron subnet-update 17c869eb-909c-464b-9572-e555cb044230 --gateway_ip 10.100.100.8 | 12:16 |
*** gampel1 has joined #openstack-dragonflow | 12:21 | |
yuli_s | I opend a bug for this | 12:23 |
yuli_s | https://bugs.launchpad.net/dragonflow/+bug/1563835 | 12:23 |
openstack | Launchpad bug 1563835 in DragonFlow "gateway_ip change event in DNAT" [High,New] | 12:23 |
yuli_s | i found another problem: | 12:25 |
yuli_s | df-db dump | grep 10.100.0.1 | 12:25 |
yuli_s | Key = 40348c8c-aae3-4831-a292-80472cfa5439 , Value = {"router_id": "e6c8cdd4-1bdc-4947-ae51-e9f81df26a45", "status": "ACTIVE", "name": "40348c8c-aae3-4831-a292-80472cfa5439", "floating_port_id": "871352b4-d1da-48b9-92b5-d141c062e877", "floating_mac_address": "fa:16:3e:42:4a:8e", "floating_network_id": "cb3f53f3-2694-4924-8b31-a7735906610c", "topic": "85104976665e47cfaefe8ad50e7de019", "fixed_ip_address": "10.0.0.3", "floa | 12:25 |
yuli_s | ting_ip_address": "10.100.201.200", "external_cidr": "10.100.0.0/16", "port_id": "91693805-d02d-4e0d-8091-1f4918cab08e", "external_gateway_ip": "10.100.0.1"} | 12:25 |
raofei | Hi yuli, current dnat doesn't process the subnet update event and router update event. it will be supported later. | 12:28 |
raofei | so please test dnat without change external subnet gateway. | 12:29 |
yuli_s | ok, great ! | 12:30 |
yuli_s | it is very strange | 12:30 |
raofei | dnat APP doesn't process these two updation event. | 12:30 |
yuli_s | my recommendations might not be fully correct. | 12:30 |
yuli_s | this is is taken from floatingip table : "external_gateway_ip": "10.100.0.1" | 12:31 |
yuli_s | this IP is taken from floatingip table : "external_gateway_ip": "10.100.0.1" | 12:31 |
yuli_s | i have restarted DF controller, | 12:31 |
yuli_s | the old value is still in use ("10.100.0.1") and not ("10.100.100.1" | 12:32 |
raofei | I think when you create external subnet first time, the gateway ip is 10.100.0.1 | 12:32 |
yuli_s | yes ! | 12:32 |
yuli_s | this where from this IP is originating | 12:33 |
*** openstackgerrit has quit IRC | 12:33 | |
raofei | so when you update this value, it doesn't change | 12:33 |
*** openstackgerrit has joined #openstack-dragonflow | 12:33 | |
yuli_s | yes, same for floatingip table "external_gateway_ip" field | 12:34 |
yuli_s | strange | 12:34 |
yuli_s | ok, found it | 12:38 |
raofei | it's not strange. | 12:38 |
yuli_s | @lock_db.wrap_db_lock() | 12:38 |
yuli_s | def create_floatingip(self, context, floatingip): | 12:38 |
yuli_s | ......... | 12:38 |
yuli_s | self.nb_api.create_floatingip( | 12:38 |
yuli_s | .............. | 12:38 |
yuli_s | external_gateway_ip=floatingip_subnet['gateway_ip'], | 12:38 |
raofei | actually, I'm going to implement updation on a new patch. | 12:39 |
yuli_s | great ! | 12:39 |
raofei | so please check the dnat without updation gateway ip | 12:39 |
yuli_s | yup | 12:40 |
yuli_s | thank you for the patience ;) | 12:41 |
raofei | u are welcome | 12:43 |
yuli_s | duo xie | 12:43 |
yuli_s | ;) | 12:43 |
raofei | @Gal, current DNAT App doesn't support configuring static gateway mac. Today's issue it's because DNAT doesn't support gateway ip updation. | 12:47 |
gampel1 | I want to merge the Redis patch any objection ? | 12:55 |
gampel1 | not the dev stack the driver | 12:55 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Add publisher's table https://review.openstack.org/297167 | 12:56 |
raofei | anybody occur this problem? | 12:58 |
raofei | https://www.irccloud.com/pastebin/WAQVrPfb/ | 12:58 |
raofei | with latest devstack | 12:59 |
raofei | I know the issue, it's because ovs kernal is not load | 12:59 |
yuli_s | btw | 13:01 |
yuli_s | neutron floatingip-delete id is not working | 13:01 |
gampel1 | yuli_s: whats the problem | 13:02 |
yuli_s | it just not working | 13:02 |
yuli_s | neutron floatingip-delete 871352b4-d1da-48b9-92b5-d141c062e877 | 13:02 |
yuli_s | Floating IP 871352b4-d1da-48b9-92b5-d141c062e877 could not be found | 13:02 |
yuli_s | Neutron server returns request_ids: ['req-38f0fcd0-a6d0-4434-8c91-8e47b3f9080b'] | 13:02 |
yuli_s | neutron floatingip-list | 13:02 |
yuli_s | +--------------------------------------+------------------+---------------------+---------+ | 13:02 |
yuli_s | | id | fixed_ip_address | floating_ip_address | port_id | | 13:02 |
yuli_s | +--------------------------------------+------------------+---------------------+---------+ | 13:02 |
yuli_s | | 40348c8c-aae3-4831-a292-80472cfa5439 | | 10.100.201.200 | | | 13:02 |
yuli_s | +--------------------------------------+------------------+---------------------+---------+ | 13:02 |
yuli_s | ops | 13:03 |
yuli_s | wrong id | 13:03 |
yuli_s | i used wrong id, it is ok | 13:03 |
gampel1 | you used the wrong id and it deleted another fip ? | 13:04 |
yuli_s | nop, i had one fid | 13:05 |
yuli_s | fip | 13:05 |
gampel1 | yuli can you please describe the problem I am not sure i understand | 13:06 |
yuli_s | <raofei> @Gal, current DNAT App doesn't support configuring static gateway mac. Today's issue it's because DNAT doesn't support gateway ip updation. | 13:08 |
yuli_s | this command is not handled correctly "neutron subnet-update 17c869eb-909c-464b-9572-e555cb044230 --gateway_ip 10.100.100.8" | 13:10 |
yuli_s | i was researching this bug | 13:11 |
yuli_s | found that df db floatingip table | 13:11 |
yuli_s | has the "external_gateway_ip": filed that is not updated | 13:12 |
yuli_s | so, I was trying to continue with the tests, so, | 13:12 |
yuli_s | i removed the flowtingip | 13:12 |
yuli_s | and got a new one | 13:12 |
yuli_s | with correct gateway now | 13:13 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Install OVS as an OS service https://review.openstack.org/296897 | 13:17 |
*** openstackgerrit has quit IRC | 13:18 | |
*** openstackgerrit has joined #openstack-dragonflow | 13:19 | |
openstackgerrit | duankebo proposed openstack/dragonflow: Fullstack test cases for topology Fix the problem caused by overlaped ip address Change-Id: Ice5993aa3700df98fbff421d4583a6ed963c9aaf https://review.openstack.org/296522 | 13:24 |
gampel1 | I want to merge the Redis patch any objection ? | 13:26 |
openstackgerrit | Merged openstack/dragonflow: Fix openflow dump exception https://review.openstack.org/298039 | 13:27 |
gampel1 | nick-ma: ping | 13:27 |
DuanKebo | I agree | 13:30 |
nick-ma | gampel1: pong | 13:40 |
*** gongysh has quit IRC | 13:42 | |
*** gongysh has joined #openstack-dragonflow | 13:54 | |
openstackgerrit | Merged openstack/dragonflow: Fix typos in Dragonflow files https://review.openstack.org/299167 | 13:56 |
nick-ma | gampel1: what's up? | 13:58 |
openstackgerrit | Merged openstack/dragonflow: Solve the problem that sometimes network can't be recycled due to some ports of that network are used by some vms https://review.openstack.org/297940 | 14:01 |
gampel1 | I want to merge the devstack patch | 14:10 |
gampel1 | You were active on that patch and i want to see that it is Ok with you | 14:11 |
gampel1 | nick-ma:I want to merge the devstack patch | 14:11 |
gampel1 | I want to merge the redis driver patch not the devstack | 14:11 |
nick-ma | ok. i tried to set up redis for testing. but due to network restriction, my dev machine cannot access ruby sites. so, i cannot run it properly and i don't have time to manually build the whole environment. | 14:15 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Do not log error on missing table during pub/sub test https://review.openstack.org/297198 | 14:16 |
nick-ma | gampel1: you can merge it. i'll find out how to run devstack for redis. maybe i need to buy a stable vpn connection to access ruby sites. it's really annoying. | 14:19 |
gampel1 | nick-ma: if we could help let us know maybe we could clone it in our offices | 14:21 |
nick-ma | gampel1: np. thanks. actually i don't think relying on ruby scripts to build redis cluster is a good idea. it installs ruby runtime and lots of packages in my machine that generally i don't need them at all. but it is the issue about devstack. | 14:25 |
gampel1 | ok let us know if we could help | 14:27 |
yuli_s | gampel1, here https://review.openstack.org/#/c/298131/ | 14:27 |
gampel1 | omer: i did not know we merged dragonflow.tests.fullstack.test_apps.TestL3Ap but it is failing | 14:27 |
yuli_s | u advised to add another patch as Omer suggested | 14:28 |
openstackgerrit | Merged openstack/dragonflow: Add Redis DB Driver and Redis PubSub Driver https://review.openstack.org/286028 | 14:28 |
yuli_s | and gave -1 to this patch | 14:28 |
gampel1 | yuli_s: why you do not fix the comments in that patch ? | 14:28 |
yuli_s | Omer reported some issue that can be found in other tests | 14:29 |
nick-ma | gampel1: the L3App is failing? I checked it before, but it works. | 14:29 |
yuli_s | so, I wanted to post it as a different patch | 14:29 |
gampel1 | yuli_s: yes why not in the smae patch as nick-ma asked you ? | 14:29 |
gampel1 | nick-ma: i see that it is failing in http://logs.openstack.org/67/297167/4/check/gate-dragonflow-dsvm-fullstack-nv/7c67e69/testr_results.html.gz | 14:30 |
yuli_s | hm, I can make a patch here, just this problem exists in other tests located in other files | 14:31 |
nick-ma | gampel1: http://logs.openstack.org/16/297316/3/check/gate-dragonflow-dsvm-fullstack-nv/dfbdb43/testr_results.html.gz | 14:31 |
nick-ma | gampel1: ok, it seems that test is not stable. | 14:32 |
*** oanson has quit IRC | 14:32 | |
gampel1 | yuli_s: no omer asked you to replace .wait_until_is_and_return with .wait_until_tru ... | 14:32 |
yuli_s | yes, exactly, this problem exists in other tests too | 14:32 |
gampel1 | yuli_s: I am not sure I understand you are submitting a change and you got comment about your change , | 14:35 |
yuli_s | i will check this again, | 14:35 |
yuli_s | and add a patch | 14:36 |
gampel1 | thx | 14:36 |
yuli_s | np | 14:36 |
*** yuli_s has quit IRC | 14:41 | |
gampel1 | omer: ping | 14:48 |
openstackgerrit | Eran Gampel proposed openstack/dragonflow: Remove test that are based on flow diff https://review.openstack.org/299452 | 14:49 |
*** irenab has quit IRC | 16:04 | |
*** irenab has joined #openstack-dragonflow | 16:06 | |
openstackgerrit | Fei Rao proposed openstack/dragonflow: Dragonflow L3 Agent for distributed DNAT https://review.openstack.org/294374 | 16:11 |
*** openstack has joined #openstack-dragonflow | 17:07 | |
*** yamamoto has quit IRC | 18:53 | |
*** yamamoto has joined #openstack-dragonflow | 19:53 | |
*** yamamoto has quit IRC | 20:03 | |
*** openstackgerrit has quit IRC | 20:48 | |
*** openstackgerrit has joined #openstack-dragonflow | 20:48 | |
*** gampel1 has quit IRC | 22:04 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!