*** yamamoto has joined #openstack-dragonflow | 00:53 | |
*** yamamoto has quit IRC | 00:59 | |
*** yamamoto has joined #openstack-dragonflow | 01:05 | |
*** yamamoto has quit IRC | 01:13 | |
*** gongysh has joined #openstack-dragonflow | 01:29 | |
*** zenoway has quit IRC | 01:50 | |
*** gongysh has quit IRC | 02:05 | |
*** yamamoto has joined #openstack-dragonflow | 02:11 | |
*** yamamoto has quit IRC | 02:17 | |
*** zenoway has joined #openstack-dragonflow | 03:06 | |
*** zenoway has quit IRC | 03:10 | |
*** yamamoto has joined #openstack-dragonflow | 03:17 | |
*** yamamoto has quit IRC | 03:22 | |
*** zenoway has joined #openstack-dragonflow | 03:30 | |
*** zenoway has quit IRC | 03:35 | |
*** yamamoto has joined #openstack-dragonflow | 03:49 | |
*** yamamoto has quit IRC | 03:54 | |
*** zenoway has joined #openstack-dragonflow | 03:55 | |
*** zenoway has quit IRC | 04:01 | |
*** hshan has joined #openstack-dragonflow | 04:46 | |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Remove redundant call to _register_as_publisher from NB API. https://review.openstack.org/303798 | 04:55 |
---|---|---|
hshan | oanson: ping | 05:03 |
*** hshan has quit IRC | 05:08 | |
*** yamamoto has joined #openstack-dragonflow | 05:17 | |
*** yamamoto has quit IRC | 05:22 | |
*** oanson has joined #openstack-dragonflow | 05:31 | |
*** zenoway has joined #openstack-dragonflow | 05:36 | |
*** zenoway has quit IRC | 05:40 | |
openstackgerrit | Merged openstack/dragonflow: Fix AttributeError for calling _register_as_publisher https://review.openstack.org/303325 | 05:44 |
openstackgerrit | Omer Anson proposed openstack/dragonflow: Lookup router_port's network_id with lswitch's topic https://review.openstack.org/298121 | 06:12 |
*** zenoway has joined #openstack-dragonflow | 06:14 | |
*** yamamoto has joined #openstack-dragonflow | 06:18 | |
*** zenoway has quit IRC | 06:18 | |
*** yamamoto has quit IRC | 06:23 | |
*** irenab_ has joined #openstack-dragonflow | 06:32 | |
*** irenab has quit IRC | 06:34 | |
*** irenab_ is now known as irenab | 06:34 | |
*** yamamoto has joined #openstack-dragonflow | 06:35 | |
*** zenoway has joined #openstack-dragonflow | 06:38 | |
*** zenoway has quit IRC | 06:42 | |
*** yamamoto has quit IRC | 06:43 | |
*** zenoway has joined #openstack-dragonflow | 06:52 | |
*** zenoway has quit IRC | 06:56 | |
*** zenoway has joined #openstack-dragonflow | 07:16 | |
*** zenoway has quit IRC | 07:21 | |
*** zenoway has joined #openstack-dragonflow | 07:54 | |
*** gampel has joined #openstack-dragonflow | 07:55 | |
*** gampel has quit IRC | 08:00 | |
*** zenoway has quit IRC | 08:01 | |
*** gampel has joined #openstack-dragonflow | 08:04 | |
*** zenoway has joined #openstack-dragonflow | 08:13 | |
*** zenoway has quit IRC | 08:17 | |
*** Shlomo_N has quit IRC | 08:34 | |
*** yuli_s has joined #openstack-dragonflow | 08:35 | |
yuli_s | raofei you here ? | 08:35 |
*** zenoway has joined #openstack-dragonflow | 08:37 | |
*** zenoway has quit IRC | 08:41 | |
openstackgerrit | Li Ma proposed openstack/dragonflow: Fix port update event loss https://review.openstack.org/303284 | 09:06 |
nick-ma | Hi all, I have a bug fix that I think it is of high priority. I did a code analysis in the bug description: https://bugs.launchpad.net/dragonflow/+bug/1567773/comments/7 and also a fix is under review: https://review.openstack.org/#/c/303284/ | 09:11 |
openstack | Launchpad bug 1567773 in DragonFlow "DHCP failed for instances" [High,In progress] - Assigned to Li Ma (nick-ma-z) | 09:11 |
nick-ma | If it is wrong, please correct me. thanks a lot. | 09:11 |
nick-ma | by the way, the secgroup patch doesn't work for me. all the flows are correctly installed but they are not working at all. it takes time for me to go over each flow one by one. if someone made it, please tell me about it. maybe there is some fault in my environment. | 09:14 |
*** yamamoto has joined #openstack-dragonflow | 09:17 | |
*** gongysh has joined #openstack-dragonflow | 09:25 | |
*** yamamoto has quit IRC | 09:25 | |
yuli_s | nick-ma, today I have a kind of similar bug (probably) | 09:39 |
yuli_s | after installations no default rules existed in table 0 | 09:40 |
yuli_s | except for in_port=1 | 09:40 |
yuli_s | i had to recreate the router interface of the private network on router1 | 09:40 |
yuli_s | i deleted it and created again, after that table:0 were filled with required rules | 09:41 |
*** CrayZee has joined #openstack-dragonflow | 09:51 | |
*** oanson has quit IRC | 10:01 | |
gampel | nick-ma: ping | 10:14 |
nick-ma | gampel: pong | 10:14 |
gampel | Hi Raly is failing now often on subnet add delete test I think it is related to your change that we merged last week | 10:15 |
gampel | will you have have time to look at this next week | 10:15 |
gampel | http://logs.openstack.org/38/280538/17/check/gate-dragonflow-dsvm-rally-nv/9db69ea/rally-plot/results.html.gz | 10:17 |
nick-ma | do you have logs? I just checked the ongoing reviews. for example, https://review.openstack.org/#/c/303397/, it is working. | 10:17 |
nick-ma | ok. | 10:17 |
gampel | It does not fail all the time but every second patch more or less on Request Failed: internal server error while processing your request. Neutron server returns request_ids: ['req-aa39596f-ec48-41e8-ae50-767c3a054d31'] | 10:18 |
gampel | Thanks | 10:19 |
nick-ma | i will fix. a tiny bug. sorry. | 10:20 |
gampel | InternalError(errno, errorvalue)\n', "DBDeadlock: (pymysql.err.InternalError) (1213, u'Deadlock found when trying to get lock; try restarting transaction') [SQL: u'UPDATE ipavailabilityranges SET first_ip=%(first_ip)s WHERE ipavailabilityranges.allocation_pool_id = %(ipavailabilityranges_allocation_pool_id)s AND ipavailabilityranges.first_ip = %(ipavailabilityranges_first_ip)s AND ipavailabilityranges.last_ip = %(ipavailabili | 10:20 |
gampel | Ok thx | 10:20 |
nick-ma | the database exception was discovered before, it seems not related with dragonflow. | 10:23 |
nick-ma | deadlock in ip allocation. | 10:23 |
openstackgerrit | Li Ma proposed openstack/dragonflow: Fix UnboundLocalError in create_subnet and update_subnet https://review.openstack.org/303819 | 10:24 |
openstackgerrit | Merged openstack/dragonflow: Stale port deletion when transaction failed https://review.openstack.org/303397 | 10:39 |
*** CrayZee has quit IRC | 10:59 | |
*** CrayZee has joined #openstack-dragonflow | 11:00 | |
openstackgerrit | Merged openstack/dragonflow: Fix UnboundLocalError in create_subnet and update_subnet https://review.openstack.org/303819 | 11:30 |
yuli_s | https://bugs.launchpad.net/dragonflow/+bug/1568498 | 11:54 |
openstack | Launchpad bug 1568498 in DragonFlow "sg test bug 1" [High,New] | 11:54 |
openstackgerrit | Eran Gampel proposed openstack/dragonflow: Mitaka Version Features Update https://review.openstack.org/303825 | 12:06 |
*** gongysh has quit IRC | 12:07 | |
openstackgerrit | Eran Gampel proposed openstack/dragonflow: Mitaka Version Features Update https://review.openstack.org/303825 | 12:09 |
todin | hmm. I think the local.conf in the vagrant environment is wrong because of the dist. DNAT changes. The concat from the the setup-controller.sh ends in the post-config stanza, but it should be in the local stanza | 12:32 |
gsagie | todin: its very likely that its the case, we havent updated it recently, but feel free to send a patch with a correction if you have something | 12:34 |
todin | gsagie: ok, I will try | 12:36 |
gsagie | todin: thanks | 12:36 |
todin | nick-ma: you are right, when the external router is reachable all the flows are there. | 12:36 |
gsagie | todin: btw, keep in mind we are in heavy integration point right now, so the master version is not that stable, we hope to stabilize it by the summit | 12:37 |
gsagie | we are getting into feature freeze | 12:37 |
todin | gsagie: no problem, I am wondering if I use df with dist. DNAT every compute node should have a br-ex and the dhcp-app in the local df-controlller? | 12:39 |
gsagie | todin: br-ex yes, the dhcp app is not needed | 12:40 |
gsagie | there is a special local.conf needed i think, yuli_s do you have an example for todin? | 12:40 |
gsagie | todin: actually the DHCP app needs to run in any compute node regardless, so yes but its not related to the distributed DNAT | 12:41 |
yuli_s | todin, give me your email | 12:41 |
yuli_s | i will mail you my local.config | 12:41 |
todin | yuli_s: see query | 12:42 |
todin | and the dhcp app is getting the information via the e.g. etcd? | 12:43 |
todin | and the neutron ml2 driver is putting the data into etcd? | 12:43 |
yuli_s | todin, send | 12:43 |
yuli_s | the sg code is buggy at this point | 12:43 |
todin | yuli_s: thanks | 12:43 |
gsagie | todin : the DHCP app like everything else in the local controller can be distributed via the DB or via our pluggable pub-sub mechanism depends on your configuration | 12:44 |
gsagie | what we do, instead of having DHCP namespace with dnsmasq is have an application in the local controller that answer DHCP offers/acks | 12:45 |
todin | gsagie: yep I saw that in the openflows | 12:45 |
todin | what is the diffrents between the DB an dhte pluggable pub-sub mech? | 12:46 |
gsagie | btw, you can also work with q-dhcp if you prefer it | 12:46 |
todin | gsagie: no, I prefer the way you did that, I like it | 12:46 |
gsagie | todin: we identified that sometimes you want different pub-sub mechanism then the one the DB offers, so the controller when it comes up first sync with the DB then register for specific topics on the pub-sub mechanism and our publishers (sitting in the Neutron servers) send the events | 12:47 |
gsagie | so for example you could use Redis + ZeroMQ as the pub-sub mechanism to get notifications on DB changes (instead of using Redis build in pub-sub) | 12:47 |
gsagie | we identified that sometime there are different requierments from these two (db and pub-sub) so we abstracted them in the design | 12:48 |
gsagie | From our testing we have very good results using ZeroMQ, we will publish them in the summit | 12:48 |
gsagie | Are you going to attend the Austin summit todin? | 12:49 |
todin | gsagie: unfortunatly not, but collegues of mine | 12:50 |
gsagie | todin: tell them to come to our design sessions, we can talk furthur about our future plans | 12:52 |
todin | gsagie: we are looking more for professional services to get us startet with our openstack plattform | 12:54 |
todin | could you offer such services? | 12:58 |
gsagie | what do you mean "professional services" ? | 12:59 |
gsagie | you have a cloud/data center and you want to deploy OpenStack? | 13:01 |
todin | I mean that your engineers are helping us to implement df in our environment. | 13:01 |
todin | gsagie: yes, we are a hosting company and want to deploy openstack in our datacenter. but right now we have some scalibilty issuses | 13:02 |
gsagie | todin: which company if i may ask? | 13:03 |
*** CrayZee has quit IRC | 13:03 | |
todin | gsagie: see query | 13:04 |
*** CrayZee has joined #openstack-dragonflow | 13:04 | |
*** gongysh has joined #openstack-dragonflow | 13:52 | |
*** gongysh has quit IRC | 14:10 | |
*** gongysh has joined #openstack-dragonflow | 14:40 | |
*** yuli_s has quit IRC | 15:22 | |
*** gongysh has quit IRC | 16:09 | |
*** oanson has joined #openstack-dragonflow | 16:48 | |
*** oanson has quit IRC | 17:11 | |
*** oanson has joined #openstack-dragonflow | 17:40 | |
*** zenoway has joined #openstack-dragonflow | 20:02 | |
*** zenoway has quit IRC | 20:29 | |
*** oanson has quit IRC | 20:53 | |
*** zenoway has joined #openstack-dragonflow | 21:25 | |
*** gongysh has joined #openstack-dragonflow | 22:13 | |
*** zenoway has quit IRC | 23:06 | |
*** zenoway has joined #openstack-dragonflow | 23:21 | |
*** zenoway has quit IRC | 23:29 | |
*** zenoway has joined #openstack-dragonflow | 23:53 | |
*** zenoway has quit IRC | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!