| *** mhen_ is now known as mhen | 02:42 | |
| genekuo | <DTEIT_> "hi guys, I have a question about..." <- Have you try upgrading your openstacksdk installation? I believe the latest version is 1.0.1 | 09:46 |
|---|---|---|
| DTEIT_ | genekuo: on F37 that's the latest avail | 09:52 |
| genekuo | <DTEIT_> "Gene Kuo: on F37 that's the..." <- use pip to install the latest one | 09:59 |
| DTEIT_ | genekuo: apart from that, is that a dependency issue from ansible or from openstack? | 10:01 |
| genekuo | DTEIT_: IIUC, Ansible modules requires openstacksdk >=1.0 | 10:03 |
| DTEIT_ | genekuo: ok thanks | 10:04 |
| jrosser | genekuo: there is information here at "Branches and Non Backward Compatibility" https://galaxy.ansible.com/openstack/cloud | 10:08 |
| jrosser | DTEIT_: ^ also | 10:08 |
| DTEIT_ | jrosser: ahhh nice, that's clear, thanks | 10:09 |
| edebeste | Hi everyone, I'm sitting with a bit of a network issue that I'm struggling to debug. I have a basic 3-controller cloud deployed with kolla-ansible. I'm using provider networks and *not* dvr. Physnet1 which is a trunked port and provider networks are created on top of it within OpenStack, tagging the appropriate vlan ID when creating. I've created a | 14:54 |
| edebeste | network for my internet traffic, say vlan 240, called "Internet" and I've created an internal VXLAN network called "INTERNAL". I've created a router between the two networks as well. Attaching "INTERNAL" to the VM works fine, I can reach the machines on vlan 240 and I can reach the internet. Floating IPs also work. However, when I attach both | 14:54 |
| edebeste | "INTERNAL" as well as "Internet" networks ("Internet" being a direct provider port) then the "INTERNAL" network works, but no traffic seems to come or go from the "Internet" network port in the VM. Creating the VM with "INTERNAL" only and then adding "Internet" as a port later DOES seem to work. I'm scratching my head here. Any ideas? | 14:54 |
| edebeste | If I create the VM with both networks, remove the "Internet" network and add it back in then things work again. Comparing ovs-vsctl show and list port shows everything looks almost exactly the same except the "tag" value iterates | 15:06 |
| edebeste | so the issue seems to crop up only when the VM is created with both networks attached, but not when adding the network afterwards | 15:11 |
| frickler | vms act weird when attached to two networks that both advertise a default route. best solution is to avoid such a scenario | 15:18 |
| edebeste | frickler, to add more info, the second network does not have dhcp. It must be set statically within the VM. There are no 2 default routes | 15:25 |
| edebeste | I also just discovered that it seems to not necessarily be network related. I just created a VM with the networks the other way around. I assigned the static IP to eth0 "(which is now the provider net on vlan 240). I then did a DHCP request on the second network (internal vxlan) and no traffic is flowing to or from the second network | 15:26 |
| edebeste | It appears that I am unable to attach two networks to my vms and have the second network work at all until I detach it and re-attach it | 15:26 |
| jamesben_ | Are there any ceph folks here? I have a bug but issues filing | 18:51 |
| *** EugenMayer40 is now known as EugenMayer4 | 19:38 | |
| *** damian___921 is now known as damian___92 | 19:39 | |
| *** arif-ali_ is now known as arif-ali | 19:39 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!