Page MenuHomePhabricator

Unable to access codfw1-dev VMs via ssh
Closed, ResolvedPublic

Description

In theory we now have publicly-accessible public floating IPs available in codfw1-dev. I've added one to a VM and am attempting to access it from my local laptop -- I get 'connection refused.'

I've confirmed that ssh-server is running on that host (I can connect locally), and that vm has a security group of 'ALLOW IPv4 22/tcp from 0.0.0.0/0'

So... not sure if this is a networking thing or a config issue in openstack or on the VM. The fact that it's a connection refused and not a timeout suggests that ssh is actively rejecting the connection but I'm out of ideas.

Event Timeline

The VM I'm testing with is:

+--------------------------------------+----------------------------------------------------------+
| Field                                | Value                                                    |
+--------------------------------------+----------------------------------------------------------+
| OS-DCF:diskConfig                    | AUTO                                                     |
| OS-EXT-AZ:availability_zone          | nova                                                     |
| OS-EXT-SRV-ATTR:host                 | cloudvirt2001-dev                                        |
| OS-EXT-SRV-ATTR:hypervisor_hostname  | cloudvirt2001-dev.codfw.wmnet                            |
| OS-EXT-SRV-ATTR:instance_name        | i-00000461                                               |
| OS-EXT-STS:power_state               | Running                                                  |
| OS-EXT-STS:task_state                | None                                                     |
| OS-EXT-STS:vm_state                  | active                                                   |
| OS-SRV-USG:launched_at               | 2020-01-03T09:41:06.000000                               |
| OS-SRV-USG:terminated_at             | None                                                     |
| accessIPv4                           |                                                          |
| accessIPv6                           |                                                          |
| addresses                            | lan-flat-cloudinstances2b=172.16.128.32, 208.80.153.186  |
| config_drive                         |                                                          |
| created                              | 2020-01-03T09:40:50Z                                     |
| flavor                               | m1.small (2)                                             |
| hostId                               | 8fdf14da370a65315e7c4fb8496e87c196aa81afe6dce16d73bf0c99 |
| id                                   | ec7d86b2-5bb1-4755-b50c-aceecc444bfa                     |
| image                                | d1b2ea32-10ca-40a5-a3fc-babc3956f049                     |
| key_name                             | None                                                     |
| name                                 | pike-test-01                                             |
| os-extended-volumes:volumes_attached | []                                                       |
| progress                             | 0                                                        |
| project_id                           | devtest                                                  |
| properties                           |                                                          |
| security_groups                      | [{u'name': u'external_ssh'}, {u'name': u'default'}]      |
| status                               | ACTIVE                                                   |
| updated                              | 2020-01-03T09:41:06Z                                     |
| user_id                              | labtestandrew                                            |
+--------------------------------------+----------------------------------------------------------+

Mentioned in SAL (#wikimedia-cloud) [2020-01-13T13:34:38Z] <arturo> [¢odfw1dev] prevent neutron from allocating floating IPs from the wrong subnet by doing neutron subnet-update --allocation-pool start=208.80.153.190,end=208.80.153.190 cloud-instances-transport1-b-codfw (T242594)

@Andrew reports via IRC this is working now.

Issue is that neutron was allocating a floating IP from the wrong subnet. This is fixed by reducing the transport allocation pool to just the IP in use in the neutron end (the other end is the core router).

More info on this:

Please reopen if required.