ManageIQ/manageiq-providers-vmware

Jason Frey

@Fryguy

:sparkles:

It's a part of the ontap stuff

But I thought that was removed ages ago

Keenan Brock

@kbrock

ikr?

thnx

never heard of it - I'll remove the view. thanks

Jason Frey

@Fryguy

Can you grep for Snia and Ontap in core just to be sure...I may be mistaken

Keenan Brock

@kbrock

a bunch of localization files

but no ruby files

Jason Frey

@Fryguy

Ah ok

Keenan Brock

@kbrock

and of course a view... which is why <kaboom>

Jason Frey

@Fryguy

Yeah localization will auto clear itself once @mzazrivec does his magic

Keenan Brock

@kbrock

nice

Jason Frey

@Fryguy

ManageIQ/manageiq#15153

Keenan Brock

@kbrock

that one wasn't so bad. Hawkular still has factories :)

ooh sorry all - thought this was providers room. not providers-vmware

Fabien Dupont

@fdupont-redhat

Hi. I'm facing a weird behavior of the inventory for a VMware VM. I have an RHEL 7 VM with 3 IP addresses configured on the same NIC: 2 x IPv4 and 1 x IPv6. I've followed https://access.redhat.com/solutions/8672 to configure the additional IPv4 address. VMware UI reports the 3 IP addresses, as well as ManageIQ. However, looking at the objects, I can't get all the IP addresses by NIC.

The GuestDevice id is 42000000000032:

irb> Vm.find_by(name: 'fdupont-test-migration').hardware.nics
=> #<ActiveRecord::Associations::CollectionProxy [#<GuestDevice id: 42000000000032, device_name: "Network adapter 1", device_type: "ethernet", location: nil, filename: nil, hardware_id: 42000000000024, mode: nil, controller_type: "ethernet", size: nil, free_space: nil, size_on_disk: nil, address: "00:50:56:a5:6b:9a", switch_id: nil, lan_id: 42000000000001, model: "VirtualVmxnet3", iscsi_name: nil, iscsi_alias: nil, present: true, start_connected: true, auto_detect: nil, uid_ems: "00:50:56:a5:6b:9a", chap_auth_enabled: nil, manufacturer: nil, field_replaceable_unit: nil, parent_device_id: nil, vlan_key: nil, vlan_enabled: nil, peer_mac_address: nil>]>

When looking at the networks, only one has the device_id set:

irb> Vm.find_by(name: 'fdupont-test-migration').hardware.networks
=> #<ActiveRecord::Associations::CollectionProxy [#<Network id: 42000000000112, hardware_id: 42000000000024, device_id: 42000000000032, description: nil, guid: nil, dhcp_enabled: nil, ipaddress: "10.19.2.32", subnet_mask: nil, lease_obtained: nil, lease_expires: nil, default_gateway: nil, dhcp_server: nil, dns_server: nil, hostname: "fdupont-test.example.com", domain: nil, ipv6address: nil>, #<Network id: 42000000000113, hardware_id: 42000000000024, device_id: nil, description: nil, guid: nil, dhcp_enabled: nil, ipaddress: "10.19.2.31", subnet_mask: nil, lease_obtained: nil, lease_expires: nil, default_gateway: nil, dhcp_server: nil, dns_server: nil, hostname: "fdupont-test.example.com", domain: nil, ipv6address: "fe80::250:56ff:fea5:6b9a">]>

Shouldn't all the network objects have a device_id ? If no, how could I collect all the IP addresses per NIC ?

Miha Pleško

@miha-plesko

for cloud providers it would be

all_ips = nic.cloud_subnet_network_ports.map(&:address)

but you're into infra provider so not sure

Fabien Dupont

@fdupont-redhat

Thanks @miha-plesko, but indeed it's not the same for infra provider. I would have expected each NIC to have to IP addresses attached to it, not to the network...

Adam Grare

@agrare

@fdupont-redhat that's interesting, let me look into this

I agree all of the networks should have an associated device

Fabien Dupont

@fdupont-redhat

Does it seem easy to fix in Hammer ?

Adam Grare

@agrare

I have to find out what's wrong first ;)

Fabien Dupont

@fdupont-redhat

Sure... :wink:

Adam Grare

@agrare

so we model it as a nice has_one network

so if there are more than one network with the same nic it overwrites it

which is why you're seeing only one with a device_id

Fabien Dupont

@fdupont-redhat

Oops. Means a schema change, doesn't it ?

Adam Grare

@agrare

so the device_id is on the network, which allows for has_many

we'd have to change the association though from a has_one :network to has_many :networks and idk what else that might break

we'd have to update the refresher also to handle this

Fabien Dupont

@fdupont-redhat

Not a small change then... Any chance to have it in 5.10.0 ? Or rather 5.10.1 ?

Adam Grare

@agrare

what's 5.10? :wink: you mean hammer-1 / hammer-2?

hammer is rc1, I doubt we'd be able to do this for hammer- without this being a severe blocker

Fabien Dupont

@fdupont-redhat

Yes hammer-1 / hammer-2...

Jared Deubel

@jdeubel

Anyone here seen an issue where we get Timed out stalled task ?

[----] I, [2019-03-21T22:21:56.010309 #6433:4cb1ea8]  INFO -- : MIQ(MiqTask.generic_action_with_callback) Task: [1000000000128] Queued the action: [Validate EMS Provider Credentials] being run for user: [admin]
[----] I, [2019-03-21T22:22:26.012863 #6433:4cb1ea8]  INFO -- : MIQ(MiqTask#update_status) Task: [1000000000128] [Finished] [Timeout] [Timed out stalled task.]

Adam Grare

@agrare

@jdeubel do you have any workers running in that zone?

Jared Deubel

@jdeubel

@agrare looking...

@agrare which worker does the credentials?

Adam Grare

@agrare

generic

Jared Deubel

@jdeubel

ems_operations?

Adam Grare

@agrare

with the ems_operations role

Jared Deubel

@jdeubel

@agrare The role is enabled on the appliances

Adam Grare

@agrare

and the generic workers are running/

Jared Deubel

@jdeubel

It looks like it yea.

Im getting more logs to see what is going on

Adam Grare

@agrare

if they are running, and they aren't so backed up that they aren't getting to that queue item in 30 mins, then there's some other reason they aren't dequeueing messages

Where communities thrive

People

Repo info

Activity