#theforeman

how does errata imports work? i can see errata in pulp "pulp-admin rpm repo content errata --repo-id="centos7-updates-x86_64""

however none of them are visible in the wui or otherwise ..

well errata for centos isnt officially supported, I assume you've used https://github.com/brdude/pulp_centos_errata_im..., and the 5 steps mentioned in the readme should be enough to see said errata under content

carl-, if the data in pulp is there but not in the UI try "foreman-rake katello:reimport"

sean_: will try that

hey forman, is there a way to modify a search after its created?

the out of sync one is set to be 35 minutes but we can only roll through all our machines once an hour

squig: check the settings page, look for puppetrun interval

yes taht gets the one from the profile but no the one in all hosts saved search out of sync

ah, thats just a bookmark, which is set by default. you can delete it and save a new one

how can I delete it :)

can I jsut write over the top of it?

Administer > Bookmarks

you can edit the existing one or delete it and save a new one

thank you, just spent 20 minutes looking for that menu :D

:)

sean_: that took care of it it seems .. now .. so katello has to actually get to know whats in pulp it is not "directly" integrated but imports from pulp

??

when you "hammer host delete" is an delete from disk issued or does it not remove the machine all together ??

In puppet I can use $facts['hostname'] to get the hostname, but how do I get the same variable from foreman, so I can change the hostname from foreman?

in vmware to clarify

carl-: if the host in foreman is associated with your VMWare compute resource, 'hammer host delete' should delete the host in VMWare too

$::foreman_interfaces. is such a variable.

Han: in foreman you can use ERB to access anything Foreman knows about, so <%= @host.fqdn %> is probably what you want - you can use that in a template or in a class parameter

dLobatog: hmmm .. well in vmware you can remove a host from inventory or do delete from disk

dLobatog: o there are 2 methods of actually removing a vm .. where the other will leave disks in place

gwmngilfen, I think I found it, if I look at the yaml output in the interface I can see all the relevant variables. So hostname would be: $::foreman_ip.name

gwmngilfen, your solution also sounds interesting.

Han: oh you mean on the puppet side? sure you can reference stuff in the enc from top scope ($::<foo>) or send it via class parameters

depends on your puppet code, of course

indeed.

carl-: as far as i know it should remove the vm entirely, that'd be consistent with the other resources we integrate with

you can "disassociate" a vm, which removes it from foreman

(thats on the host edit page, IIRC)

gwmngilfen: mm ok .. well "grep -inr destroy_vm /usr/share/foreman/"

def destroy_vm(uuid) i can see that everywhere but not in the vmware.rb

is it somehow different how it works ?

i dont actually use vmware myself, so i'm working on assumptions. let me see if i can find out

oooh thanks

so it looks like vmware is the only one that doesn't override the base destroy_vm method in app/models/compute_resource.rb (line 190)

you can see it expects such a method to exist on line 417 of vmware.rb

actually, i'm wrong, vmware.rb does define destroy_vm

looks like tat was added in 1.14, so if you're on an older version you may not see it

carl-: ^

but then for older versions my initial comment about using the base class' mathod is true :)

gwmngilfen: hmm .. i cant find it here .. it just calls it where i am :) .. hmmm ..

what version are you on?

gwmngilfen: foreman-1.11.0.65-1

oh man, old

... mmmmm

also, where are you getting that version from? iirc we only do x.y.z numbers, the 65 is odd

gwmngilfen: well its satellite actually thats why its old and .. .... old :)

ah, satellite

righto :)

anyway

hehehe :)

thats where it's defined

which really just finds the vm in Fog and then calls .destroy on it

well hmmm .. ok . i dont know if understand exaclty what it does .. or what the fog is for that matter :) however thanks alot anyway

carl-: so fog is the layer that handles compute interactions for us (github.com/fog iirc) - we just tell Fog to go delete the VM, and it handles it. As I say, I would expect a full delete if you just click Delete in the UI, but I assume that's not what you're seeing?

well someone who worked now said they could see actually that disks remained, that suggests otherwise .. he asked me how it worked and i had assumed that it did a complete delete .. however i cannot see the code actually

lemme see if i can find someone more vmware knowledgeable :P

carl-: if it all goes quiet, try the mailing list - that has a wider audience and is less timezone sensistive :)

aaah :) nice

I'm checking fog-vmware to make sure..

dLobatog++

carl-: as far as I can see it just makes a call to vSphere using the instance UUID, which refers to the inventory host only

carl-: I'm asking a guy in #foreman-dev now..