#theforeman

Hello

is anyone around?

I'm trying to figure out how to use PXE provisioning with libvirt VMs without controlling DHCP/DNS (I want to use an external router and have the VMs on the same network as the hypervisor)

I'm just a hobbyist here, so I'm looking online for tutorials that may illuminate this for me, but none that I've found talk about not using the foreman as a DNS/DHCP authority

hesperaux: whats the question? :)

ohadlevy! I remember your nick

ohadlevy, I was able to configure my pfsense box (external machine) to provide the tftp IP of my foreman server and the filename (/pxelinux.0). With that, when I try to provision a host, it boots PXELinux, but the provision script, etc. is not automatically set up. If I manually modify /srv/tftp/pxelinux.cfg/default on my server to replicate the options generated by the puppet website under Hosts->Provisioning Templates, it

successfully installs Linux

The question I have is... how might the TFTP server/DHCP server/whatever be misconfigured that would cause the PXE config not to be properly generated?

I looked in /var/lib/ and there is no tftp folder in there, and the tftp process running on the foreman VM is pointing to /srv/tftp/ instead

ALso, is it normal for there to be 0 puppet classes installed by foreman-installer? I manually installed the stdlib module into the production environment and was able to import it successfully, but I figured there would be some already installed

Hi! How much RAM should I provide to a foreman instance that'll manage 30-50 nodes?

2 GB maybe? but I guess it depends on many factors

the instance currently has 2G, but the "Passnger RackApp" process takes already a lot of that.

I just had out of memory errors doing a reverse search in bash

is it the foreman or the puppet passenger instance?

the puppetmaster is on a different node, so foreman only

that is.... interesting

ha, that's what I through

I'm on 1.11

probably best to open an issue, I'm not an expert of passenger internals ;)

ok

#13801 maybe?

mmoll: #13801 is http://theforeman.org/issues/13801 "Bug #13801: foreman passenger memory leak - Foreman"

yes, sounds fimilar mmoll

another question: how can it happen, that certain hosts show "Last execution failed" even though the execution ran fine and search for errored hosts doesn't show them?

there were/are some issues about the new combined host status, best ask again tomorrow at day time

ok

I thought only katello hosts still have that problem, but I didn't follow that

It got better with 1.11 (I think I first saw it after upgrading to 1.10.1 or .2), but a 3 hosts still have it

yeah, I saw something similar

iirc my problem also was stemming from using salt, puppet and remote_execution at the same for the same hosts

I started using puppetdb around that time

that shouldn'

no salt in use, but remote_execution since a few weeks

t influcence the host status in foreman