#theforeman

ricou: I might give it a try if that helps

timogoebel++

mbacovsk_: created a redmine for this #18304

mbacovsk_: #18304 is http://theforeman.org/issues/18304 "Feature #18304: Support updating network interfaces for VMWare - Foreman"

timogoebel: great, thanks!

weird, @host.sp_ip always return true even if the host does not have any BMC interface attached , at least on VMware VM

ricou++

what could be the cause, that a host keeps pulling in a wrong puppet env, altough it seems no longer to be assigned?

Error: Notice: Local environment: 'KT_SCM_Test_CentOS_7_3_PuppetSandbox_9' doesn't match server specified node environment 'KT_SCM_Test_CV_CentOS72_3', switching agent to 'KT_SCM_Test_CV_CentOS72_3'.

agent run: https://paste.fedoraproject.org/541130/48578881/

I have no idea why the agent wants to switch the environment to 'KT_SCM_Test_CV_CentOS72_3'

the host should use 'KT_SCM_Test_CentOS_7_3_PuppetSandbox_9'

yaml says and assignments say so :/

lzap would it be possible for you to add mdadm to the discovery image?

lzap that way I could add support for mdraid

Raboo: no problem send me exact package names

lzap ok, will check centos repos

looks like mdadm it is

lzap yes, it's 'mdadm'

ok added

will be part of my image deployment patch, still working on that

lzap perhaps it's good to have xfs utils if not included and other filesystem utils needed to create filesystems and partitions

xfsprogs right?

lzap i'm building a option to run a custom partition script with the image installer

sure

sorry, yes xfsprogs

done

I also added "ssm" which is also quite capable thing

that way folks can use partition tables in foreman to create custom partition setups

Raboo: just curious, did you test out the cloud-init plugin?

timogoebel not yet, i was not able to get around the whole ruby version thing. I have a plan!

my plan is to updated bundler.d/Gemfile

and then run a foreman update which will trigger a bundle install with the correct ruby version

timogoebel currently i'm adding support for custom partition setups. I noticed that it was needed when i tried to setup two older HP nodes with no raid controller.

Raboo: cool

mmoll: what are the steps required to add an additional plugin from git to a .deb foreman install? is that easily doable?

lzap if the image grows to large, perhaps it's time to look at alpine linux? :)

Raboo: nope if a hardware is certified for RHEL, discovery will run on it!

lzap another idea about the discovery image is to add an option for remote syslog

yeah multiple users asked for that already

its slightly bigger now by 25 MBs :-(

I don't want to strip out any drivers or firmware...

lzap for me it doesn't matter since my pxe fetches via http, but i pity those who use TFTP

yeah I was prototyping using NBD remote diskless stuff but it was not great

basically the only thing i fetch via tftp is pxelinux.cfg/ boot configs and fdi_image extentions

lzap i think it's better to document that pxelinux compiled with http support is recommended

ipxelinux.0

Raboo: oh really? does it work?

Raboo: oh you said iPXE not PXE I see

feel free our discovery docs are at foreman

hello. i have some problem with image based provisionnon with Vmware provider. could someone help me please ? Network information (ip,gateway,hostname) are ok. Now i want to execute some custom action like spacewalk resgistration and puppet install

the run-cmd from my user_data seems to not be executer

executer

*executed

I'm trying to provision a centos 7 machine, but dracut-initqueue fetches something big and then fails because the ramdisk is full - what is it trying to fetch?

ewoud: update.img ?

lzap basically ipxelinux is pxelinux with http support

Raboo: yeah I know it very well :-) I just misread it

does i need cloud-init in my template in order to execute the run-cmd or do i have to go with the snippet remote-execution-ssh?

ewoud: I think that's product.img and update.img

lzap: why are they so big and how can I fix this?

ricou i'm working on a image based installer for foreman

ricou that uses cloud-init

ewoud: if this is 7.3 note memory requirements are higher now

lzap: ah, it increased?

so it's not yet implemented on foreman ?

ewoud: discussed on the foreman thread, here is also relevant link https://ekohl.nl/Screenshot_rhel7.2_2017-01-30_...

uh

ricou I don't know if there are other ways to install cloud-init images, but i built my own way at least.

ricou https://github.com/deltaprojects/foreman_discov...

ewoud: its actually install.img I thinmk

ricou but I would advise to wait a day or two, I'm going to push and update with a major change soon

ewoud: not update.img thats something other

and run-cmd will work?

ewoud: you better come to my talk about bare-metal image deployment at CfgMgmt :-)

its now more relevant with the new hungry Anaconda

lzap: I'm trying to prepare a talk for the centos dojo which includes foreman discovery so your talk won't be in time ;)

haha yeah

its bummer this anaconda memory bumps

ricou runcmd in cloud-init, yes i use that, look ath the readme with the example

luckily I'm only doing an introduction talk so I'll refer to your talk for those attending cfgmgtcamp

what permission should /etc/puppetlabs/puppet/ssl/client_key.pem have?

ricou https://github.com/deltaprojects/foreman_discov...

I currently see a lot of errors in puppetserver logs with permission denied :/

lzap: so yes, with 1.5GB it does continue

ok it looks fine

if you want i test something i'm your man

Raboo : if you want i test something i'm your man

Raboo: ewoud: https://www.youtube.com/watch?v=63EkGNGWH3A&...

lzap is it your udpcast?

Raboo: yup but I had a good comment on that - udp might be restricted so I am integrating HTTP transport this week

ricou testing is always welcome. I've rewritten big parts of the script and will push it once i've tested it.