2017-06-01 07:52
hi guys, i have tried to install some times, but no networks coming up or adjustable in the web interface. each time i have tried to add one, i can not configure this to fits to my needs. I need following design:

2017-06-01 07:54
!(/home/de1235092/Documents/Projekts/Talanx/rebar/system_layout_V01.png)

2017-06-01 07:54
(how can i insert a picture?)


2017-06-01 07:56
which asses mode i should use? host or forwarder?

greg
2017-06-01 12:46
HOST - note the capitalize the word on the command line. Create a network with category admin and group lab for the deploy network. set conduit to dhcp. For BMC, create a network with category bmc and group lab for the bmc network. set conduit to bmc.

greg
2017-06-01 12:47
DigitalRebar runs on an Admin node. The admin node can have any amount of networks it needs, but there are two requirements. That the admin node has routable access to the BMC network and that nodes and the admin node can have bidirectional communication on a admin network. [11:06] The Admin node will be required to be in the L2 network for all DHCP networks unless DHCP relays/BOOTP forwarders can be configured to relay on its behalf. [11:07] The DigitalRebar admin node will only serve DHCP from configured networks for networks in the bmc and admin categories (I believe). You can through the cli/api force the DHCP server to do more, but it will not necessarily be reflected on the node in DigialRebar. [11:09] DigitalRebar Networks are defined by a category and a group. The group is an arbitrary string that allows networks of different categories to be applied to nodes in the same group. [11:09] The category defines what the network is used for. This is a free form field that has two special values. [11:10] The first is admin - networks in this category are assumed to be the PXE booting networks for the bare metal system. These networks when containing a network range "dhcp" will server those addresses to unregistered nodes [11:11] The DigitalRebar will assign an IP from the host range once the node is discovered and treat that address as a more static address (though it can still be delivered by DHCP). [11:12] The other reserved category is Bmc. This is used by the IPMI roles / management system to direct IPMI configuration. It too should have a host range that will assign addresses from the pool. It can also use DHCP to deliver those bound addresses. [11:15] Other categories are for use of the administrator. These can be applied to nodes for configuration post installation (or pre, but most useful post) and will assign addresses from the host range to nodes. The group is used to figure out which sets of categories go together. For example, if DigitalRebar discovers a node on network admin-rack1 and it is told to configure ipmi, it will look for a bmc-rack1 network to draw its IPMI address from. Falling back to the first bmc network, I think. or none. [11:15] Also, the system attempts to last octet alignment across networks if possible. [11:15] What interface a network is configured on is determined by the conduit. [11:16] The conduit is only applicable to managed nodes (not the admin node). [11:17] For non-bmc networks, the conduit can be dhcp, or [-+?][1,10,100,40][gm][0-100] (e.g. 10g1 - which means second 10g found on box). [11:18] dhcp means that what ever the box dhcp on is the interface used. This is really only useful for the admin network. [11:18] For bmc networks, the conduit can be dhcp or bmc. DHCP says use DHCP to get the address for the bmc, other set the bmc statically. [11:19] conduits can be multiple interfaces separated by a comma. This will generate a bonded set [11:20] Vlans can also be added on top. [11:20] In general, admin and bmc don't use conduits or vlans. [11:21] When an admin or dhcp network is configured for DHCP correctly, you should be able to see in the DHCP nav section of the UX that configuration and current leases. You may have to hard refresh the UX page to get the latest data. [11:21] Networks also define routers. The pref of the router determines if it becomes the default gateway. Lower is high priority. [11:22] Sooo - when the network is configured the available interfaces are consulted and the lowest preference is used to set the default gateway.

greg
2017-06-01 12:47
Dump from another slack conversation.

zehicle
2017-06-02 17:38
@meshiest the link to the gitter chat on the rebar.digital site is wrong -> points to digitalrebar/digitalrebar but should be digitalrebar/core

meshiest
2017-06-02 17:38
has joined #community201706

2017-06-02 17:48
Howdy! I'm using the rebar-provision quickstart on osx 10.12.5, and after tools/discovery-load.sh runs, and a subnet is assigned, boxes are able to pull DHCP, but fail on pulling the next file in sequence: The traffic that comes via tcpdump is: 38 RRQ "lpxelinux.0^A^DM-^?M-^?M-|" octet blksize 1456"

2017-06-02 17:48
the failed message on the server console is the same

2017-06-02 17:48
*for the same file

2017-06-02 17:49
Any thoughts on where to look to see what's going on?

vlowther
2017-06-02 17:54
hm

vlowther
2017-06-02 17:55
I don't think the extra control characters should be part of the RRQ request.

vlowther
2017-06-02 17:55
What are the boxes?

zehicle
2017-06-02 17:55
thanks @meshiest

2017-06-02 17:56
The boxes are a bit older, based on https://www.asus.com/Commercial-Servers-Workstations/KFSN5DIST/specifications/

2017-06-02 17:56
provision is running locally on my mac, late 2012, retina, 10.12.5

2017-06-02 17:56
I added the route to point directly at said mac goo

2017-06-02 17:56
*too

vlowther
2017-06-02 17:57
hm

vlowther
2017-06-02 17:58
Is the firewall on the mac set to allow incoming TFTP to dr-provision?

2017-06-02 17:58
firewall is off

2017-06-02 17:59
tcpdump shows the packets getting there

vlowther
2017-06-02 18:02
ok

vlowther
2017-06-02 18:03
and lpxelinux.0 is present in drp-data/tftpboot ?

vlowther
2017-06-02 18:03
(it should have been automatically when the provisioner started up

2017-06-02 18:03
yes

2017-06-02 18:04
https://gist.github.com/bunchc/c5909f3289481c0cd9cf6002167c11f2

vlowther
2017-06-02 18:06
ok, that looks fine.

vlowther
2017-06-02 18:07
Have you created a network in dr-provisioner via the UI?

2017-06-02 18:07
Yup: https://i.imgur.com/MIUxBlQ.png

vlowther
2017-06-02 18:08
ok

vlowther
2017-06-02 18:08
What interface is en3?

2017-06-02 18:09
Thunderbolt ethernet

2017-06-02 18:09
hooked up to the same switch as the box pxe booting

vlowther
2017-06-02 18:11
ok

vlowther
2017-06-02 18:11
I am seeing if I can duplicate here.

vlowther
2017-06-02 18:12
Don't have a Mac handy to test with, but I do have a beefy Linux box.

2017-06-02 18:13
I can spawn an ubuntu VM if we need, was hoping to avoid that layer tho.

2017-06-02 18:19
Victor: I need to step away for a few hours. Will be back tho.

vlowther
2017-06-02 18:22
hm

vlowther
2017-06-02 18:22
I get 13:21:00.186474 IP (tos 0x0, ttl 64, id 1045, offset 0, flags [none], proto UDP (17), length 69) 192.168.124.41.13686 > m4723.tftp: [udp sum ok] 41 RRQ "lpxelinux.0" octet blksize 1432 tsize 0

vlowther
2017-06-02 18:23
that is with tcpdump -vvv

greg
2017-06-02 18:45
@bunchc. Make sure the subnet option with pxelinux.0 doesn't have extra content in it

2017-06-02 21:33
Greg: re: subnet option, the bit in the UI?

2017-06-02 21:42
Looks like it's getting the bits from somewhere, started over from the top and am now getting two different sets of control chars

2017-06-02 21:42
lpxelinux.0M-^? in addition to what it have before

2017-06-02 21:47
The latest with tcpdump -vvv https://gist.github.com/bunchc/4d9f65cc698bda950f8ae57ebe635c00 looks like it both sends lpxelinux.0 and the one with control charachters

2017-06-02 22:12
Looks like it's on the NIC end, am able to boot other boxes mostly fine

2017-06-02 22:13
Thanks!

zehicle
2017-06-02 23:46
cool - I'd be interested to know what made the NIC do that

2017-06-03 03:08
I'm not sure /why/ it's sending control charachters, other than it looks to be part of an older tftp spec. I'll dit a little bit once I get it sorted out.

2017-06-05 10:06
Hello, I've been using cobbler to set up bare metal nodes, and discovered DR as a potential alternative. I am looking through the table of contents in the docs and not finding much reference to bare metal installs. Is there a handy link?

2017-06-05 10:36
I manager to get it working after watching.

2017-06-05 10:37
https://youtu.be/LhqbfcCOgwY

2017-06-05 10:38
This was also helpful

2017-06-05 10:38
https://youtu.be/5YWMlYYuu-s

2017-06-05 13:57
Thank you, Simon, I'll take a look!

2017-06-05 20:04
@svallebro those videos will bring in the full Rebar infrastructure which does a lot beyond cobbler. If you want to start simpler with just replacing cobbler, check out the Provision subproject too: https://github.com/digitalrebar/provision

2017-06-05 20:04
sorry should have flagged @eegilbert instead of @svallebro !

2017-06-06 09:00
Thanks, Rob.

2017-06-07 05:01
I created the support ticket, https://rackn.freshdesk.com/support/tickets/49 . I a awaiting for response from the support. could you please help on the same...

2017-06-07 10:35
Hi all, I`m trying to install Digital rebar on clean Ubuntu 16.04 installation (bare metal) and I`m stuck at "wait for admin convergence". I used "curl -fsSL https://raw.githubusercontent.com/digitalrebar/digitalrebar/master/deploy/quickstart.sh | bash". What am I missing ?

greg
2017-06-07 13:46
@zdebinski - how big is the machine you running the admin node on?

greg
2017-06-07 13:46
You need at least 6GB of memory and ideally 4+ virtual CPUs.

2017-06-07 13:48
It should be enough I have 32GB of ram and eight cores

2017-06-07 13:51
I removed apparmor and managed to launch ui but install process failed timing out and ui was not running properly

2017-06-07 13:52
It`a also failing when dnsmasq-base is installed

greg
2017-06-07 13:59
Digitalrebar runs things that conflict with dnsmasq and ntp. The install ansible should kill all of those.

2017-06-07 14:02
I know I saw it tries to stop the services but still installation is failing if I`dont remove it

2017-06-07 14:27
@deepuashokan85 have you registered for RackN support? I'm not showing you in our system.

2017-06-07 14:40
@zehicle_twitter "cmd": "killall dnsmasq", "failed": true, "msg": "[Errno 2] No such file or directory", "rc": 2}" it should be killall5 not killall

2017-06-07 14:41
that is why it was failing

greg
2017-06-07 14:41
oh - ok - hmm - I bet that varies by OS type. :neutral_face:

2017-06-07 14:45
Ubuntu 16.04 - killall5 belongs to sysvinit-utils

2017-06-07 15:22
I`ve managed to finish installation, but adding deployments or adding nodes doesn`t work.

2017-06-07 15:22
evproxy_1 | 2017/06/07 15:20:13 requested url = /api/v2/deployments revproxy_1 | 2017/06/07 15:20:13 translated url = https://rebar-api-service/api/v2/deployments rebar_api_1 | 2017-06-07 15:20:13.514 [20207] [ERROR] EXCEPTION: param is missing or the value is empty: name rebar_api_1 | 2017-06-07 15:20:13.514 [20207] [ERROR] BACKTRACE: rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/strong_parameters.rb:251:in `require' rebar_api_1 | /opt/digitalrebar/core/rails/app/controllers/deployments_controller.rb:65:in `block in create' rebar_api_1 | /opt/digitalrebar/core/rails/lib/api_helper.rb:118:in `block in retriable_transaction' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activerecord-4.2.5/lib/active_record/connection_adapters/abstract/database_statements.rb:213:in `block in transaction' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activerecord-4.2.5/lib/active_record/connection_adapters/abstract/transaction.rb:184:in `within_new_transaction' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activerecord-4.2.5/lib/active_record/connection_adapters/abstract/database_statements.rb:213:in `transaction' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activerecord-4.2.5/lib/active_record/transactions.rb:220:in `transaction' rebar_api_1 | /opt/digitalrebar/core/rails/lib/api_helper.rb:116:in `retriable_transaction' rebar_api_1 | /opt/digitalrebar/core/rails/app/controllers/deployments_controller.rb:50:in `create' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/implicit_render.rb:4:in `send_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/abstract_controller/base.rb:198:in `process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/rendering.rb:10:in `process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/abstract_controller/callbacks.rb:20:in `block in process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:117:in `call' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:117:in `call' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:555:in `block (2 levels) in compile' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:505:in `call' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:505:in `call' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:92:in `__run_callbacks__' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:778:in `_run_process_action_callbacks' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/callbacks.rb:81:in `run_callbacks' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/abstract_controller/callbacks.rb:19:in `process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/rescue.rb:29:in `process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/instrumentation.rb:32:in `block in process_action' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/notifications.rb:164:in `block in instrument' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/notifications/instrumenter.rb:20:in `instrument' rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/activesupport-4.2.5/lib/active_support/notifications.rb:164:in `instrument'

2017-06-07 15:22
rebar_api_1 | /var/cache/rebar/gems/ruby/2.1.0/gems/actionpack-4.2.5/lib/action_controller/metal/instrumentation.rb:30:in `process_action' reba

2017-06-07 15:24
I dont have time to debug it further, I will switch to kargo

greg
2017-06-07 15:32
digitalrebar uses kargo and drives it. You need to add a provider if you are using it against cloud. If you are using metal, you have to define networks and OS install images. Thanks for trying it.

zehicle
2017-06-07 15:34
Your error is that name was not defined in the call

2017-06-07 15:36
I know :) but I provided it :worried:

greg
2017-06-07 15:36
Did you create a provider?

2017-06-07 15:37
I need to deploy kubernetes on metal

greg
2017-06-07 15:37
Did you have nodes discovered by the system?

2017-06-07 15:37
When I want to choose provider metal does not show

2017-06-07 15:38
I will back to it tommorow, thanks for the help :)

greg
2017-06-07 15:38
Yes - installing on baremetal means that you would select use existing nodes that have been discovered by DigitalRebar. These systems would need to PXE boot our discovery image and get added to DigitalRebar to be then consumed by the deployment wizard for kubernetes.

aaron.feng
2017-06-07 17:13
has joined #community201706

dubie
2017-06-07 17:45
has joined #community201706

ajones
2017-06-07 17:48
has joined #community201706

pdelarosa-riot
2017-06-07 17:48
has joined #community201706

eblack
2017-06-07 18:01
has joined #community201706

mpainter
2017-06-07 18:04
has joined #community201706

rcameron
2017-06-07 20:18
has joined #community201706

armagan
2017-06-07 21:36
has joined #community201706

2017-06-08 07:13
@zehicle What if I have dedicated servers that can`t be booted with PXE, that have system already installed and have only public IPs ? is it possible to use digital rebar in this case ?

greg
2017-06-08 13:31
Yes - you would "join" them to DR. There are join scripts in the digitalrebar/deploy directory that can be used as a starting point for that process.

rstarmer
2017-06-08 18:24
hey, how do I set DRP to use a local mirror (or rather the install ISO) to provision a node (using ubuntu-16.04, pre-downloaded, and that all seems to work as far as getting the system booted). I.e., I?m working in a somewhat disconnected environment, and am not sure how to proceed if I need to have a local mirror rather than just using the upstream ones

rstarmer
2017-06-08 18:25
we can use a proxy, but the upstream connection from the lab is _SLOW_ and so we?d rater use/build a local one if that?s the best approach.

rstarmer
2017-06-08 18:26
also, how do I add a profile (for ssh keys, for example) to a bootenv? It seems like the template is there (e.g. ubutntu-16.04 looks like it would use access_keys parameters)

jgartrel
2017-06-08 19:20
has joined #community201706

2017-06-08 19:38
there is a way to do it, I know we discussed it - I could not find the docs so I'm looking around

zehicle
2017-06-08 19:39
@vlowther do you remember how to set the repos for DRP to a local source?


vlowther
2017-06-08 19:51
set the local_repo param to true.

vlowther
2017-06-08 19:52
that will use just the repos off the expanded Ubuntu ISO.

vlowther
2017-06-08 19:54
If you want other repos from somewhere other than the exploded ISO, I don't think wer have a pre-baked template that does that.

rstarmer
2017-06-08 20:07
thanks guys, will have a look, and come back with questions.

rstarmer
2017-06-08 20:08
Ah, yes, ok so that template is a part of my ubuntu bootenv, but how do I set the parameter to be true?

rstarmer
2017-06-08 20:09
do I update the bootenv with a JSON/yaml blob?

rstarmer
2017-06-08 20:14
I do need help here, I don?t have a good model for how to pass parameters, for example, do I just update the bootenv with the parameters I?d like added?

rstarmer
2017-06-08 20:19
So with that template, if I want to add it to my ubuntu template, I can create a new template section, but it?s not clear where it ?lives?. I.e., what?s the path?

rstarmer
2017-06-08 20:24
ok, I see where those templates live and how they are ingested into files, but it isn?t clear how I set the local_repo parameter for that bootenv. I?ve tried a YAML object like ```Params: local_repo: true```

rstarmer
2017-06-08 20:25
but that gets rejected when I do: ``` drpcli bootenvs update bootenvs/ubuntu-16.04.yml local_repo.yaml```

rstarmer
2017-06-08 20:25
local_repo.yaml is where I put the Params: bit

vlowther
2017-06-08 20:28
yeah, that parameter has to be set on the machine, not the bootenv.

rstarmer
2017-06-08 20:29
do I set it directly on the machine? and how might I make that a default (possible?)

rstarmer
2017-06-08 20:32
I think I got it: ``` drpcli machines set "{uuid}" params "local_repo" to "true"```

rstarmer
2017-06-08 20:49
shouldn?t ``` drpcli machines params "{uuid}" params < params.yml``` work?

greg
2017-06-08 20:51
the first one

rstarmer
2017-06-08 20:53
but how do I pass a bunch of parameters?

rstarmer
2017-06-08 20:53
like access keys

rstarmer
2017-06-08 20:54
do I have to write out a text json blob rather than nice simply yaml?

greg
2017-06-08 20:56
I often do this: ```drpcli machines set j param rebar-access to "$(cat fred)"```

rstarmer
2017-06-08 20:56
yeah, but I wanted to set local_repo, and access_keys, and,?.

rstarmer
2017-06-08 20:57
well, it turns out params does happily read JSON, just not yaml (at least as text via the CLI)

rstarmer
2017-06-08 20:57
So, the second question is: how do I set that as a default on the bootenv, I don?t want ot have to do this for every machine after the fact :slightly_smiling_face:

greg
2017-06-08 20:58
global profile

greg
2017-06-08 20:58
drpcli profiles show global

greg
2017-06-08 20:58
Path the params there and all nodes gett them.

rstarmer
2017-06-08 21:00
I?ll give that a shot.

rstarmer
2017-06-08 21:01
kewl

rstarmer
2017-06-09 20:01
what?s the best way to change the hostname/ip addresses of discovered servers in DRP? Do I have to manually register the nodes, or can I update things somehow (my experience seemed to indicate that nodes didn?t like this to be changed and ended up out of sync with the system)

greg
2017-06-09 20:04
You can change the name on the machine and reboot into sledgehammer. The name is what is used as FQDN.

rstarmer
2017-06-09 20:04
will try. might be that I erased the domain when I tried last time

rstarmer
2017-06-09 20:04
how about IP address?

greg
2017-06-09 20:05
IP address comes from lease to mac to machine.

greg
2017-06-09 20:05
You can reserve an address by creating a resesrvation of mac to IP.

rstarmer
2017-06-09 20:06
how do I do that?

rstarmer
2017-06-09 20:06
drpcli?

greg
2017-06-09 20:06
yes

rstarmer
2017-06-09 20:07
can I just change the lease and reboot the node?

greg
2017-06-09 20:07
to get a different IP?

rstarmer
2017-06-09 20:08
yeah

greg
2017-06-09 20:08
Can try it. Not sure if that will work, haven't tried it.

rstarmer
2017-06-09 20:08
customer just said he didn?t care, but I like order (sometimes :)

rstarmer
2017-06-09 20:08
ok, may be moot, but I may try anyway.

rstarmer
2017-06-09 20:14
and ipmi control is not part of DRP right?

greg
2017-06-09 20:15
correct - it is not

rstarmer
2017-06-09 20:17
thn

rstarmer
2017-06-09 20:17
x

rstarmer
2017-06-12 17:30
default user for sledgehammer? Seems it should be root, but RocketSkates doesn't seem to work (we were able to add an ssh key, but wanted to get proper terminal access defined as well)?

2017-06-12 18:30
https://www.youtube.com/watch?v=uUWU-4ObGIY I was following this demo, came to the end with the same results BUT the machine was not discovered. The last line on the bootme machine on the video is: Connecting to 192.168.124.11:8091 I get the same, but then... wget: can't connect to remote host (192.168.124.11): Operation timed out Failed to download stage2.img for 70.... Which actually makes sense as the bootme vm is not on that subnet... so I'm confused... did the demo fail to discover the machine too, (seems like it would have)... or I'm confused... I dug around abit and figured out that 192.168.124.11 is the default admin node... but I was doing just the provisioner

zehicle
2017-06-12 18:31
@rstarmer I believe it's rebar

rstarmer
2017-06-12 18:36
@zehicle thanks

rstarmer
2017-06-12 18:36
will give that a shot.

zehicle
2017-06-12 18:41
@ctrees, the demo was missing a step that is added in the docs


zehicle
2017-06-12 18:41
On Darwin, you may have to add a route for broadcast addresses to work. This can be done with the following comand. The 192.168.100.1 is the IP address of the interface that you want to send messages through. The install script will make suggestions for you. sudo route add 255.255.255.255 192.168.100.1

2017-06-12 18:42
Thanks I did the route, (as it was hit in demo ;) I'll go check the docs... thanks1

2017-06-12 18:49
@ctrees the demo ends at the same IP request... checking on my end

2017-06-12 18:50
yea I for sure am not tracking... I checked https://127.0.0.1:8091 and it was not live...

2017-06-12 18:51
that should be the tftp endpoint ?correct?

2017-06-12 18:52
try w/o http

2017-06-12 18:52
that's for http static files, would be the same as the tftp root

2017-06-12 18:52
yup... (duh) that worked...

2017-06-12 18:55
the broadcast works but it's the 192.168.124.11 how and why is that even in there... I started to track that down which is why I came here... seems like it should be 127.0.0.1 OR 192.168.56.1 (or whatever was the vmlan network) ?

2017-06-12 18:56
I think that's the sledgehammer default - should be overridden from DHCP settings, I'm looking for the doc

vlowther
2017-06-12 18:58
This is DRP, right?

2017-06-12 18:58
yes

2017-06-12 18:58
yes

vlowther
2017-06-12 18:58
ok

vlowther
2017-06-12 18:59
What command line did you use to launch drp?

2017-06-12 18:59
sudo ./dr-provision --file-root=`pwd`/drp-data/tftpboot --data-root=drp-data/digitalrebar

2017-06-12 19:00
no background is all...

vlowther
2017-06-12 19:00
ok, that is fine.

vlowther
2017-06-12 19:01
Did you add a network to drp via the web UI?


vlowther
2017-06-12 19:03
hm

vlowther
2017-06-12 19:04
so sledgehammer should have been making requests back to 192.168.66.1


2017-06-12 19:04
yea... my thinking too

vlowther
2017-06-12 19:05
assuming that is the actual IP address on vboxnet1

2017-06-12 19:05
but then the same IP came up in the video BUT... it did not wait till the machine got discovered

2017-06-12 19:05
yea... see it in the route table...

2017-06-12 19:06
I kind of assumed that the 124.11 was the sledge default... so I started to hunt where you override that... that's where I got lost in scripts... I was looking for the startup.sh thing ...

2017-06-12 19:08
I went back to the video... figure I messed a config (just like the route)... BTW I love when 'common' mistakes are made and you show how to debug.... of course they make for longer video... but they sure help me more than 'doing it right the first time'... ;-)

vlowther
2017-06-12 19:09
Are you in Sledgehammer right now?

2017-06-12 19:10
just a sec... have to find that window... I'll screen it if so

vlowther
2017-06-12 19:10
no need

vlowther
2017-06-12 19:10
jsut paste the contents of /proc/cmdline


2017-06-12 19:11
@ctrees I'll make sure the docs get updated when we resolve this, thanks for your patience

2017-06-12 19:11
I have a heck of a time getting paste from vbox

vlowther
2017-06-12 19:12
ok

vlowther
2017-06-12 19:12
from there, run cat /proc/cmdline


vlowther
2017-06-12 19:14
hm

vlowther
2017-06-12 19:17
You are running drp on a mac?

2017-06-12 19:17
yea... I was going to start to play with how drp passes in configs... as this gets to the root of what I was attempting to replace... all our pxe/ks and internal IPAM stuff... seems like a great fit..

vlowther
2017-06-12 19:17
Not in a VM?

2017-06-12 19:17
yup

vlowther
2017-06-12 19:17
yeah

2017-06-12 19:17
mac mini

2017-06-12 19:17
16GB

vlowther
2017-06-12 19:18
so you will need to run drp with the --static-ip option


vlowther
2017-06-12 19:19
We have code in place that works reliably on Linux to figure out what IP address recieved a given packet so that we can handle running properly in a multi-homed environment.

vlowther
2017-06-12 19:19
but that code is less reliable on macs.

2017-06-12 19:20
Well... the 'quick fix' 4me is to custom the pxe to put the 192.168.66.1 IP in for default ?

2017-06-12 19:21
... wow... your telling me you on the fly ID that stuff in the DNS... that's cool...

vlowther
2017-06-12 19:21
Just pass --static-ip=192.168.66.1 to dr-provision at startup

vlowther
2017-06-12 19:21
no, not DNS

vlowther
2017-06-12 19:22
We inspect the raw IP packets for incoming DHCP and TFTP traffic to determine the IP address that the incoming packet was destined for

2017-06-12 19:23
I about did that... checking now ( --static-ip )

vlowther
2017-06-12 19:23
so we can set the address we live at appropriately in future responses that require it.

vlowther
2017-06-12 19:23
On Linux this works reliably.

vlowther
2017-06-12 19:24
On macs we sometimes get all zeroes or other nonsensical data, so we have to fall back to whatever --static-ip is set to

2017-06-12 19:24
rebooting bootme now...

2017-06-12 19:25
working

2017-06-12 19:26
looks good

vlowther
2017-06-12 19:26
this allows us to work properly in multihomed scenarios where a system lives on networks A and B but there is no route between A and B

vlowther
2017-06-12 19:26
and it allows us to work without needing a restart when network interfaces come and go.

2017-06-12 19:27
yea... I'm pretty jazzed about you approach... working up a demo for my boss/es

vlowther
2017-06-12 19:28
I am also impressed to see a mac mini server still being used in the wild. :slightly_smiling_face:

2017-06-12 19:28
oh heck... THATS my GOOD box... ;-)

2017-06-12 19:29
but I've got 3 piles of HPE C7000 I'm going to 'resurect' for devops

2017-06-12 19:32
If I get the AFS (Andrew, not Apple) setup into a workload... I'll for sure shove that back to you/community as we need to support AFS for more years and CERN is dropping it...

2017-06-12 19:32
I'm updating docs to improve notes on Mac

2017-06-13 00:11
Hello, I'm not sure if this is the right room to ask questions. But anyways, I'm seeing if I can try out digital rebar with vagrant. Following these directions here: http://digital-rebar.readthedocs.io/en/latest/deployment/install/vagrant.html

2017-06-13 00:12
The Vagrantfile points to a bad link, but looking around I found it at https://github.com/digitalrebar/digitalrebar/blob/master/deploy/Vagrantfile

2017-06-13 00:13
The Vagrantfile also references some scripts in there I do not have in this section: # # Admin nodes eat themselves without swap # base.vm.provision "shell", path: "scripts/increase_swap.sh" base.vm.provision "shell", path: "quickstart.sh"

2017-06-13 00:14
Where can I find these files? Or can I just comment that out?

2017-06-13 00:26
The errors I'm seeing when trying to bring it up is: # $ vagrant up base --provider=libvirt =========================================================== Welcome to Digital Rebar Vagrant Machine types available: client (run remote DR) base (DR server) node1[-20] (node for local test) Documentation: https://github.com/digitalrebar/doc/blob/master/deployment/vagrant.rst export REBAR_ENDPOINT and REBAR_KEY to use existing Digital Rebar Server (default = Vagrant admin) TRIGGERS REQUIRED: vagrant plugin install vagrant-triggers see http://www.rubydoc.info/gems/vagrant-triggers/0.2.1 REBAR CLI REQUIRED: rebar cli must be on your path see https://github.com/digitalrebar/doc/tree/master/cli Maintained by RackN, Copyright 2016 =========================================================== To monitor > https://192.168.99.100 (Digital Rebar) After the system is up, you can start the nodes using `vagrant up /node[1-20]/` Bringing machine 'base' up with 'libvirt' provider... There are errors in the configuration of this machine. Please fix the following errors and try again: shell provisioner: _ `path` for shell provisioner does not exist on the host system: /home/choyj/realdr/scripts/increase_swap.sh _ `path` for shell provisioner does not exist on the host system: /home/choyj/realdr/quickstart.sh Vagrant: * Unknown configuration section 'trigger'.

2017-06-13 03:40
@jack-likes-to-code it's been a while since we've updated those vagrant scripts. lately everyone runs docker locally, so they are OK w/ the compose approach on their laptop or the Ansible install from a VM or server

2017-06-13 03:40
I do know that you'll need to install the Vagrant triggers for it to work

zehicle
2017-06-13 03:56
@rstarmer I found the default user for debian in the docs: "rocketskates"

rstarmer
2017-06-13 04:36
? yeah, I tried the documented ansewrs already, but they don?t appear to work. I wonder if our addition of an ssh auth token changes the bootstrap code?

greg
2017-06-13 13:21
ssh keys are for root.

greg
2017-06-13 13:22
rocketskates/RocketSkates

greg
2017-06-13 13:22
should be the login user.

2017-06-13 13:54
I'll update the docs to clarify and see about putting it an faq also

2017-06-13 17:40
Hi @zehicle I started reading about/trying out digital rebar yesterday so I'm noobish in this area as my team is interested in using that for cloud deployment. I believe we(SUSE) have a meeting with you on Thursday to go over digital rebar, but I wanted to give it a try before going into the meeting. I was able to get it to run on a vagrant VM yesterday just following section 1.1.2. I could try again just getting it to run via the compose approach, but is there a way to use local VMs as nodes?

2017-06-13 17:42
ls

vlowther
2017-06-13 18:17
Sure, as long as the VMs are attached to thee same bridge that the admin node is serving DHCP on, and there are no conflicting DHCP servers.

vlowther
2017-06-13 18:19
If you are running on a Linux box, digitalrebar/core/tools/docker-admin will bring up the DR admin containers, and digitalrebar/core/tools/kvm-slave will spin up QEMU instances that attach to the docker0 bridge and PXE boot by default.

2017-06-13 18:20
oh! I'll have to look into that... installing Ubuntu 16.04 on another box right now. Will give this a try. Thx!

vlowther
2017-06-13 18:21
Those are the tools I use in my daily development.

vlowther
2017-06-13 18:22
has not teested them on opensuse in a while, tho.

vlowther
2017-06-13 18:22
Arch Linux is my daily driver.

2017-06-13 18:23
Currently using Ubuntu for now...

vlowther
2017-06-13 18:24
I won't tell anyone else. :slightly_smiling_face:

2017-06-13 18:28
lol, I'm a transplant from HPE that was acquired by SUSE 3 months ago... Luckily, SUSE is pretty open about these things. :)

2017-06-13 20:05
@jack-likes-to-code there's a video in the training sessions where we show how to do the docker-admin & kvm-slave path

2017-06-13 20:06
@zehicle I was just going to ask you about this

2017-06-13 20:06
can you point me to that?

2017-06-13 20:06
you should also make sure to try DR Provision - it's the newer model deployment

2017-06-13 20:07
I got to the point of bringing up digital-rebar, but I went ahead and ran the kvm-slave to bring up a node and it sits there with: 2017-06-13 13:01:21 -0700: 26660 - PXE booting node (0)

2017-06-13 20:07
should i have run the docker-admin script first?

2017-06-13 20:08
Not familiar with DR Provision. Can you point me to that documentation?

2017-06-13 20:09
https://youtu.be/OBK1Gkv0YH8?list=PLXPBeIrpXjfh2lXdXkNnzAuc7_SUtYJR-&t=243

2017-06-13 20:09
For DRP http://provision.readthedocs.io/en/latest/

2017-06-13 20:09
thanks, I'll take a look

2017-06-13 20:10
@jack-likes-to-code the challenge w/ the kvm-slave script is that is assumes your using docker0 as the bridge. you'll need to adjust the script to use another bridge if you used Vagrant to bring up the admin now

2017-06-13 20:10
not using vagrant anymore. Just installed ubuntu on a box and ran the quickstart script on it

2017-06-13 20:11
that brought up digital rebar and I can see the UI

2017-06-13 20:11
you can change the bridge by setting the "OCB_BRIDGE" environment variable

vlowther
2017-06-13 20:11
jack-likes-to-code: Yes, you need to run docker-admin and wait for Digital Rebar to come up before running kvm-slave

2017-06-13 20:11
I think kvm-slave should work in that case

vlowther
2017-06-13 20:11
Otherwise the VM will not be able to PXE boot and just sit there forever.

2017-06-13 20:12
ok

2017-06-13 20:12
BTW > Victor is using our Slack bridge, which impersonates by Gitter account. So there are really two people :)

2017-06-13 20:12
I was wondering why victor comes up in all your msgs

zehicle
2017-06-13 20:13
this is how my (Rob's comments) show up by typing from Slack

2017-06-13 20:14
For people who prefer Slack, we're happy to provide invites to that channel. either way, the comments are mirrored

2017-06-13 20:16
@zehicle running docker-admin gives me the following error: ERROR: for forwarder driver failed programming external connectivity on endpoint compose_forwarder_1 (b3cc8cd5024e4c714ec3c14c70fe65d8f745ca7d55943a82cc1df5e9c0916151): Bind for 0.0.0.0:3000 failed: port is already allocated Traceback (most recent call last): File "<string>", line 3, in <module> File "compose/cli/main.py", line 63, in main AttributeError: 'ProjectError' object has no attribute 'msg' docker-compose returned -1 Bringing containers up. Exiting the shell will kill and remove the containers It looks like docker-proxy is already listening on port 3000

2017-06-13 20:17
is it because I'm already running digital rebar? Perhaps I should bring that down first before running docker-admin?

2017-06-13 20:35
yes. from digitalrebar/deploy/compose, use docker-compose stop then docker-compose rm -f

2017-06-13 20:44
@zehicle done. I kicked off docker-admin. Looks like ntp service has been sitting around: ```dev@Z640-extra:~/digitalrebar/deploy/compose$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 5699fd40e04f digitalrebar/dr_rebar_api:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_rebar_api_1 cb604b7ea9c0 digitalrebar/dr_goiardi:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_goiardi_1 c4e400eddb06 digitalrebar/cloudwrap:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_cloudwrap_1 87050fbf06bf digitalrebar/dr_trust_me:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_trust_me_1 1557912d03d7 digitalrebar/dr_webproxy:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_webproxy_1 faf9eaaa8484 digitalrebar/dr_rev_proxy:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_revproxy_1 d8853af62a83 digitalrebar/dr_rebar_dhcp:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_dhcp_1 c3b1b4323d2c digitalrebar/dr_provisioner:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_provisioner_1 faf2f86e6631 digitalrebar/logging:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_logging_1 6f7e5d0761e9 digitalrebar/dr_dns:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes compose_dns_1 d7880c396047 digitalrebar/dr_forwarder:master "/sbin/docker-entr..." 3 minutes ago Up 3 minutes 0.0.0.0:443->443/tcp, 0.0.0.0:3000->3000/tcp compose_forwarder_1 555d43930ce9 digitalrebar/dr_postgres:master "/docker-entrypoin..." 3 minutes ago Up 3 minutes compose_postgres_1 9696410fe96d digitalrebar/rule-engine:master "/sbin/docker-entr..." 3 minutes ago Up 31 seconds compose_rule-engine_1 7a49fbda0ce0 gliderlabs/consul "/bin/consul agent..." 3 minutes ago Up 3 minutes compose_consul_1 de8cea4fbee6 digitalrebar/dr_ntp:master "/sbin/docker-entr..." About an hour ago Up About an hour 0.0.0.0:123->123/tcp, 0.0.0.0:123->123/udp compose_ntp_1

2017-06-13 20:47
can't seem to figure out preformat rebar/rebar1 creds no longer work. What are the new creds when running docker-admin?

zehicle
2017-06-13 20:49
it takes a little while before the user is created

2017-06-13 20:50
thx. i guess i was too impatient

2017-06-13 20:57
still no cigar with kvm-slave. i am able to login to the UI, but kvm-slave still gets stuck at PXE booting node (0). You mentioned that docker0 is the assumed bridge. How do I determine that? `dev@Z640-extra:~/digitalrebar/deploy/compose$ docker network ls NETWORK ID NAME DRIVER SCOPE 1a30c0576175 bridge bridge local 918aa569f74b host host local 902721abd2e5 none null local

2017-06-13 21:00
Trying this again: ``` dev@Z640-extra:~/digitalrebar/deploy/compose$ docker network ls NETWORK ID NAME DRIVER SCOPE 1a30c0576175 bridge bridge local 918aa569f74b host host local 902721abd2e5 none null local ```

2017-06-13 21:03
It seems everything is up on docker0: ``` dev@Z640-extra:~/digitalrebar/deploy/compose$ brctl show bridge name bridge id STP enabled interfaces docker0 8000.0242b3fb13c2 no veth1c3107d veth3ad8868 veth3d311d6 veth42dbc38 veth5268663 veth73ba994 veth74f95ac veth91e9b8a vethc2366ce vethe6c4c02 vethf3a9141 vm-88964-0 vm-88964-1 vm-88964-2 virbr0 8000.000000000000 yes ```

2017-06-13 21:04
there's a IP bridge that you need... forgot to mention!

2017-06-13 21:04
sudo ip a add 192.168.124.4/24 dev docker0

2017-06-13 21:05
http://digital-rebar.readthedocs.io/en/latest/development/dev_env/kvm-slaves.html?highlight=kvm-slave

2017-06-13 21:06
then restart the VMs

2017-06-13 21:37
Thanks! That worked! More reading!

2017-06-13 21:38
:)

2017-06-14 23:22
I see my nodes came up with sledgehammer, then I figured I'd try and deploy Ubuntu on there, but it got stuck trying to find an NTP server. I walked away and eventually the kvm consoles went blank and I couldn't get them out of the blank state. DR shows the nodes as green. It doesn't appear the NTP service is running on there when running docker-admin. How do I bring that up? ``` dev@Z640-extra:~/digitalrebar/deploy/compose$ docker-compose ps WARNING: The DR_TAG variable is not set. Defaulting to a blank string. Name Command State Ports ------------------------------------------------------------------------------------------------------------- compose_cloudwrap_1 /sbin/docker-entrypoint.sh Up compose_consul_1 /bin/consul agent -config- ... Up compose_dhcp_1 /sbin/docker-entrypoint.sh Up compose_dns_1 /sbin/docker-entrypoint.sh Up compose_forwarder_1 /sbin/docker-entrypoint.sh Up 0.0.0.0:3000->3000/tcp, 0.0.0.0:443->443/tcp compose_goiardi_1 /sbin/docker-entrypoint.sh Up compose_logging_1 /sbin/docker-entrypoint.sh Up compose_postgres_1 /docker-entrypoint.sh postgres Up compose_provisioner_1 /sbin/docker-entrypoint.sh Up compose_rebar_api_1 /sbin/docker-entrypoint.sh Up compose_revproxy_1 /sbin/docker-entrypoint.sh Up compose_rule-engine_1 /sbin/docker-entrypoint.sh Up compose_trust_me_1 /sbin/docker-entrypoint.sh Up compose_webproxy_1 /sbin/docker-entrypoint.sh Up ```

2017-06-15 00:35
very strange! you should be able to docker-compose ntp up -d. When you install, using --con-ntp should work; however, it should be included in your default set

2017-06-15 00:35
check the NTP role in the system deployment

2017-06-15 00:36
and see how that is set, but it should have have turned green if the service was missing unless you picked external ntp.

2017-06-15 13:34
@jack-likes-to-code FWIW, NTP was in your earlier docker ps. Very strange that it's not in the docker-compose ps list!

2017-06-15 16:29
we're tracking down a very critical typo in the docs... default password is "rocketsk8ts" - it is correct in the source, but not updated in the generated doc pages

2017-06-15 16:30
NOTICE: DOCS ARE UPDATED

greg
2017-06-15 17:21
I need to move tip to get latest updated. We may need to cut a release to update stable.

zehicle
2017-06-15 17:32
I had RTD rebuild and it worked

2017-06-15 18:38
@zehicle docker-compose ntp up -d resulted in a "No such command: ntp". No ntp in docker-compose.yml either.

greg
2017-06-15 20:33
@Jack-likes-to-code: did you modify common.env? That can alter NTP operations.

2017-06-15 20:40
@zehicle I haven't touched that file. Don't know what it's used for

2017-06-15 20:42
Here's what I have for the NTP section in that file: ``` # # NTP Parameters # # Should we run NTP as intermediate or only # If EXTERNAL_NTP_SERVERS is not specified, # the local time in the container will be used. # NTP_RUN_PROXY=YES # This defines the upstream NTP servers # If NTP is running, it will forward to these. # If NTP is not running, then this is injected set. # Comma separated list of ips or names EXTERNAL_NTP_SERVERS= ```

2017-06-15 21:11
under digitalrebar/deploy/compose look at the init_files.sh script. That changes that compose.yml. Something in your system much have tweaked inputs to that file when you restarted. it could have been an environment variable.

2017-06-15 21:12
i do remember when I first did the quickstart, the ntp container was started. When I shut it down, it did not stop/remove the ntp service so i manually did that via docker

2017-06-15 21:13
then when I started DR via docker-admin, the ntp service did not come back up

2017-06-15 21:14
who kicks off init_files.sh? Should I kick it off again?

zehicle
2017-06-15 22:01
the install script uses it at the end of the ansible run

zehicle
2017-06-15 22:02
the run-in-system script can be run multiple times

zehicle
2017-06-15 22:02
but will tear down the running system

2017-06-15 22:02
ok. I may try that another time, then.

zehicle
2017-06-15 22:02
if you want to keep your config, then you need to modify the config files

zehicle
2017-06-15 22:02
those are input into the install

2017-06-15 22:02
It's fairly easy to setup, so no harm done in resetting it up

2017-06-19 18:23
Hi, It's possible to use CoreOS as OS? Thanks

greg
2017-06-19 19:17
For what piece of code?

2017-06-19 19:20
The master and nodes (I already have an Ubuntu box with docker-compose running)

greg
2017-06-19 19:42
okay - so DR as opposed to DRP.

greg
2017-06-19 19:43
I've been playing with CoreOS and DRP. It is possible. DR should be similar

greg
2017-06-19 19:43
I created a custom bootenv that served the kernel and initrd and then served an ignition file that start the components I wanted.

greg
2017-06-19 19:43
It could have installed instead.

greg
2017-06-19 19:44
I don't have anything checked in. It was more a play-with-it-to-see if it could be done.

greg
2017-06-19 19:44
Nothing real.

greg
2017-06-19 19:44
Right now

spector
2017-06-26 15:02
has joined #community201706

spector
2017-06-26 15:02
Spector here

2017-06-27 12:39
Hello! I have been looking at the OpenStack Helm project and ran into digital rebar.. It looks pretty impressive! I noticed however that the quick start guide does not include bare metal provisioning. My goal is to get to this point: https://www.youtube.com/watch?v=6xuVm9PJ2ck Any advice?

zehicle
2017-06-27 13:41
the docs will take you through an physical install and there are videos (http://digital-rebar.readthedocs.io/en/latest/deployment/README.html). We (RackN) would be to have a call with you to talk about the environment before you start. We find that a short call can save a lot of time in networking.

2017-06-27 13:50
I am much more of a video person (which I am super thankful that y'all have so many!). Do y'all have any videos covering the setup with bare metal?

2017-06-27 23:04
it's different for each person due to their setup - that's why we suggest a live meeting.

2017-06-29 22:32
I did the quick config on a 32GB Ubuntu 16.08 clean install: curl -fsSL https://raw.githubusercontent.com/digitalrebar/digitalrebar/master/deploy/quickstart.sh | bash -s -- --con-provisioner --con-dhcp --admin-ip=192.168.88.228/24

2017-06-29 22:36
Of which the 192.168.88.228 was the DHCP IP handed out on the .88 sub by a mikrotik router.... I'm still waiting for the ansible playbook to end, but I had started to poke around on the web UI... I had earlier installed the provisioner ONLY on another machine on another test network

2017-06-29 22:38
Your comments above lead me to believe the provisioner is not installed ? http://provision.readthedocs.io/en/stable/doc/quickstart.html

greg
2017-06-29 22:41
You are mixing docs.

greg
2017-06-29 22:42
Your command line is for digitalrebar and not digitalrebar provision (the doc link). You want the http://digitalrebar.readthedocs.io/en/stable/doc/quickstart.html

greg
2017-06-29 22:42
WIth that command line, you should have the provisioner and the dhcp server

2017-06-29 22:47
OK... I though so, just was not sure if the other is better to use for metal ? I noticed you've broke it out into an independent github repo... but you've got such flexibility with all these services it, I have to check the 'spinning top'

greg
2017-06-29 23:06
Yeah - we are slowly splitting things out for better control and non-monolithic services. Slowly.

jacob
2017-06-30 16:14
has joined #community201706

2017-06-30 16:29
OK... I 'think' I understand my DHCP/PXE problem... see if I can explain as I'm sure it's just me missing something. I've got a HPE C7000 that I'm testing with... I pulled all the blades and drives, put in one blade and one new drive, loaded Ubuntu on that and did the quick config mentioned above. I think my issue is that the fabric switches (Cisco 3020), OnboardAdmin Module, and iLO use the little MikroTik to grab management IP's via DHCP. I listened to your RackN Digital Rebar Training (006): Configure DHCP and Network, where I extracted the fact that you need to put an IP on the admin DHCP network. So, I went into the UI, DHCP Subnets, admin-internal where the subnet was 192.168.99.0/24. Given that network, I added 192.168.99.2 IP to enp3s0 (static) and ifup. I went out to the mikrotik and disabled it's DHCP. I then inserted another blade, set it to PXE boot.... thinking this 'should' work.

2017-06-30 16:32
The new blade is now in a PXE boot-loop, not finding the DHCP/PXE that I 'think' is running on 192.168.99.2 enp5s0

2017-06-30 16:33
OR... is the 'ADMIN_IP' need to change from 192.168.88.2 enp3s0 (as installed) to 192.168.99.2 enp5s0

2017-06-30 16:35
I'll go through that video again and take another look at the docs... that video does have a lot of deep 'gems' plus I know it showed how to get to the logs...

2017-06-30 16:39
What I might do is put DR on a machine OUTSIDE of the c7000 and let the DR machine capure all the DHCP... I may be fighting all the management IP DHCP from the OA, iLO and interconnect switches of the c7000.