2017-08-03 22:19
Hi, 3 days ago, you said something about Job/Task provisioning, what did you mean by that?

greg
2017-08-03 22:20
The tree and been picking up changes for this the last few days and tip is changing quite a bit for it.

greg
2017-08-03 22:21
Tasks are a collection of templates that can be ?run? in a bootenv environment by a runner. (The runner is built into the cli.)

greg
2017-08-03 22:22
A job is a run of a task on a machine. These have dates and logs and status. They can be retried and sequenced.

greg
2017-08-03 22:22
The idea is that you can use these to do advanced templatized actions on machines during various stages of provisioning.

greg
2017-08-03 22:23
RackN is going to make content bundles of supported features with this system.

greg
2017-08-03 22:23
But that is all for now. More is going on that soon.

2017-08-03 22:25
And a different question, is the use-case of _not_ installing an OS to disk, but simply always start through PXE, supported in your design?

greg
2017-08-03 22:26
Yes - in fact, we prefer and recommend that machines always PXE so that DRP has better control over the system.

2017-08-03 22:26
say I'd want a Ceph OSD image always being run by a node, but running it from mem?

2017-08-03 22:26
Ah ok

greg
2017-08-03 22:27
I have some toy unsupported and unpublished CoreOS templates that are baiscally that. There is also a linux kit variant in the tree currently that does something similar.

2017-08-03 22:27
So in time, DRP would have to be really high-available, since it will be crucial after a calamity, when bringing machines all up

2017-08-03 22:29
Yeah, I've been wanting to check out Linux kit. As far as you know, does LinuxKit itself assume this same model of always starting afresh?

greg
2017-08-03 22:29
well - kinda. Depends upon where and what. In the case, locally installed OS, not as much. The boot order set to PXE then Harddisk handles that kind of outage. You can?t provision new, but existing system function. In the case of immutable RAM boot OS, it is more crucial and an HA form is needed. now, we have plans for how that would work and strangely enough DRP is pretty close to HA now. It is not quite there yet.

greg
2017-08-03 22:30
LinuxKit in general assumes an immutable OS layer with docker running in that layer and containers on top.

greg
2017-08-03 22:30
It is in fact that docker OS immutable layer.

greg
2017-08-03 22:31
It could be installed, but appears to be more targeted at CoreOS style downloadable OS.

2017-08-03 22:32
But furthermore, you're telling me to wait a bit for all the new functionality which is in the pipeline :) ?

greg
2017-08-03 22:32
You could, but I think you can do a lot with what is there.

2017-08-03 22:33
Ow, and one last thing, something which I couldn't find in the docs, what's the difference between the default and the unknown bootenv?

greg
2017-08-03 22:35
default bootenv is the boot environment that machine gets when it is create without a bootenv specified.

greg
2017-08-03 22:35
unknown bootenv is the boot environment handed out by the DHCP server for a machine that isn?t know.

greg
2017-08-03 22:36
unknown bootenv populates the default.ipxe, default elilo, and default lpxelinux file. It is the fail through it nothing else is specified.

greg
2017-08-03 22:37
It is what is known as an UnknownOnly boot env. Only unknown machines will boot from it (usually, everything as nuances).

greg
2017-08-03 22:37
Examples of these are ignore and discovery.

2017-08-03 22:37
Ok, I get the last one. When would the default one kick in when using the webUI, because if I create a machine manually, I _have_ to choose a bootenv anyway.

greg
2017-08-03 22:38
The default bootenv is what the machine?s bootenv becomes whtn the machine is created.

greg
2017-08-03 22:38
An example of this sledgehammer (or any other non-UnknownOnly bootenv).

greg
2017-08-03 22:39
If you create a machine and specify a boot env (like ubuntu-16.04-install). that will be the machines boot env. If the machine boots, it will get directed to that boot env and ubuntu will be installed.

greg
2017-08-03 22:43
If you don?t create the machine and don?t specify a bootenv in the object, it will get the default. We usually set it to sledgehammer so that the machine can continue to be discovered dynamically.

2017-08-03 22:43
So: 1. discovery (assuming i had the unknown bootenv set to sledgehammer for example). 2. Once discovered it will assign it the default bootenv?

2017-08-03 22:43
ehh the second 1. should've been 2.

greg
2017-08-03 22:43
unknown is set to discovery.

greg
2017-08-03 22:43
then it will set the node to sledgehammer (it is a bit pedantic about that).

greg
2017-08-03 22:44
We could have left it blank and let the default take over.

greg
2017-08-03 22:44
Actually, just a second..

greg
2017-08-03 22:45
Yeah - I remembered ?rightly?. the discovery template control.sh creates the machine with sledgehammer.

greg
2017-08-03 22:45
We didn?t want to force install things without user intervention by accident.

2017-08-03 22:46
Sorry, but then I still don't understand at which step the default bootenv is set, and/or used.

greg
2017-08-03 22:47
only when the machine is manually created by the user without a bootenv in the specified JSON.

2017-08-03 22:48
Aah, so not through webUI

2017-08-03 22:48
in the webUI you're forced to choose a env

greg
2017-08-03 22:48
well - the UI knows to put something.

2017-08-03 22:53
THen the super last thing, how much mem does starting sledgehammer take? i'm little constrained mem wise in my small POC setup. I can only give 768M to a test VM which does PXE. Though it crashes with a low-mem warning

greg
2017-08-03 22:54
Yeah - that is probably to small. 1.5 G or so. Sledgehammer is ?bloated? with tools and environment pieces to identify components on real servers. So, it is a little heavier that what you are trying on.

greg
2017-08-03 22:54
Sorry

2017-08-03 22:54
That's okay, was just wondering

2017-08-03 22:54
Thanks, so far again. Bedtime for me now

greg
2017-08-03 22:55
Later

zehicle
2017-08-03 23:15
listening to your vm

zehicle
2017-08-03 23:18
no need to call - thanks for the detailed voice mial

lae
2017-08-04 00:20
does sledgehammer not have serial console support?

greg
2017-08-04 00:26
it does, but you need to tell it which one.

greg
2017-08-04 00:27
Or really, it doesn?t, but it will configurable. It depends upon which version of templates you have.

greg
2017-08-04 00:27
you need to console=? on the bootparams line. The problem is that varies depending upon your environment. Sooo, you need to add it in your bootenvs.

lae
2017-08-04 00:27
just did a fresh install of 3.0.5 and tried adding console=ttyS1,115200n8 to sledgehammer, but nothing shows up :<

lae
2017-08-04 00:28
(and bootup stopped showing on tty0)

greg
2017-08-04 00:28
okay - you may need to add it to discovery as well.

lae
2017-08-04 00:28
yeah, I made the changes to discovery, too, though once the machine's been discovered it just goes straight to sledgehammer I assume

greg
2017-08-04 00:28
yes

greg
2017-08-04 00:29
unless in packet?.

lae
2017-08-04 00:29
this is on-premise

greg
2017-08-04 00:29
should be okay then. is it just sledgehammer console?

lae
2017-08-04 00:30
the existing debian install on the system I'm testing it on outputs to serial console fine, if that's what you're asking

greg
2017-08-04 00:31
well - I?ve been playing with it and it is strange. I haven?t found a set of options that work consistently everywhere.

greg
2017-08-04 00:32
I have a coming change that has this as bootparam in sledgehammer and discovery.

greg
2017-08-04 00:32
```rootflags=loop root=live:/sledgehammer.iso rootfstype=auto ro liveimg rd_NO_LUKS rd_NO_MD rd_NO_DM provisioner.web={{.ProvisionerURL}} rs.uuid={{.Machine.UUID}} rs.api={{.ApiURL}} -- {{ if .ParamExists \"kernel_console\"}}{{.Param \"kernel_console\" }}{{ end }}```

greg
2017-08-04 00:33
After the ? seems to be required and then console=ttyS1,115200n8 or console=ttyS0,115200n8 seems to work.

greg
2017-08-04 00:33
now bios settings can mess with this, though if you have an ubuntu install that works. console line from that should work.

greg
2017-08-04 00:34
For ubuntu I had to put the console string on both sides of the --

lae
2017-08-04 00:34
will try in a bit

lae
2017-08-04 01:06
no cigar

lae
2017-08-04 01:12
hm

lae
2017-08-04 01:15
oh, this machine's actually ttyS2 - never mind then, it works. (looks like I was looking at the wrong item in my inventory whose serial-over-LAN console is ttyS1, so I got confused)

greg
2017-08-04 01:15
whew

2017-08-04 03:30
Working on t-shirt design - please feel free to vote https://99designs.com/contests/poll/j4amxx

2017-08-04 12:15
@zehicle , I am getting the below error when I tried to upload the iso image using rebar script [root@test10 ~]# ./rebar provisioner isos upload CentOS-6.8-x86_64-bin-DVD1.iso as CentOS-6.8-x86_64-bin-DVD1.iso 2017/08/04 09:45:41 ID not set panic: ID not set goroutine 1 [running]: log.Panic(0xc4200dfa48, 0x1, 0x1) /usr/lib/go/src/log/log.go:322 +0xc0 github.com/digitalrebar/digitalrebar/go/rebar-api/api.(_Client).UrlTo(0xc420054be0, 0xa6c620, 0xc420140140, 0x0, 0x0, 0x0, 0xa64900, 0xc42014c300) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/api/client.go:87 +0xb4 github.com/digitalrebar/digitalrebar/go/rebar-api/api.(_Client).Read(0xc420054be0, 0xa6c620, 0xc420140140, 0x0, 0x0) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/api/types.go:132 +0x6a github.com/digitalrebar/digitalrebar/go/rebar-api/api.(_Client).Fetch(0xc420054be0, 0xa6c620, 0xc420140140, 0x0, 0x0, 0x0, 0x24) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/api/types.go:145 +0x9c github.com/digitalrebar/digitalrebar/go/rebar-api/api.Session(0x88068a, 0x16, 0x0, 0x0, 0x0, 0x0, 0x73da86, 0xc4200fbcb0, 0xc42011d3d0) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/api/client.go:184 +0x34d main.main.func1(0xc420138d80, 0xc42011d500, 0x3, 0x3) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/rebar/main.go:282 +0x2fb github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra.(_Command).execute(0xc420138d80, 0xc42011d3b0, 0x3, 0x3, 0xc420138d80, 0xc42011d3b0) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra/command.go:619 +0x4d5 github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra.(_Command).ExecuteC(0xa941c0, 0xc4200dff60, 0xa94301, 0xc4201438c0) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra/command.go:722 +0x339 github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra.(_Command).Execute(0xa941c0, 0xc4200dff58, 0x1) /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/vendor/github.com/spf13/cobra/command.go:681 +0x2b main.main() /home/victor/gocode/src/github.com/digitalrebar/digitalrebar/go/rebar-api/rebar/main.go:353 +0x35d [root@test10 ~]#

greg
2017-08-04 12:40
Make sure that you have REBAR_KEY and REBAR_ENDPOIINT to valid value.

greg
2017-08-04 12:40
or pass -E and -U and -P on the cli

2017-08-04 12:47
@deepuashokan85 if you are focused on provisioning, we strongly recommend starting with the DR Provisioner (https://github.com/digitalrebar/provision) instead of the full Rebar stack. Provision is specifically designed for the DHCP/PXE automation phases and can be integrated back into Rebar.

lae
2017-08-05 01:10
congrats on the 1000th commit for provision btw

greg
2017-08-05 02:33
we?ve been flying. :slightly_smiling_face:

zehicle
2017-08-05 03:24
!! Wow. That's a big millstone

greg
2017-08-05 04:05
Warning - swagger committed a change today that broke our swagger generation. I?m looking at how to fix it now.

greg
2017-08-05 04:05
This effects people trying to build tip.

greg
2017-08-05 16:46
back working

lae
2017-08-07 20:31
does provision just extract the `IsoFile`?

lae
2017-08-07 20:32
I have some ramdisks and kernels for some in house operating systems that I need to create bootenvs for - I'm guessing I just need to tar them up and define Initrd and Kernel correctly

greg
2017-08-07 20:33
Yes - it explodes it in the tftpboot directory so it can be served.

greg
2017-08-07 20:33
Yes. tars are valid ?isos? - look at the sledgehammer tarball

lae
2017-08-07 20:33
yeah, that's where I got my assumption from. thanks

lae
2017-08-09 21:32
How can I configure DRP so that machines boot locally by default?

lae
2017-08-09 21:32
``` $ drpcli prefs set unknownBootEnv local Error: BootEnv local cannot be used for the unknownBootEnv ```

lae
2017-08-09 21:33
and setting defaultBootEnv to local doesn't seem to actually set it to be local

greg
2017-08-09 21:33
local => ignore

greg
2017-08-09 21:33
local is the machine-specific bootenv

greg
2017-08-09 21:33
ignore is the discovery bootenv

lae
2017-08-09 21:34
ah

lae
2017-08-09 21:34
Thanks

2017-08-10 00:31
Greetings folks! I discovered that Canonical broke the quickstart process by releasing 16.04.3, and opened a bug and pull request to (hopefully) address. Let me know if I need to sign paperwork or whatnot.

greg
2017-08-10 01:17
Thanks. I'll pull it shortly just so it is in the tree

greg
2017-08-10 01:19
I have a set of changes that are coming that move content in to a separate repo to handle changes like this cleaner. I have fixed it there already. That report has automated tests to catch this errro. And the like. We are trying to get ahead of these issues.

greg
2017-08-10 01:39
@edolnx - my comments are for DRP. I?ve pulled it into DR.

wdennis
2017-08-10 16:43
Hi @greg - updated to latest stable this morning (v3.0.5) and I'm experiencing problems with both UI and drpcli

wdennis
2017-08-10 16:44
@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F6MRS9R0E/image_uploaded_from_ios.jpg and commented: Here is what I see from drpcli:

lae
2017-08-10 16:44
ah

lae
2017-08-10 16:44
moving assets to a separate repo?

lae
2017-08-10 16:45
(I had forked it to our local git repo so I could modify assets/)

wdennis
2017-08-10 16:45
UI just doesn't connect in browser...


greg
2017-08-10 16:46
list -> show

wdennis
2017-08-10 16:47
Aha, that works...

greg
2017-08-10 16:47
@lae - yeah - we are having a hard time supporting everything and updating everything. So we are going to split them into separate repos with more version control stuff and layering.

wdennis
2017-08-10 16:47
Should wrong keyword cause an abort tho?

greg
2017-08-10 16:47
no

greg
2017-08-10 16:47
:slightly_smiling_face:

wdennis
2017-08-10 16:48
OK, QA hat on here...

wdennis
2017-08-10 16:49
If I type 'drpcli bootenvs bogus ...' I do get a Usage: reply

greg
2017-08-10 16:49
yeah - list may be doing something ?strange?. I have cli test for most of it.

wdennis
2017-08-10 16:50
'drpcli bootenvs list ...' causes the panic

greg
2017-08-10 16:50
drpcli bootenv list

greg
2017-08-10 16:50
works?

greg
2017-08-10 16:50
I know what is going on. I?ll fix it and add a test.

greg
2017-08-10 16:50
The list command lets you do index filtering.

wdennis
2017-08-10 16:50
Yes, works

greg
2017-08-10 16:51
```drpcli bootenvs list Available=true```

greg
2017-08-10 16:51
I don?t test if you don?t send an equals.

wdennis
2017-08-10 16:51
OK, so now about the UI...

greg
2017-08-10 16:51
yeah

wdennis
2017-08-10 16:52
I do see TCP listeners on port 8091 and 8092

greg
2017-08-10 16:52
https://<ip>:8092

wdennis
2017-08-10 16:52
But if I try that URL, no go...

greg
2017-08-10 16:53
that one is strange

greg
2017-08-10 16:54
that address is part of that host? ```ip a``` shows it?

wdennis
2017-08-10 16:55
Yes, confirmed using correct IP

greg
2017-08-10 16:55
cool - okay - firewall rules?


greg
2017-08-10 16:57
iptables getting in the way?

wdennis
2017-08-10 16:58
Maybe that's it - had full DR running on this node prev, Docker uses iptables-- when I did a 'iptables-L' I see 1,000 rules :)


wdennis
2017-08-10 16:59
You know the command to del all chains & flush again?

wdennis
2017-08-10 17:00
I have also removed docker-engine* pkgs now

greg
2017-08-10 17:02
iptables -F

greg
2017-08-10 17:02
is what I do

greg
2017-08-10 17:02
I think.

wdennis
2017-08-10 17:03
N/m, sorted it - combo of -F and -X flags

wdennis
2017-08-10 17:04
And now I have a UI :)

greg
2017-08-10 17:04
Yeah

wdennis
2017-08-10 17:04
(And found a bug)

greg
2017-08-10 17:04
Yep - I?ll add it to more list.

wdennis
2017-08-10 17:04
Oh, tabs on UI now - slick

greg
2017-08-10 17:05
Which one did you get?

greg
2017-08-10 17:05
it is undergoing a lot changes and will be different. More on that later.

wdennis
2017-08-10 17:05
3.0.5 (stable)

greg
2017-08-10 17:05
yeah, but it my mind that doesn?t help. :slightly_smiling_face:

lae
2017-08-10 17:06
from april

wdennis
2017-08-10 17:06
Well, it's better than "one long page with everything" design :)

lae
2017-08-10 17:07
oh yeah, would ya suggest to run tip

wdennis
2017-08-10 17:08
I'm actually using it in production so rather be on 'stable'

greg
2017-08-10 17:08
stable is fine.

greg
2017-08-10 17:09
tip is usually stable (the unit tests are good for the core pieces). The templates and such not as much. Hence, some of the changes coming.

greg
2017-08-10 17:10
Tip is passing unit tests, and has one cavaet on upgrade. subnets can be enabled/disabled now without deleting them. The data migration makes it off by default. :disappointed:

lae
2017-08-10 17:10
I'm running my own version of the templates - unless you mean handling of the templates has been changed a bit

greg
2017-08-10 17:10
templates are good - the actual values in templates are the issue.

greg
2017-08-10 17:11
We are working on a content layering system that hopefully will be simple but let you have your own content versions that we can still update. Almost there.


wdennis
2017-08-10 17:15
Still getting this when I click "API Help" links in UI tho

greg
2017-08-10 17:17
okay - bug - that swagger ui isn?t working again.

lae
2017-08-10 17:19
you'll need to pass in the current IP, @wdennis

lae
2017-08-10 17:19
that's a CORS issue

greg
2017-08-10 17:20
oh - yeah - the IP in the browser URL and the IP in the form need to be the same.

greg
2017-08-10 17:20
That is the real bug to fix.

wdennis
2017-08-10 17:22
Ah

zehicle
2017-08-10 23:53
@wdennis if you can give me a little summary of the iptables fix, I'll make sure it gets into the FAQ

lae
2017-08-11 00:50
@greg curious, what does the ce- prefix stand for on the templates in provision-content?

lae
2017-08-11 01:02
also I'm not really a fan of the inconsistent syntax within DRP/provision-content - like strings quoted in sledgehammer.yml but not in other bootenvs, and some variables using hyphens while others use underscores. minor issue, but it makes it confusing to figure out what style to stick to

greg
2017-08-11 01:10
Yeah. I'm on phone. Will need to type longer answer

wdennis
2017-08-11 13:29
@zehicle - Basically it's: iptables -F iptables -X iptables -t nat -F iptables -t nat -X iptables -t mangle -F iptables -t mangle -X (-F flushes the chain rules, -X removes all but default chains)

greg
2017-08-12 21:31
@lae ce- stands for community edition

greg
2017-08-12 21:33
With regard to the quotes and not, the underscores and hyphens, some is laziness, some is ruby-isms.

greg
2017-08-12 21:36
The original components operated in a dynamic ruby environment and hypen and underscores had to with class separators. It is not an issue now. These are independent from the DR content now.

greg
2017-08-12 21:37
A clean up pass and consistency pass should be done.

lae
2017-08-13 06:30
Ah I see

lae
2017-08-13 06:32
@greg I can submit a PR if you'd like - I did one pass regarding this for my company's copy already

greg
2017-08-13 15:06
okay - that would be great.

2017-08-14 18:12
Hi

zehicle
2017-08-14 18:13
hello

2017-08-14 18:15
Need a help.

2017-08-14 18:15
I am trying to install digital rebar

2017-08-14 18:15
and getting error as

2017-08-14 18:15
TASK [gem install kvm slaves] **************************************************************************************************************** ...................fatal: [149.56.217.200]: FAILED! => {"changed": true, "cmd": ["gem", "install", "json", "net-http-digest_auth"], "delta": "0:00:00.111974", "end": "2017-08-14 21:46:42.611851", "failed": true, "rc": 1, "start": "2017-08-14 21:46:42.499877", "stderr": "ERROR: Loading command: install (LoadError)\n cannot load such file -- zlib\nERROR: While executing gem ... (NameError)\n uninitialized constant Gem::Commands::InstallCommand", "stderr_lines": ["ERROR: Loading command: install (LoadError)", " cannot load such file -- zlib", "ERROR: While executing gem ... (NameError)", " uninitialized constant Gem::Commands::InstallCommand"], "stdout": "", "stdout_lines": []} to retry, use: --limit @/root/digitalrebar/deploy/digitalrebar.retry

2017-08-14 18:16
what would be the issue?

2017-08-14 18:22
Any help will be appreciated

2017-08-14 23:06
@ukris7_twitter could you provide more information about your environment and how you started to run the script?

2017-08-14 23:07
Also, we are STRONGLY recommending people start with Digital Rebar Provision in all use cases. This is much easier to get running initially and will help validate your environment before you bring up the more complex dockerized Digital Rebar environment.

lae
2017-08-15 16:20
@greg reading/editing BootParams is kind of a pain with the current multiline strings, what do you think about having each param on a newline?

lae
2017-08-15 16:20
e.g ``` BootParams: ' priority=critical console-tools/archs=at console-setup/charmap=UTF-8 console-keymaps-at/keymap=us popularity-contest/participate=false passwd/root-login=false keyboard-configuration/xkb-keymap=us netcfg/get_domain=unassigned-domain console-setup/ask_detect=false debian-installer/locale=en_US.utf8 console-setup/layoutcode=us keyboard-configuration/layoutcode=us netcfg/dhcp_timeout=120 netcfg/choose_interface=auto url={{.Machine.Url}}/seed netcfg/get_hostname={{.Machine.Name}} root=/dev/ram rw quiet -- {{if .ParamExists "kernel-console"}}{{.Param "kernel-console"}}{{end}} ' ```

greg
2017-08-15 16:49
I?m fine with it as long as it posts in. I mean need to change processing to remove the newlines. We may do that today, but I?d have to look

lae
2017-08-15 16:57
I don't think you need to make any changes to provision (I assume you meant s/mean/may/ there) - you already have newlines in some existing profiles in provision and it works fine - I also just tested this particular change and the pxelinux.cfg append line gets generated correctly

lae
2017-08-15 17:00
In other news, I opened https://github.com/digitalrebar/provision-content/pull/14 regarding style when you get a chance

greg
2017-08-15 17:03
okay - cool - I?lll look at it. :slightly_smiling_face:

2017-08-15 20:38
Hi

zehicle
2017-08-15 20:40
@lae are you interested in also adding style recommendations to the docs to match? that would help codify your changes

2017-08-15 20:40
@sherinsha hello

2017-08-15 20:40
Hi Rob

2017-08-15 20:41
TASK [Get Docker] ********************************************************************************************************** fatal: [149.56.217.200]: FAILED! => {"changed": false, "dest": "/tmp/docker.sh", "failed": true, "msg": "Request failed: <urlopen error [Errno 1] _ssl.c:492: error:14077410:SSL routines:SSL23_GET_SERVER_HELLO:sslv3 alert handshake failure>", "state": "absent", "url": "https://get.docker.com/"} to retry, use: --limit @/root/digitalrebar/deploy/digitalrebar.retry PLAY RECAP ***************************************************************************************************************** 149.56.217.200 : ok=9 changed=3 unreachable=0 failed=1

2017-08-15 20:41
I stuck at this stage while doing DigitalRebar installation

2017-08-15 20:43
anyone experienced this before

lae
2017-08-15 20:43
@zehicle sure, how do you want me to go about that? `docs/` folder with a `style.md` or just a quick `README`?

lae
2017-08-15 20:43
or a contributions doc

zehicle
2017-08-15 20:50
we need a doc/dev directory - I'll see about consolidating that. in the meantime, could you add doc/dev-style.rst?

zehicle
2017-08-15 20:51
the docs are autogenerated by read-the-docs, so all rst

2017-08-15 22:56
@sherinsha try installing Docker first, then run the script. something about your Docker install is failing

2017-08-15 22:57
okay

2017-08-15 22:57
is there any way to reset the rebar user password via SSH ?

2017-08-15 22:58
you can use the rebar CLI to do that on the system or remote

2017-08-15 22:59
do u remember the exact command for the same

2017-08-15 23:44
Thanks. Manged to reset the user password

2017-08-16 01:14
Digital Rebar install done. How can I deploy a bare metal server though it. All online documentations are confusing. Is there any simple step by step documentation?

zehicle
2017-08-16 02:00
you need to configure your admin network to be on the right NIC and have a host dhcp range. There are some videos for that.

2017-08-16 02:01
but couldnt find a straight forward documentation in any

zehicle
2017-08-16 02:02
HOWEVER, if your primary goal is to provision metal machines, we strongly recommend starting with Provision

2017-08-16 02:04
cd ~/digitalrebar/deploy workloads/add-provider.sh --admin-ip=IP --provider=Baremetal

2017-08-16 02:04
is this to create provider ?

zehicle
2017-08-16 02:05
you don't have to create the baremetal provider - it's there by default if you installed with con-provisioner & con-dhcp

lae
2017-08-16 02:05
Made an Arch Linux package for drpcli https://aur.archlinux.org/packages/drpcli

2017-08-16 02:06
@sherinsha really... start with Digital Rebar Provision. https://github.com/digitalrebar/provision

2017-08-16 02:06
okay

2017-08-16 02:07
it only takes 5 minutes to get running and will do all the metal provisioning much more easily. What use case are you trying to accomplish?

2017-08-16 02:07
okay ty

2017-08-16 02:08
once you have that working, it's possible to feed into Digital Rebar if you need features from that orchestration

2017-08-16 02:08
okay

2017-08-16 02:09
however, with v3.1 DRP coming out very soon (look at the master branch on tip) most of the primary workflows can be done in DRP

2017-08-16 02:09
cool

zehicle
2017-08-16 02:10
@lae sweet! thanks

lae
2017-08-16 02:14
Do you think we could get tarball releases for DRP? zip seems kind of unconventional

lae
2017-08-16 02:14
(in consideration for 3.1)

greg
2017-08-16 05:15
We could, but zip works on all the platforms and tar / bsdtar uncompresses them and is required to be installed.

2017-08-16 11:50
hello to everyone!

2017-08-16 11:52
could anyone please help me. I am testing digitalrebar with kvm virtual mcahines. When I boot a VM from network I get stuck on mounting stage2.img with the following error failed to mount xxxxx.squashfs as Stage2 initramfs

2017-08-16 11:52
and then i drop out into shell

greg
2017-08-16 13:52
Are you using DRP or DR?

greg
2017-08-16 13:54
Actually, for either case, make sure that the VM can reach the admin node. Check how many nics you are booting and make sure that you are getting the nic you expect. For example, vagrant adds an extra nic and that gets confusing.

2017-08-16 15:10
Sorry, redeployment of digital rear nodes Olvera the problem. It created a file stage2.img if a different size somehow and now everything works. Thank you for response.

2017-08-16 15:11
Sorry, redeployment of digital rebar nodes solved the problem.

2017-08-16 15:16
great! Glad to know that it's working

shane
2017-08-16 15:34
has joined #community201708

vishwanathj
2017-08-16 16:35
has joined #community201708

greg
2017-08-16 16:52
@lae - Your pull request worked in unit tess and are in. :slightly_smiling_face: Thanks! Now, to get DRP to use that repo. :slightly_smiling_face:

lae
2017-08-16 17:22
yay

lae
2017-08-16 17:26
@greg so - I noticed that in the bootenvs for Debian and Ubuntu in digitalrebar's assets (which carried over into DRP's ubuntu bootenv) set OS.Family to debian and ubuntu, respectively - to me this is kind of confusing since I think of them as both in the debian family (similar to RHEL/Centos/Fedora all being of the redhat family) - ansible sees it this way and uses a separate variable, `ansible_distribution` to distinguish the different distros in the same family. I also note that in `tools/install.sh` the same semantics that ansible uses are also used https://github.com/digitalrebar/provision-content/blob/master/tools/install.sh#L93

lae
2017-08-16 17:27
Does DRP use `OS.Family` itself or do only the templates use it?

greg
2017-08-16 17:28
It is mostly info. I don?t think it is used anywhere. It could be expanded in templates.

lae
2017-08-16 17:29
Ah okay - so yeah, I was thinking of expanding usage of OS.Family into templates - for example making it simple to have different partitioning templates while retaining the primary preseed/kickstart files - and that the partitioning templates would be identified by their layout rather than being a preseed or kickstart template - and within those partitioning templates we test for OS.Family

lae
2017-08-16 17:30
also actually, it is used in the preseed template right now to test whether or not we're installing Debian or Ubuntu

greg
2017-08-16 17:36
okay - I see what you are going to do, I think.

greg
2017-08-16 17:42
We haven?t tested this , but you could also do somethingl ike this:

greg
2017-08-16 17:43
{{$partlayout := .Param ?machine-part-config?}} {{template $partlayout .}}

greg
2017-08-16 17:43
where machine-part-config would be a string parameter that contains the name of a template that represents the partitioning layout you want.

lae
2017-08-16 17:44
I was actually in the midst of typing something of a similar fashion out but got distracted by work

lae
2017-08-16 17:44
but basically yeah

lae
2017-08-16 17:44
and then have a default part-config

greg
2017-08-16 17:45
This is the one area where the job/task system doesn?t help. The kickstart / preseed tweaks.

greg
2017-08-16 17:46
This would be a good addition to the community templates.

lae
2017-08-16 17:47
(this is actually something I used to do with Cobbler previously, ha - although I had separate centos/debian templates)

greg
2017-08-16 17:52
yeah - we wanted to make sure something like this was functional to make it easier to move over from cobbler.

2017-08-16 18:37
Is there a way to join already existing VM Linux "Node" thats riding on vmware?

greg
2017-08-16 19:04
to which system? DRP or DR. Yes. kinda.

2017-08-16 19:04
dr is what I installed

greg
2017-08-16 19:07
The quick short answer is not really in a run this command form.

2017-08-16 19:08
can use pxe instead?

greg
2017-08-16 19:11
So - one option would be to pxe boot the machine to sledgehammer. That will create the machine in DR. copy off the authorized_keys file in the root directory. Then put that in the root directory of your machine after you reboot back to its OS. Then you can mark the machine available and it will be operated against. That is kinda slow.

greg
2017-08-16 19:14
There is also the digitalrebar/deploy/scripts/join_rebar.sh - but it is old. It does mostly what needs to be done.

2017-08-16 19:14
yea I tried tweaking it a bit, the api has changed It looks like, I was close but wanted to find out another way,

2017-08-16 19:15
I'll try and pxe the vm and see if that works

2017-08-16 20:36
@hornjason what are you trying to make happen?

2017-08-16 20:37
im trying to spin up vm's inside vmware...to do a POC ... instead of metal. thats all

2017-08-16 20:37
then run install k8s/openstack

2017-08-16 20:38
that install has decayed - will not work out of the box at this point

2017-08-16 20:38
sorry to tell you

2017-08-16 20:39
np, at least I know now, I noticed you had a video using osx/virtual box. is this a usable option at this time

2017-08-16 20:39
for testing provisioning, yes.

2017-08-16 20:40
right ... let me give that a try and play with join_rebar see what I can do

2017-08-16 20:41
we're about to drop v3.1 of DR Provision - that will include a simple dynamic inventory that can install k8s via kubespray

2017-08-16 20:42
I'll upload the doc into master for DR Provision and post the link

2017-08-16 20:49
https://github.com/digitalrebar/provision/blob/master/doc/integrations/ansible.rst

2017-08-16 20:50
excellent, i just registered a already running vm like join_rebar.sh does. hopefully this will work as well

2017-08-16 20:50
thanks rob

2017-08-16 20:52
ahh so this will add a provisioner so you can deploy from rebar

greg
2017-08-18 18:48
: Hi all. We are in the process of changing how model validation and a few other things work internally. The master of the tree will have some issues for a little while. tip is works. stable is unchanged. This should be back shortly.

monkey
2017-08-18 18:52
has joined #community201708

2017-08-21 13:48
Hi all, I?m playing with Digital Rebar for the first time, it looks like an awesome tool! I?ve watched a few videos and now I?m following the guide http://provision.readthedocs.io/en/stable/doc/os-support/linuxkit.html I?m having an issue right out of the gate though, the first think to run is `curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/master/tools/install.sh | bash -s -- --isolated --rs-version=tip install` which gives me back `sudo ./dr-provision --static-ip=10.10.1.10 --file-root=/home/since/rebar-digital/lk-dr-trial/dr-provision/drp-data/tftpboot --data-root=drp-data/digitalrebar --local-store="" --default-store="" &` but when I run that I get back `unknown flag `local-store?` :worried:

greg
2017-08-21 14:10
Oh wait - just fixed that last night or Saturday night. I think. Should be local-content and default-content. May need to blow away dir and start over. Sorry. Tip has been in flux the last few days. Should be back together.

greg
2017-08-21 14:11
Of course if you tried the install stable and got this message that is a bug.

2017-08-21 14:16
@zehicle I just started clean, thought I?d try the most basic thing first, so I went to http://provision.readthedocs.io/en/stable/doc/quickstart.html and ran the first command shown `curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/master/tools/install.sh | bash -s -- --isolated install` whcih gave me back `sudo ./dr-provision --static-ip=10.10.1.10 --file-root=/home/since/drp-data/tftpboot --data-root=drp-data/digitalrebar --local-store="" --default-store="" &` and if I run that then I also get `unknown flag `local-store'`

2017-08-21 14:18
I?m guessing the version on the "Quick Start? is stable, so is this a bug?

zehicle
2017-08-21 14:18
we've changed some flags in tip

zehicle
2017-08-21 14:18
so the quick start script is straddling

zehicle
2017-08-21 14:19
the simplest solution is to `./dr-provision --help` to make sure the flags are right

2017-08-21 14:25
Thanks, looks like those flags have gone.

zehicle
2017-08-21 14:32
could you check ./dr-provision --version?

zehicle
2017-08-21 14:34
the quckstart should be using 3.0 by default. we're just about to release 3.1 from tip to stable, so there may be some accidental leakage.

greg
2017-08-21 14:40
It is a bug. Too aggressive on my flag littering.

greg
2017-08-21 14:40
Just don?t use local-store/local-content and default-store/default-content on 3.0 systems.

greg
2017-08-21 14:41
It is a bug in the install script.

greg
2017-08-21 14:42
Actually, instead of me fixing it. The stable tree and docs should use: ```curl -fsSl https://raw.githubusercontent.com/digitalrebar/provision/stable/tools/install.sh```

greg
2017-08-21 14:43

greg
2017-08-21 14:44
Hopefully, that will give you enough to deal with the issue:

greg
2017-08-21 14:44
Three things should work:

greg
2017-08-21 14:44
1. change initial curl to stable from master.

greg
2017-08-21 14:44
2. Change the initial comment to add ```--drp-version=tip``` - this would get the tip which is functionally equivalent.

greg
2017-08-21 14:45
to stable.

zehicle
2017-08-21 14:45
@greg should I update the quickstart to show stable & tip install choices?

greg
2017-08-21 14:45
Yeah.

greg
2017-08-21 14:45
We shouldn?t have the install script cross the ?streams? as it were.

greg
2017-08-21 14:46
we tried and it works for point releases ,but will get hairy as we go forward.

greg
2017-08-21 14:46
Since we can access the trees, it is better to have a consistent command line.

greg
2017-08-21 14:46
This is stable install: ```curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/stable/tools/install.sh | bash -s -- --isolated install```

greg
2017-08-21 14:47
This is tip install: ```curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/tip/tools/install.sh | bash -s -- --isolated install --drp-version=tip```

greg
2017-08-21 14:48
We should avoid master because it isn?t necessarily completely vetted like a tip is. I move tip slightly independently of master. This way straddling features and PRs can be brought together into tip.

zehicle
2017-08-21 14:50
makes sense. I like people make a specific choice and being clear about it

greg
2017-08-21 14:51
and then dropping --isolated install makes it ?production?

greg
2017-08-21 14:52
Also I didn?t want to fix it because the install.sh script now has different methods for getting content.

greg
2017-08-21 14:52
Tip uses the validated community content repo. stable pulls from assets.

2017-08-21 14:59
@zehicle Thank you! I thought I was being stupid, I?ll work thought it like you said.

greg
2017-08-21 15:08
We?ve got to get a better bot. :slightly_smiling_face: Or get everyone into slack. @Simon-Ince - you?ve been talking to two of us. :slightly_smiling_face:

2017-08-21 15:11
Thank you everybody!

greg
2017-08-21 15:13
More for awareness than praise. @zehicle and @greg will get peoples attention.

2017-08-21 15:30
@Simon-Ince happy to talk 1x1 and also pull you over to the Slack channel too

2017-08-22 14:53
Hi there, can anyone help me with a weird issue we are having running dr-provision. It looks like it's only binding to IPv6 ports, and leaving the v4 address alone.


2017-08-22 14:54
I'm getting something that looks like ^^

2017-08-22 14:56
can you provide the start-up parameters and the type of system you are running on? you can use ``./drpcli info get`` to recover the info

2017-08-22 14:57
it would be helpful to know if you specified a static ip too

2017-08-22 14:58
Thanks @zehicle , The startup command I'm running is `./dr-provision --static-ip=10.10.1.10 --file-root=/home/dave/digital-rebar/drp-data/tftpboot --data-root=drp-data/digitalrebar`

2017-08-22 14:59
Nothing fancy, just want it to bind to ip4 address and access the UI

greg
2017-08-22 14:59
That looks normal. The kernel will bind to both by default And it looks like that

2017-08-22 15:00
I thought `8091` and `8092` were just on v6 addresses?

2017-08-22 15:01
(Unless I'm reading this completely wrong, which is quite possible)

2017-08-22 15:36
@zehicle All sorted.. gumble grumble firewall.. Thanks for your help anyway :-)

2017-08-22 15:37
if you can share the troubleshooting steps, I'll put it into the docs

greg
2017-08-22 15:52
:::8091 is unspecified IP (both). Great that you?ve worked it out.

2017-08-22 17:04
Hey Folks - I'm having a heck of a time getting DR up and running. Tried the quickstart but DHCP provisioning wasn't enabled and trying to enable it would fail with ansible errors. Tried to follow the docs, and it all seems to start but I cannot login to the control panel. I feel like I am missing something obvious.

zehicle
2017-08-22 17:06
Dr or Dr provision?

zehicle
2017-08-22 17:08
If you are using ansible, then DRv2. Check out the DR P install. Its much lighter.

zehicle
2017-08-22 17:10
For DR v2, check the docker log, it may not be all the way started if you cannot login

2017-08-22 17:12
Are there docs for DR P?


2017-08-22 17:16
Thanks!

zehicle
2017-08-22 17:16
for boot provision, the provision is much simpler, smaller and easier to understand

zehicle
2017-08-22 17:17
with some of the new additions (tasks, plugins, content) , we think it covers all of the key use cases

2017-08-22 17:19
:+1:

2017-08-22 17:28
setting up a lab cluster of 10 Intel NUCs... doesn't seem to be able to get through pxe with `dr-provision: dr-provision2017/08/22 17:16:14.566078 sending block 0: code=0, error: TFTP Aborted`

2017-08-22 17:29
trying to debug it, but I can't seem to change the preferences in the UI

2017-08-22 17:30
two questions... where are the logs stored and how do I persist the debug levels in preferences?

2017-08-22 17:31
running tip v3.0.4-tip-235-9f07bd96cefa80f66b9d10062f50f1048a7cf8ff

2017-08-22 17:43
we're working on UX - CLI is safer right now to set things, especially for tip

2017-08-22 17:44
are you running this in a container?

vlowther
2017-08-22 17:45
We log to stderr by default -- curerntly, we don't save logs to disk by ourselves.

2017-08-22 17:45
no... just install.sh without isolated

2017-08-22 17:46
systemd

vlowther
2017-08-22 17:46
We rely on stderr being handled appropriately by your init or container system for persistent logging right now.

2017-08-22 17:48
cool... I'll set debug using drpcli

2017-08-23 17:01
Does version 3 only use linux kit to deploy k8s ? Is the concept of deployments that was in V2 not available in v3?

2017-08-23 17:10
Yes, the roadmapping of 2 vs. 3 seems completely divergent.

2017-08-23 17:29
Both use Kubespray Ansible. For v3, there's no additional wrappers at this point


zehicle
2017-08-23 17:30
the v3 roadmap started with the Provision/DHCP services being split into stand alone things.

zehicle
2017-08-23 17:31
with the ability to register machines back into v2 using the API driven calls (which is what is what v2 does too)

zehicle
2017-08-23 17:32
the plan for v3 is that DRP becomes the new metal provider to be on part w/ the v2 cloud providers.

zehicle
2017-08-23 17:32
early feedback from v3 DRP is that it's sufficient for most uses cases people wanted

zehicle
2017-08-23 17:33
and using the ansible / terraform tools directly against DRP seems to be a more easily understood way to consume them

2017-08-23 17:34
@tpagden @hornjason I hope that helps

2017-08-23 17:45
@hornjason to answer the specific question: DR P (the current delivered part of v3) has no deployment concept. It does have profiles that can be used to provide grouping mechanisms for machines.

2017-08-23 17:48
@zehicle I apologize, I'm not keeping up on a few things then. So the v2 app catalog, backed with ansible - that's no long considered to eventually be in v3? I have the DRP v3 running, however, compared to the v2 dashboard... maybe there's some step I'm missing to get back to the functionality that v2 seems to offer.

2017-08-23 18:07
we're working on a new dashboard for v3.1 release - private beta starting very soon of it

2017-08-23 18:08
but some of the features of v2 will be slower to emerge - we've found that people were having trouble getting the whole stack working and wanted something simpler to get started

2017-08-23 19:16
Gotcha, thanks for the response. If I continued to use v2, is there some form of migration or transfer of stewartship to v3 should v3.x emerge to have all the functions?

2017-08-24 05:06
it's time to start having bi-weekly meetings to discuss roadmap and design.

2017-08-24 05:10
@tpagden we have several plans for v2 -> v3 migration and integration. it really depends on which features you are depending on. Some of the annealer uses cases (especially for physical workflows) are much simpler in the DR P task systems. I'd be happy to talk here or 1x1 about migration plans - it really depends on which features of v2 you are implementing.

wdennis
2017-08-25 14:00
@zehicle @greg I think publishing roadmap/design would be great - would like to know futures of DRP and DR, and what the proposed use-cases are for each. Have always thought that since DRP used Sledgehammer to initially boot new nodes, it could be used to create an inventory of machines and their hardware/firmware (more useful for those of us using self-owned bare metal.) And after OS installation, have the ability to run an Ansible playbook against the node(s). Or will stuff like this be reserved for "full-on" DR? On a non-technical note, hope all you folks in TX remain safe with the coming hurricane/storms, and recover quickly!

zehicle
2017-08-25 15:31
@wdennis we DO have an Ansible dynamic inventory generator - checkout out /integrations/ansible and there's even docs for it

zehicle
2017-08-25 15:31
we'll get a 3.1 release notes and then start community meetings back up to draft a roadmap

zehicle
2017-08-25 15:32
(there's a Terraform provider in the works too... that will require 3.1 workflow capabilities).

wdennis
2017-08-25 16:50
@zehicle Thanks, looking fwd to the docs


zehicle
2017-08-26 16:57
note on using VirtualBox - the vmboxnet0 is not created in the system until a machine using it is booted

zehicle
2017-08-26 17:49
ALL - we're getting ready for the v3.1 release from master/tip. if you have time to test & validate functionality, please take some time to play with it. Especially if you have v3.0 bootenvs and templates

2017-08-29 01:54
Please help me updating bootenv. I am setting up one for windows server. There is a bootenv for this os in default deployment of DR. I am trying to add missing files, wimboot file in particular. So I have copied the file into windows-2012r2/install directory, but it is still not found by the provisioner. What actions should be taken after adding missing files for bootenv. Thank you.

greg
2017-08-29 02:56
umm - that is about to be removed because we can?t really support it right now. We haven?t built or tested it for a long time. It also requires you to have built wimboot images and the like.

greg
2017-08-29 02:56
I really need to put out the next release because it isn?t really supported.

2017-08-29 18:29
@lion_kg_twitter Windows support is complex (especially in v2) and we're pulling it out of community resources because it's not something that can be community supported at this time. RackN does provide support for Windows on DRP for customers where we can work 1x1 with people around their specific environments.

2017-08-29 18:32
@lion_kg_twitter reading your question another way... are you just asking about DRP file upload? You should be able to use the CLI files upload and then see the file and then see the file in the CLI files list

lae
2017-08-29 20:55
@zehicle I think there should maybe be a note, or upgrade task really, for migrating content from `/var/lib/dr-provision/` to `/var/lib/dr-provision/digitalrebar`

greg
2017-08-29 21:05
The old path should have worked still with the old options.

greg
2017-08-29 21:05
hmm - sigh. I?ll check

lae
2017-08-30 02:06
Is it intentional for `bootenvs install` to install all templates in the templates directory, despite the bootenv being installed not referencing them?

greg
2017-08-30 02:20
Yes. Templates can ref templates and we don't do on demand loading

zehicle
2017-08-30 14:31
new v3.1 content API allows for multiple items to be bundled together (and also makes them read only so they can be upgraded)

greg
2017-08-30 20:43
@lae - do you think this would be reasonable for a change to install.sh to address the production directory change? ``` if [[ ! -e /var/lib/dr-provision/digitalrebar && -e /var/lib/dr-provision ]] ; then sudo mkdir -p /var/lib/dr-provision/digitalrebar sudo mv /var/lib/dr-provision/* /var/lib/dr-provision/digitalrebar fi ```

lae
2017-08-30 21:21
I suppose - although I'm using an Ansible role myself for deploying DRP

lae
2017-08-30 21:21
I'm probably the only one, though

lae
2017-08-30 21:22
that `mv` command will try to move `digitalrebar` into itself, which I believe errors out

lae
2017-08-30 21:23
``` [musee@birdy tmp]$ mv hello/ hello/ mv: 'hello/' ?????????????? 'hello/hello' ???????? [musee@birdy tmp]$ echo $? 1 ```

lae
2017-08-30 21:23
yeah

lae
2017-08-30 21:25
you could rename the existing directory (`mktemp`), then recreate `/var/lib/dr-provision/`, then move the renamed directory into the newly created directory as `digitalrebar`

greg
2017-08-30 21:26
@lae - you are pulling the zip and doing it all yourself?


lae
2017-08-30 21:27
effectively

lae
2017-08-30 21:29
does an install and some configuration afterwards, like removing the default user (if specified) and creating our own admin user

greg
2017-08-30 21:39
nice!

vlowther
2017-08-30 21:43
sweet.

lae
2017-08-30 21:49
if wanted I can clean this up a bit and publish it but

lae
2017-08-30 21:51
I'd like to not use `get_url` and `unarchive` to get around making downloading/extracting the DRP release appear idempotent to ansible (there's a bug with regards to using `unzip` in the `unarchive` module)

lae
2017-08-30 21:54

lae
2017-08-30 21:56
ah nah it was this that I had read previously https://github.com/ansible/ansible/pull/24580#issuecomment-302592533

lae
2017-08-30 21:56
(so preferably I'd want releases to be tarballs)

greg
2017-08-30 22:04
that is why we use bsdtar. It process zips and does the right thing, usually.

zehicle
2017-08-30 22:42
@lae would you be interested in contibuting this to the project?

lae
2017-08-30 22:46
I have already :sweat_smile:

lae
2017-08-30 22:47
(just a PR to provision-content so far)

zehicle
2017-08-30 22:48
you're a rock star! thanks