How do I get the static url for an iso?

in a template?

I'm having to do some non standard things to install this iso

so i'm booting into sledgehammer and them i'm going to dd it

but I can't figure out how to download the iso

@hadees - what version of the product are you using ?

are you trying to pull the ISO down from the provisioning server to the target server ?

if you are trying to pull down ISOs from your DRP server - it looks like you should be able to grab them from a similar URL: http://<drp_server>:8091/isos/ubuntu-16.04.3-server-amd64.iso

http://<drp_server>:8091/isos/sledgehammer-b3c09ebd5a9c228c66d8a617b6f5d10ccbe1c273.tar

@hadees did you want to DD the image or is that a workaround? we've been working on some advanced integrations that use DD from Packer or similar instead of kickstart.

@zehicle I ended up just uploading it as a file

that worked out better anyway because it was compressed

i kept running out of room if i had the whole iso

although streaming the iso for dd would be awesome

uploading it as file is a good idea if you just want to download it to the machines. the ISO upload is designed for the system to manipulate into boot images, not for direct retrieval

@shane - is correct. There are some additional things that are available too. If you are doing work in a template, you can do `{{ .ProvisionerURL }}` Which dynamically expands to `http://<drp_server>:<drp_webport>`. If you are running in sledgehammer, the cmdline for the kernel also has the provisioner Url. That is if you don?t want to hardcode things.

@greg Having an issue with my DRP install - was working fine, but then after rebooting the DRP admin node, I set a target machine for the desired bootenv, but when PXE-booted, TFTP is failing...

It's a 3.0.5 installation btw

@wdennis - does your pxe target get a DHCP addr from the DRP admin node ?

It actually gets one from the subnet router, which then hands it next-server

did you set dr-provision service to start on boot: ```sudo systemctl daemon-reload && sudo systemctl enable dr-provision```

Next-server is the DRP admin nodes IP, and file name is 'lpxelinux.0'

Yes, DRP is started and running

if the `dr-provision` service is running - it's easy to test the TFTP

from a host outside of the drp server - simply do: ```tftp <drp_server> tftp> verbose Verbose mode on. tftp> get lpxelinux.0 getting from drp_server:lpxelinux.0 to lpxelinux.0 [netascii] Received 92766 bytes in 0.1 seconds [5926590 bit/s] tftp> quit```

you should be able to snag that file (assuming now FW or iptables/etc entries interfering

if that is working - I'd suggest chucking `tcpdump` on your network interface and listen for incoming requests to make sure your next-boot hand off is working, and your pxe client is actually able to route to the drp server ```sudo tcpdump -i <network_interface_name> port 67 or port 68 or port 69```

@shane Yup, damn 'iptables' rules...

Had full DR on this node prior, that had Docker and spawned the wealth of iptables rules...

:slightly_smiling_face:

Bites me every time I (infrequently) reboot the DRP server...

what distro you using ?

I use an "iptables-restore" tool in Ubuntu to save/restore rules on reboots ... if I'm not using UFW

For DRP admin node, CentOS 7.3.1611

looks like centos has iptables-restore in yum repo too

older doc - but still relevant: https://www.centos.org/docs/5/html/5.1/Deployment_Guide/s1-iptables-saving.html

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F70PFCYNS/image_uploaded_from_ios.jpg

Hmmm, something still awry- now I have no machines showing up in inventory!!

is it set to use a different bootenv than local ? ```drpcli machines show <UUID> | grep BootEnv "BootEnv": "local",```

Ah, wrong value used for DRP "--data-root="

are you running in "production" mode ?

in production mode - data-root is in `/var/lib/dr-provision`

Oh hell no :)

(meaning the installer mode - not really in "production") :slightly_smiling_face:

Kicking DRP off by hand

if you install with the ```?isolated``` flag to the installer, it drops everything in your local users home directory

other wise it does "production" mode - which is the ```/var/lib/dr-provision``` and ```/var/lib/tftpboot``` directories (Ubuntu-land)

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F6Z7NH96C/image_uploaded_from_ios.jpg

Yeah I did --isolated

All is working now :sweat_smile:

fabulous !

@shane Are you a part of the RackN gang, or a user?

I'm the FNG at rackn

LOL

I'm still the FNG here at my job, 11 years in :stuck_out_tongue_winking_eye:

Well, nice to virtually meet you :)

You too!

Was able to meet Greg, Rob and Victor at DevOpsDays Austin this year

In person :)

I'm sorry you had to go through that ...

They were nice enough to this poor old chap

:slightly_smiling_face:

How close are we to 3.1 goodness?

...being the FNG... I'll I'll need to check w/ the rest of the gang - give me a few mins

(FNG == day 3 :slightly_smiling_face: )

Threw you right in there, huh

nothing like trial by fire ... I'm used to it - I've known the team for a long time - so they didn't cut me any slack

Looks like they put you on Slack instead of cutting you any... <drumroll /><cymbal />

doh!

Be here all week...

@wdennis - the unofficial answer is "about 2 weeks-ish time frame" ...

:slightly_smiling_face:

Lets plan a release meeting next week to review in open

I would like to create a 5 node cluster like: http://node.mu/2016/12/19/5-node-nano-itx-kubernetes-tower/ but also test DigitalRebar on a 6th node - what motherboards can you recommend ?

Hey @maymann - there aren't any real requirements for a mobo for a Digital.Rebar provisioning node. Pretty much anything that can support a Linux distribution will work - if it'll run Ubuntu/CentOS ... etc ... you'll be just fine

obviously you'll need a network interface for the OS and imaging activities ...

with that in mind - any nano/mini/ITX mobo capable of running Linux and supporting a NIC - will work for you

Our pre-built binaries do have a somewhat limited support for processor architecture that we compile for - but since our binaries are implemented in Go Lang - you should be able to compile if you end up using a hardware architecture we don't currently pre-compile for.

currently compiled binary support is as follows: ``` bin/linux/amd64/ bin/linux/386/ bin/darwin/amd64/ bin/darwin/386/ bin/windows/amd64/ bin/windows/386/``` Notes: "darwin" being a Mac OS X platform - if you're not familiar with that designation "386" being a 32 bit build and "amd64" being a 64 bit build

Hi all - just started playing with DRP and found the Ubuntu bootenvs file is outdated. On my harddrive: assets/bootenvs/ubuntu-16.04.yml -- iso doesn't exist on the mirror - 16.04.2 needs to be updated to 16.04.3 in the yml file. I'd push a patch myself but it would probably take me an hour (or more) to figure out how :(

Hi @aimee - welcome !

we have updated the repo with that change, but it hasn't yet been pushed in to a newly cut version yet

zehicle: thanks! I noticed the SHA value has to be updated as well. Also for Debian the SHA is listed as BAD in the UI

Aimee - to short cut around the problem - you can update your current assets bootenv file with the following: ``` IsoFile: "ubuntu-16.04.3-server-amd64.iso" IsoSha256: "a06cd926f5855d4f21fb4bc9978a35312f815fbda0d0ef7fdc846861f4fc4600" IsoUrl: "http://mirrors.kernel.org/ubuntu-releases/16.04/ubuntu-16.04.3-server-amd64.iso"```

this change has been made in `master`, though - so the next release will have it

Thanks zehicle. I still have some coding know-how left... haven't completely gone over to the sysadmin/management dark side. (LOL). I have to say I'm liking DRP a lot - at least I understand more of it compared to MAAS (blech), Foreman, RackHD. I'm looking forward to using it more and maybe even contributing if I can dust off my hacking skills.

awesome! we look forward to helping you along your journey ... and look forward to any contributions back to DRP you can help with :slightly_smiling_face:

And I like the complement. Our goal has been to be simple

Thanks zehicle - it will be an interesting journey - hoping to use DRP along with the Moby Project's Infrakit to manage/provision my cloud native POC lab.

Also as a reminder. We are generally slack based but have a bot that relays from IRC and gitter to slack and back. The user IRC and gitter see talking is zehicle, but that is a munging. Like this is @shane and @greg right now. :slightly_smiling_face:

I was just going to ask if you were on slack - noticed the relay. I don't have slack installed on this machine - gee I'd have to roll across my office to use slack...

We can hook you up if you want slack access

Hopped onto Gitter but yes, please, hook me up with Slack

send an email - with your email. :slightly_smiling_face:

Awesome - thanks gentlemen!

has joined #community201709

@aimeeu ... welcome ... :slightly_smiling_face:

Thanks @shane

has joined #community201709

I've connected *1 new room* _#community1 (RackN)_ on Slack. See https://sameroom.io/maps/eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE1MDQ5OTgwMDAsInJvb21faWQiOjI1ODYyNjksImlkZW50aXR5X2lkIjoxOTU3NzJ9.itIpOyu50yC-lispTjJXy4HUxf4LxMHYfN_-9EN7DTQ

I've connected *1 new room* _digitalrebar/core_ on Gitter. See https://sameroom.io/maps/eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE1MDUwNjY0MDAsInJvb21faWQiOjI1ODYyNjksImlkZW50aXR5X2lkIjoxOTU3NzJ9.Y9bMe0x7GdDlOCuHRTyRY9Vk_afqqB2YLLyxrYtWWmw

testing new Eng user from Gitter

seems to go to both #community and #community1

working on it

I've got the connection from Gitter to Slack working

checking from alias

woot!

Slack to IRC is not working yet...

boo

mostly because I messed up the auth... soon

I have a few nube questions. I have been reading the docs and am excited to use Digital Rebar. Its nice! Question. If I wanted to bind user access to say a local LDAP or AD using SAML or just LDAP is there a way to do that?

hey cbitter78_twitter -welcome !

Thanks

What is your use case for the LDAP/AD auth - are you looking to build a self-service solution for provisioning - or are you looking to control auth to multiple provisioning endpoints ?

More of the self service model Where in I can have a set of inventory in a data center and then via access / roles / etc and perhaps bind those to AD users / groups

Sure - makes sense.

so I can give the Dev group access to racks 2 - 3 and prod racks 4 - 10

Presumably you're looking at the Digital Rebar Provisioning (DRP) solution right now ?

yes.

The current release version is 3.0.5 - have you played with it yet - and if so, what version ?

Am am doing read the docs discovery right now. ;)

we're just about to release 3.1 in to the community (probably another 2 weeks) ...

the LDAP/AD auth solution is a current near term road map feature - but that will be a RackN enterprise feature

Cool.

another nube question. RackN Enterprise is Digital Rebar Paid?

we definitely have it on the roadmap, and it's a frequently asked feature both for Self Service and mult-endpoint management

Is there a better forum to ask questions about RackN?

do you use Slack ? we can give you a slack invite, and we can drop to a D.M.

I do use Slack

you can send the invite to cbitter78@aol.com

I've connected *1 new room* _#digitalrebar (irc.freenode.net)_ on IRC. See https://sameroom.io/maps/eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE1MDUyMjg0MDAsInJvb21faWQiOjI1ODYyNjksImlkZW50aXR5X2lkIjoyMDI5MTJ9.GTm3UATkvTUPKo1P-8QVH4nRl2NpWsichPgwxT96w-Q

testing IRC to Slack and Gitter

@cbitter78_twitter - invite sent, sorry for the delay ...

has joined #community201709

there's a native Slack IRC gateway, if an admin enables it for this server

(^ followup to discussion with @aimeeu on friday)

morning, I have a quick question about provisioning, does Rebar support Cloud-Init?

hey Telmo - what are you trying to do with cloud-init ? what's the use case ?

we do have some current support for cloud-init, which is going to emerge in a release or two - at the moment it's not very well productized, but we're getting there

we're hoping you can help out with your use case - to help guide where we go with it

we are looking into provisioning alternatives and I am working on the market analysis for product capabilities. The use case is to have "Code as Infrastructure" that is portable across solutions. Having the Cloud-Init scripts in a CVS repository where host or hosts groups specific configurations that can be accessed by physical provisioning and "Cloud Provisioning" so hosts/instances build are as identical as possible

@Telmo, what other tools are you using? Ansible? Terraform? Chef?

@zehicle uploaded a file: https://rackn.slack.com/files/U02DHRR2L/F71LXURUM/img_20170911_210555.jpg

10 please

I'll have them at Hashicorp Fest next week if anyone is in town!

We're looking for people who want to help beta test our Terraform to Metal provider

Hello! How do I update DR without loosing all current settings and nodes? Thank you.

Hey lion_kg_twitter - welcome

what version are you currently running - and what version are you looking to upgrade to ?

shane, where can I see current version?

lion_kg_twitter - if you are running Digital Rebar Provision - then you can get the version via `drpcli version` - if you have a Digital Rebar v2 server - you'll have a "rebar" command - which would be `rebar version`. Your binaries may be installed in different locations - depending on how you did the original install.

just starting out..do i need both a V2 and a V3 (DPR) servers? or could i just start out with v3?

@smartekb - highly recommend you ONLY use DRPv3 server ...

we are very close to releasing the DRPv3 ver 3.1 release - likely next week - right now - the current released version is 3.0.5

thx Shane. I'm able to pxe boot a VM using sledgehammer...fter the VM boots up, have not been able to get it to provision an OS (after setting its bootenv to CentOS)

did you load the ISO ? you need to run `drpcli bootenvs uploadiso <iso_name>`

by default DRP is distributed without any "content" - you need to initialize/add content to be able to deploy it

which version are you running? (`drpcli version`)

i have the iso..it shows all green in the ux. I'm running version 3.05

i'll re-run that command ago, but i seem to have issues with the uploadiso command

can you provide the command/output of your uploadiso attempt ?

in 3.0.5 - you should have an "assets" directory and a "tools" directory in the base working directory when you did the install

i do have those

if you do: ```cd assets ../tools/discovery-load.sh ``` that is a helper script (needs to run from the `assets` directory) that will load content using the YAML definition files in the `assets/bootenvs` directory

you can also JUST load a specific piece of content via: `drpcli bootenvs install bootenvs/centos-7.3.1611.yml` for example

that `tools/discovery-load.sh` is just a helper - which only runs that one command for different content

once the bootenvs load is complete - you should see it in the green UI

once you have content loaded ... you need to: * *you must* also add the Discovery and Sledgehammer bootenvs - they're required * make sure you have a Subnet defined (via UI "Subnets" or "drpcli subnets" commands * boot a machine w/ PXE enabled that hits the DRP endpoint (your provisioning server) * then assign a "bootenv" to the machine * reboot your server - and it should kick the PXE provisioning process for with the given

Version: v3.0.5-0-2b326b01a5ef733f3fe599cac2c7aaa6e914b17f

yep - the above process should be good for your version

one note - there is a small bug in the content - for Ubuntu 16.04 - it tries to get 16.04.2 which is replaced by 16.04.3 in the repos

centos will work fine

`root@cn-ddev03:/opt/drp/assets# ../tools/discovery-load.sh No assets directory to work from. root@cn-ddev03:/opt/drp/assets# ls bootenvs isos profiles startup templates root@cn-ddev03:/opt/drp/assets# ls bootenvs/ centos-6.8.yml debian-8.yml esxi-6u2.yml lk-sshd.yml redhat-7.0.yml ubuntu-14.04.yml centos-7.3.1611.yml discovery.yml lk-k8s-master.yml local.yml scientificlinux-6.8.yml ubuntu-16.04.yml debian-7.yml esxi-650a.yml lk-k8s-node.yml redhat-6.5.yml sledgehammer.yml windows-2012r2.yml`

that was terrible..lemme clean that up

oops .... sorry, run that from the parent directory of "assets" ... :slightly_smiling_face: my fault

`/opt/drp/assets# ../tools/discovery-load.sh No assets directory to work from.`

is that discovery load working for you now ?

it did..with a couple errors about local not being available..i'm going to the ux to find out

root@cn-ddev03:/opt/drp# drpcli bootenvs uploadisos centos-7.3.1611 Access CLI commands relating to bootenvs Usage: drpcli bootenvs [command]

seems there is no upoadisos command in the binary?

if you don't have the ISO local to your `drpcli` command - use the `drpcli bootenvs install /opt/drp/assets/bootenvs/centos-7.3.1611.yml`

if you have the ISO local - do: `drpcli isos upload <ISO_FILE> as centos-7.3.1611`

context deadline exceeded ..does that mean a timeout?

Yes, timeout with download/upload

Is your internet bandwidth constricted?

If so, you might download the iso separately, then, use the "drpcli isos upload...." command

i have the ubuntu 16.04 iso downloaded locally..can i use the isos upload command to bypass the bug?

Yes, use the "isos" command

In a short bit, I can provide you the yaml with the updated content for Ubuntu - at lunch now....

cool..thx

`root@cn-ddev03:/opt/drp/assets# drpcli bootenvs install bootenvs/centos-7.3.1611.yml 2017/09/14 14:30:21 Installing bootenv centos-7.3.1611-install 2017/09/14 14:30:21 Uploading isos/CentOS-7-x86_64-Minimal-1611.iso to DigitalRebar Provision Error: Failed to fetch bootenv: centos-7.3.1611-install: context deadline exceeded `

timing out..even on a local iso upload

Is your drpcli command running on your endpoint you are uploading content to?

yea

What is the OS distro/version of your server

ubuntu 16.04

Can you provide filesystem space info (df - hl)

...and did you install with "--isolated" mode - or production mode (i.e. not specifying the "--isolated" flag)?

production mode

/dev/mapper/cn--ddev03--vg-root 41G 25G 15G 64% /

smartekb - 2 follow up questions: 1. what type of drives is your DRP install (default location would be /var/lib/dr-provision - if you didn't relocate it) 2. can you provide the output from this quick DD command? `time sh -c "dd if=/dev/zero of=ddfile bs=8k count=250000 && sync"; rm ddfile` (running from the same directory/filesystem as your DRP install location (eg /var/lib/dr-provision)

root@cn-ddev03:/var/lib/dr-provision# time sh -c "dd if=/dev/zero of=ddfile bs=8k count=250000 && sync"; rm ddfile 250000+0 records in 250000+0 records out 2048000000 bytes (2.0 GB, 1.9 GiB) copied, 7.10381 s, 288 MB/s real 0m14.448s user 0m0.068s sys 0m6.828s

We believe you _might_ be hitting a timeout value in the drpcli client - if you're up for testing our latest version (not "stable") - then this issue very well might go away ....

if you install again from "tip" - with production mode - then it'll update the existing version .... the only thing you might need to do is re-enable any subnet(s) you might have created previously

there is a new version feature that lets you "enable"/"disable" subnets, and the new "tip" version sets a subnet to "disable"

thats what I did earlier today :smile:

i'l do again

can you please run "drpcli version" ? you showed me version 3.0.5 previously

the "tip" will output "3.0.4-tip-...." number - which _looks_ older than 3.0.5 - but it's not :slightly_smiling_face:

Version: v3.0.5-0-2b326b01a5ef733f3fe599cac2c7aaa6e914b17f

right - that's older than "tip" version - the changes in "tip" should have a fix in it for you

but if you do "drpcli version" with "tip" - it'll show a version string of "3.0.4-tip .... " though it looks older - it's actually a newer version :slightly_smiling_face:

how do i do this update? wat ive been doing is not working..i'm still showing 3.05

Change the curl url to "tip", instead of "stable"

and add --drp-version=tip

```curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/tip/tools/install.sh | bash -s -- --isolated install --drp-version=tip```

except ... don't use "--isolated" since you're doing a "production" install :slightly_smiling_face:

:slightly_smiling_face: oops.. Thanks

Okay better version: ```curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/tip/tools/install.sh | bash -s -- --drp-version=tip install```

Greetings folks - can I get some details on how to build an OS install template? It's not super clear to me with the docs for DRP and if it helps I just need CentOS 7

@edolnx - if you are fine with starting with a stock "minimal" install - you can use the "ce-centos" template from the community content repo

That should be fine. Thanks!

I can at least look at it to learn/modify as I go

are you running 3.0.5 ? (`drpcli version`)

Stable, yeah `Version: v3.0.5-0-2b326b01a5ef733f3fe599cac2c7aaa6e914b17f`

did you do the "discovery-load.sh" - or insure you installed the Sledgehammer an Discovery bootenvs already ?

Yes, did the discovery-load

cool - did you do an "isolated" install mode or "production" ?

I did not do isolated, so production?

:slightly_smiling_face: yep

I've also had to modify the systemd script to set the IP address and certs correctly

ok - cool

did you're content end up installed in /var/lib/tftpboot ?

I think so, my target node booted is was discovered

`drpcli bootenvs install assets/bootenvs/centos-7.3.1611.yml` (Need to run that from the base directory you did the original install in - or give the fully qualified path to the "assets" directory)

if you have a Centos 7 minimal ISO file on hand already - you can use the "isos" command to just upload the ISO - without doing the download again

`drpcli isos upload <ISO_FILE> as centos-7.3.1611`

`[kumulus@koiab dr-provision-install]$ drpcli bootenvs install assets/bootenvs/centos-7.3.1611.yml Error: Error determining whether bootenvs dir exists: stat bootenvs: no such file or directory`

sorry edolnx - you need to run that command from the "assets" directory - so: ```cd <somewhere>/assets drpcli bootenvs install bootenvs/centos-7.3.1611.yml```

That looks better!

woot !!

i must say, the new interface UI is pretty

but does it really phone home?

thx! our UI guys will be happy to hear - just a note ... that UI is ... definitely ... still in "Alpha" stage - there are lots of sharp edges that can cut you

i'm bleeding all over already

definitely don't recommend relying on it yet :slightly_smiling_face:

but the url bar appears as if we are hitting rackN..that's not true, is it?

the UI is running as a SaaS/Portal in our cloud environment

the DRP endpoint (your provisioning server/service) does NOT reach out to the SaaS portal - we do not phone home from DRP endpoint

just looks like?

however - by accessing the SaaS/Portal from your browser, we can transfer content from the SaaS/Portal to your DRP endpoint

the Saas/Portal is hosted on our side

since this isn't released yet - we haven't published the docs and pretty pictures that help explain the flow

(note my edit: "by accessing the SaaS/Portal from your browser")

access and use of the Saas/Portal goes like this: * user hits DRP Endpoint UI * redirected to RackN SaaS/Portal * DRP endpoint never reaches _out_ to the SaaS/Portal * content transfers happen as a PULL from the SaaS/Portal to your browser * content is pushed from your browser to DRP endpoint at your request/authorization * that's how content is updated via transfer from the SaaS/Portal to your DRP endpoint

does that help to make a little more sense on the flow and access ?

the redirection happens by itself

?

well .. you have to open up the URL and point to your DRP Endpoint ... (for example if it's running locally on your laptop: https://127.0.0.1:8092 ) - then, it redirects to the (not official resting place) https://rackn.github.io/provision-ux/ location

@rackneng so, what are the solutions that will be available for organizations who need a 100% on-prem solution and want a pretty UI?

The UI can be served from the DRP instance itself and can be packaged for that. Since we expect high churn in the short term, we are starting with a web accessed solution.

The only current requirement is that the browser being used to manage DRP has access to the internet (web proxies work), but we understand the potential need for completely internalized solution.

thanks, will keep that in mind as we begin/continue our evaluation

so, is it possible to have it NOT redirect?

don't go to the UI url :slightly_smiling_face:

ha!

the 3.0.5 "green ui" that used to be there is deprecated - the new UI will be "the way forward" ... in the future there may be an option to provide the new UI as fully embedded in the DRP endpoint side - and not a SaaS/Portal hosted by rackn - however, that's not the case for now

if the community wishes to build/maintain a UI for the DRP endpoint - they certainly are welcome to

we'd happily welcome any commits in that respect

I'd add that we believe strongly in decoupling the UX from the DRP service & CLI. The design for the service is to keep it very small, stable and focused. We are seeing use cases where 1000s of DRP endpoints are deployed for embedded edge or top of rack infrastructure. Since the service is API driven, there's no need to bundle UI code into the service.

Can I get an invite to the slack channel?

Hey @stanchan - you can sign up at https://www.rackn.com/support

@stanchan I've got your email - invite sent

NOTE: Changes to Swagger on Friday have caused build failures. We are investigating...

has joined #community201709

if i wanted to deploy windows, what would i have to do to the windows ISO to get windows installation?

`drpcli bootenvs update some-os-install - < bootenvs/some-os-install.yml` should refetch the "ISO" if the IsoUrl changed, right? in the context of tarballs. I note ce-sledgehammer has no hash

I?ll need to answer these a little later. I?ll come back to it

I meant to type IsoUrl/IsoFile instead of just IsoUrl

but alright

@lae updating the bootenv does not pull the iso

by design, Provision is push only. it never reaches out by itself. There is an upload for the master (soon 3.1) CLI that will pull then push

yeah, I figured updating wouldn't/shouldn't pull the iso, but yeah I didn't see a relevant alternative

don't have the need to upload a new iso right now since I just wiped/recreated a bootenv but - is the workflow supposed to be `drpcli bootenvs update os-install - < bootenvs/os-install.yml` and then `drpcli bootenvs uploadiso os-install`?

@lae yes

that's the new CLI command that uses the information from the bootenv to pull the ISO

@smartekb_twitter there are a few ways to do Windows deploys - generally, it's something we discuss on a call b/c it's environment & process specific.

@lae is correct. Changing the IsoURL, IsoFile, and IsoSha in the bootenv will make the bootenv unavailable until the iso matching those values is uploaded.

@smartekb_twitter - what @zehicle said. There are many ways and we don?t support any specific one openly today. We have many custom ways that we?ve worked through, but nothing consistent enough to support for the community. Hopefully that will change with some time and users.

Hi @greg @zehicle - interested in learning more about the Terraform integration; may have a use for it...

:slightly_smiling_face: I hope to have a video shortly and some information. We are almost there.

No rush ;)

the provider will manipulate machines in DRP from a pool. It will allow you to transition machines through install process.

You will be able to do something this: ``` provider "drp" { api_user = "rocketskates" api_password = "r0cketsk8ts" api_url = "https://147.75.73.159:8092" } resource "drp_instance" "one_linux_node" { count = 1 stage = "centos-7.3.1611-install" description = "Linux node installed by centos 7.3" } ```

That will take a machine and install centos on it.

hi all, i have a problem with a error message. from my point of view, the problem isn't a failure, it's a warning, may be. can anybody explain if i'm wrong or the message not fits? thanks!

TASK [wait for admin convergence [1 upto 20 minutes]] ************************** fatal: [10.241.236.92]: FAILED! => {"changed": true, "cmd": ["/root/digitalrebar/deploy/scripts/wait_for_rebar.sh"], "delta": "0:20:09.411997", "end": "2017-09-20 16:06:32.287249", "failed": true, "rc": 1, "start": "2017-09-20 15:46:22.875252", "stderr": "Took too long for system deployment to appear", "stdout": "Loaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-\n : manager\nThis system is receiving updates from RHN Classic or Red Hat Satellite.\nPackage epel-release-7-7.noarch already installed and latest version\nNothing to do\nLoaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-\n : manager\nThis system is receiving updates from RHN Classic or Red Hat Satellite.\nMetadata Cache Created\nLoaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-\n : manager\nThis system is receiving updates from RHN Classic or Red Hat Satellite.\nPackage jq-1.5-1.el7.x86_64 already installed and latest version\nPackage curl-7.29.0-35.el7.x86_64 already installed and latest version\nNothing to do\nWaiting on system deployment", "stdout_lines": ["Loaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-", " : manager", "This system is receiving updates from RHN Classic or Red Hat Satellite.", "Package epel-release-7-7.noarch already installed and latest version", "Nothing to do", "Loaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-", " : manager", "This system is receiving updates from RHN Classic or Red Hat Satellite.", "Metadata Cache Created", "Loaded plugins: product-id, rhnplugin, search-disabled-repos, subscription-", " : manager", "This system is receiving updates from RHN Classic or Red Hat Satellite.", "Package jq-1.5-1.el7.x86_64 already installed and latest version", "Package curl-7.29.0-35.el7.x86_64 already installed and latest version", "Nothing to do", "Waiting on system deployment"], "warnings": []} to retry, use: --limit @digitalrebar.retry

i see no reason for a fatal failure and a script brake.

The V3.1.0 release is out. Use the STABLE link to get the images.

the v3.1.0 link doesn?t have the correct version in the string. I has the right code, I think.

- FYI ^^

i have sync my local git today and copy the code to the target server. i think this must be the actual version.

@theta-my - you are using DRv2. My post was about DRP (Digital Rebar Provision). Sorry, I wasn?t clear.

ok,

@zehicle uploaded a file: https://rackn.slack.com/files/U02DHRR2L/F775N9WNA/v3_1_release_feature_list.md and commented: we're working to setup the community processes so that this becomes a community document

:+1:

we could use some stickers :joy:

@lae - direct message me your address and I?ll see what I can do.

how do i download iso's again? example debian 8

Which release?

in DRP v3.1.0 - ```drpcli bootenvs uploadiso <bootenv name>```

in either DRP v3.1.0 or v3.0.5 ```drpcli bootenvs install bootenvs/debian.8... ```

@greg is incrementer useful for an end user?

anyway, pushed an update for 3.1.0 to the aur package for drpcli

so I ran into a little bit of a snag using `uploadiso` (I think this command should give some feedback indicating success and not just return), but that's because I removed `IsoSha256` from my bootenv, but updating the bootenv didn't remove that key and so it kept the old one/couldn't verify the new tarball

just leaving `IsoSha256` empty in the bootenv file worked though

Congrats RackN team on another release! Will have to update and play with it tomorrow...

How does hosted UI work - http://xxx.rackn.com URL?

If you hit port 8082 on your 3.1 endpoint, you'll be redirected to the hosted endpoint

Via https

that's pretty spiffy

@lae it should not get filled in if empty. Legacy bootenvs are missing it. Once I can update sledgehammer the base bootenvs will have Shas

nah this was a custom bootenv I built for our appliance operating systems

Okay 310 doesn't fill it anymore

I had added `IsoSha256` while troubleshooting why the bootenv wasn't updating - then got rid of it when I realised how to use uploadiso

but the previous value was stuck until I had added `IsoSha256` to the yaml but left it empty (which then removed it from the installed bootenv)

anyway, good job on the UI, it's looking a lot better (I can actually add profiles for those bootenvs with requiredparams now :smile: )

input's kind of cut off here though (and maybe other similar dialogs) https://up.lae.is/i/1505959595-ecc06.png

It's a reference example

i've deployed a cent-os-7 to a host..is there a default login/password?

usually the recommended practice would be to inject an SSH key

prolly missed that in the docs..could u send me where i can find that?

Shane is correct. The root password is set by a parameter it defaults to RocketSkates

Hi

hello

Howdy

can someone show me/point me to how to inject ssh key ? cos the root/RocketSkates is not working for me to login...I'm installing centos-7

@smartekb_twitter - check the Doc FAQ page at: http://provision.readthedocs.io/en/tip/doc/faq-troubleshooting.html#add-ssh-keys-to-authorized-keys

@shane Having a bit of a problem upgrading from 3.0.5 to 3.1.0

Yeah - @shane has been finding some issues.

What are you seeing? Did you have a ?production? install?

I ran the curlbash in my existing "drp" dir where my prior 3.0.5 install was (isolated) - it seemed to run OK, but when I restated the dr-provision binary, I see I'm still on 3.0.5 :hushed:

add --force

to make it download the image.

but before you do that.

To the bash flags?

Yeah, but wait a moment

Standing by

Can you show my your curl bash command?

Give me a few mins

even betterer

Eating lunch - I have some fixes to install.sh

I too will eat the lunch then

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F77HXTS3E/image_uploaded_from_ios.jpg

Once Shane is happy, you will want to change stable to tip in the curl part and probably add --force to the bash part. You will still download stable bits, but you will get the fixed install.sh.

Ah, makes sense

@shane Ready to go when you are :)

copy

You currently have a 3.0.5 install in "production" mode (or "system" mode) - correct ?

and you want to upgrade to current 3.1.0 mode

just want to make sure we're on the same page on what you need to do

No, it's in 'isolated' mode

Want to upgrade to 3.1 in isolated

ah! much easier path ... :slightly_smiling_face:

That sounds good :)

ok @wdennis -- with the following notes: * stop dr-provision server first if it's running (install.sh will bomb out if it's running) * backup your content first ... just in case :slightly_smiling_face: Give this a go: ```export VER=tip curl -fsSL https://raw.githubusercontent.com/digitalrebar/provision/${VER}/tools/install.sh | bash -s -- install --isolated --upgrade``` This will download the current "tip" version of `install.sh`, but perform a `stable` version upgrade (since `--drp-version-VER` isn't specified)

Backup in progress ;)

Here we go...

Just sitting at "Installing Vesion stable..." -- that expected?

it's likely downloading content - you have a slow network connection ??

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F78ETPDD4/image_uploaded_from_ios.jpg

next steps are: ``` curl -sfL -o dr-provision.zip https://github.com/digitalrebar/provision/releases/download/$DRP_VERSION/dr-provision.zip curl -sfL -o dr-provision.sha256 https://github.com/digitalrebar/provision/releases/download/$DRP_VERSION/dr-provision.sha256```

so - it's downloading

Not super slow - on commercial cable internet

Ok, finally completed

All above done....

Now start DRP the way I used to?

yep

Gah - still says Version: v3.0.5-0-....

make sure you didn?t copy drpcli and dr-provision into /usr/local/bin or /usr/bin or somewhere.

No, I'm running it from cwd which is the isolated dir I installed into

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F78LVR7JB/image_uploaded_from_ios.jpg

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7732B1MX/image_uploaded_from_ios.jpg

So I think the problem is that my binary in ./bin/linux/amd64/dr-provision is dated July 3 2017

in the base directory you did the install - run the binary that was installed to `bin/linux/amd64/dr-provision` (assuming 64 bit linux, of course)

that should have been updated :slightly_smiling_face:

hmmm - it exited early - it looks like

They are symlinks

Try this: ``` rm -f dr-provision.zip rm -f dr-provision.sha256 rm -f sha256sums ```

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F78F2PQ5C/image_uploaded_from_ios.jpg

Then do the install command again, but add `--debug` as well to the bash section

@greg OK, done

Ack

Looks like it is downloading 'dr-provision.zip'

But, I got nothing happening when I do an 'iftop'

Is the install script still running?

Yes, but looks like I may have a local network issue... 44% packet loss measured :white_frowning_face:

Trying to resolve that, stand by

Ok, looks better, retrying

Looks like the DRP server is talking to AWS S3 now :)

Working but only pulling 50Kb/sec :disappointed:

OK, file date on dr-provision binary is 9/20/2017 now

Yay, v3.1.0-0-... on startup now :slightly_smiling_face:

How to set the DRP auth token for the remote UI?

should redirect you to an auth.

@greg @shane Looks like I have to set up the ISOs, machines, etc all over again?

I did get the remote UI

And am in it

But all my machines, ISOs, etc are not showing up

umm

okay - wait

UI shore is pretty tho :)

stop dr-provision

OK done

```sudo ./dr-provision --static-ip=192.168.1.158 --disable-dhcp --base-root=/home/dradmin/drp/drp-data --local-content="" --default-content=""```

Ok, let me try again

Phew :sweat_smile:

Now to commence the learnin'

So is there UI docs written yet?

oh no

:slightly_smiling_face:

previous usage should work. famous last words

TODO ? :stuck_out_tongue_winking_eye:

yes

If you've been using the new RackN UX then you've noticed the login buttons - those should be working now. There are a few changes in progress to save login sessions and allow storing endpoints if you are a registered user. There is NO COST for registering. The login is NOT required to use the endpoint admin features; however it will be required to access RackN content in the future.

http://rackn.github.io down?

yes - fix in process.

fix in is place

A n00b here. Any pointers on how to figure out why a machine is getting a dhcp lease from dr-prov, but not completing the pxe boot?

Hello @nzsouthernman, welcome! What version of Digital Rebar Provision are you running ?

check "drpcli prefs list" - make sure that you have set the unknown bootenv and default bootenv

Running whichever version is 'stable' as of this morning.

defaultboot is slegehammer, unknown is discovery.

Excellent -that should be the 3.1.0 release - you can verify this with `drpcli version`

Version: v3.1.0-0-b70cf8ee1f61844a6d64070a8b272c2bec512204

:)

Currently running host & pxeclient on vmware, on an isolated vswitch that has a firewall between it & our core.

you can also take a look at things with the UI - if you haven't seen it already - simply point your web browser to your DRP endpoint w/ https and port 8092

The UI is running as the RackN portal - and your web browser proxies the connection to your DRP endpoint - your DRP endpoint does NOT reach out to the Portal

Yeah, got the ui running (it's very nice btw), and checked the ISO's and earlier found that the initial sledgehammer I uploaded from the cli wasn't what the ui wanted. Downloaded and uploaded via the ui the sledgehammer b68 version and that appears to have made it happy.

presumably you see your pxeclient show up under the `Machines` inventory ?

*made the ui happy. The pxeboot still doesn't seem to respond still.

wha bootenv does it show ?

Nope. I only see the MAC under the Networking\leases as evidence that something's working.

I had dhcp logging going and syslog showed the event.

could you please run `drpcli info get` for us?

{ "api_port": 8092, "arch": "amd64", "dhcp_enabled": true, "file_port": 8091, "id": "00:50:56:9e:ec:bf", "os": "linux", "prov_enabled": true, "stats": [ { "count": 0, "name": "machines.count" }, { "count": 1, "name": "subnets.count" } ], "tftp_enabled": true, "version": "v3.1.0-0-b70cf8ee1f61844a6d64070a8b272c2bec512204" }

[https://files.gitter.im/digitalrebar/core/oJpG/image.png

Possibly something wrong with the DHCP config?

it's a UX issue - can you get the subnet list from the CLI? I'll look at the UX render issue. It would help to have your subnet entry

sorry - you're seeing a current UX bug right now (UX is still in tech preview status) ... can you please do from CLI ?

:slightly_smiling_face:

:D

[ { "ActiveEnd": "192.168.1.200", "ActiveLeaseTime": 60, "ActiveStart": "192.168.1.100", "Available": true, "Enabled": true, "Errors": [], "Name": "ens192", "NextServer": "192.168.1.10", "OnlyReservations": false, "Options": [ { "Code": 0, "Value": "[object Object]" }, { "Code": 1, "Value": "255.255.255.0" }, { "Code": 2, "Value": "[object Object]" }, { "Code": 3, "Value": "[object Object]" }, { "Code": 4, "Value": "[object Object]" }, { "Code": 5, "Value": "[object Object]" }, { "Code": 28, "Value": "192.168.1.255" } ], "Pickers": [ "hint", "nextFree", "mostExpired" ], "ReadOnly": false, "ReservedLeaseTime": 7200, "Strategy": "MAC", "Subnet": "192.168.1.10/24", "Validated": true } ]

Hmm, maybe I should delete and recreate my subnet from drpcli?

what is your "Reservation" strategy? Do you have reservations required ?

No, no reservations required. host is 192.168.1.10, fw is 192.168.1.1, dhcp scope set 192.168.1.100-192.168.1.200. It's a test ip range, so pretty much anything is fine.

can you add a TFTP client to your endpoint (eg "yum install tftp" or "apt -y install tftp-hpa") then check that TFTP is working for you with this basic test: ```tftp 127.0.0.1 get default.ipxe```

even better would be a client in the same vswitch that has access to your endpoint - then hit your endpoint IP (192.168.1.10) for same test

Testing what? browser to ui? I can bring up a linux vm on that vswitch pretty quickly.

testing basic TFTP to the drp endpoint

and that your tftpboot is serving the `default.ipxe` file

Adjusted dhcp to look like this;

[ { "ActiveEnd": "192.168.1.200", "ActiveLeaseTime": 60, "ActiveStart": "192.168.1.100", "Available": true, "Enabled": true, "Errors": [], "Name": "ens192", "NextServer": "192.168.1.10", "OnlyReservations": false, "Options": [ { "Code": 1, "Value": "255.255.255.0" }, { "Code": 3, "Value": "192.168.1.1" }, { "Code": 6, "Value": "192.168.1.1" }, { "Code": 28, "Value": "192.168.1.255" } ], "Pickers": [ "hint", "nextFree", "mostExpired" ], "ReadOnly": false, "ReservedLeaseTime": 7200, "Strategy": "MAC", "Subnet": "192.168.1.10/24", "Validated": true } ]

No pxeboot stuff though. Which option should that go in?

{Code: 3, Value: ip}, {Code: 6, Value: ip}, {Code: 15, Value: 'example.com'}, {Code: 67, Value: 'lpxelinux.0'}, ]

where IP is generally from the drpcli interfaces list

sorry about the UX bug... will get that updated tonight

No worries about the UX, that stuff happens in products under dev. Current DHCP schema follows;

{ "ActiveEnd": "192.168.1.200", "ActiveLeaseTime": 60, "ActiveStart": "192.168.1.100", "Available": true, "Enabled": true, "Errors": [], "Name": "ens192", "NextServer": "192.168.1.10", "OnlyReservations": false, "Options": [ { "Code": 1, "Value": "255.255.255.0" }, { "Code": 3, "Value": "192.168.1.1" }, { "Code": 6, "Value": "192.168.1.1" }, { "Code": 15, "Value": "burnside.school.nz" }, { "Code": 28, "Value": "192.168.1.255" }, { "Code": 67, "Value": "lpxelinux.0" } ], "Pickers": [ "hint", "nextFree", "mostExpired" ], "ReadOnly": false, "ReservedLeaseTime": 7200, "Strategy": "MAC", "Subnet": "192.168.1.10/24", "Validated": true }

FYI - your DRP endpoint does have a built in API documentation set - based on swagger: https://127.0.0.1:8092/swagger-ui/#/

Unfortunately, still no pxeboot from my test client. The ubuntu VM I popped onto the subnet can browse ok now though.

Replace localhost w/ your endpoint URL - and you may need to update the Swagger URL bar with the correct IP addr

(though, admittedly - it doesn't document the DHCP option codes :slightly_smiling_face: )

Hmmm, the test vm can browse externally, but can't access either the ui, or the swagger ui.

is your test vm able to tftp to the drp endpoint ? (also swagger-ui is HTTPS ... won't work on unencrypted http)

Gotchya. Swagger works on https. Initially tried to redirect me to 127.0.0.1, but changing the redir to 192.168.1.10 brought up the ui. Will test tftp to the endpoint shortly.

would you mind sharing the parameter list you started DRP with? it you did not set the static-ip then it could cause this problem

Also, it would be helpful to have your drpcli interfaces show list

[ { "ActiveAddress": "192.168.1.10/24", "Addresses": [ "192.168.1.10/24" ], "Index": 3, "Name": "ens192", "ReadOnly": true } ]

systemd started the daemon, I'll dig around and see if I can find what it passed to the initial run of it.

This was what I told it; sudo systemctl daemon-reload && sudo systemctl enable dr-provision

This is what systemd has to say about the service;

[Unit] Description=DigitalRebar Provision Integrated DHCP and File Provisioner Documentation=http://provision.readthedocs.io/en/latest/ After=network.target [Service] ExecStart=/usr/local/bin/dr-provision LimitNOFILE=1048576 LimitNPROC=1048576 LimitCORE=infinity TasksMax=infinity [Install] WantedBy=multi-user.target

can you please change the ExecStart stanza (in `/etc/systemd/system/dr-provision.service`) to append the following: ```--static-port=192.168.1.10```

NOTE: static-ip

then `sudo systemctl daemon-reload && sudo systemctl restart dr-provision`

doh :slightly_smiling_face: thanks @zehicle - yeah (--static-ip=192.168.1.10) ... cut-n-paste of wrong line

? dr-provision.service - DigitalRebar Provision Integrated DHCP and File Provisioner Loaded: loaded (/etc/systemd/system/dr-provision.service; disabled; vendor preset: enabled) Active: failed (Result: exit-code) since Tue 2017-09-26 13:36:51 NZDT; 1s ago Docs: http://provision.readthedocs.io/en/latest/ Process: 3294 ExecStart=/usr/local/bin/dr-provision --static-port=192.168.1.10 (code=exited, status=1/FAILURE) Main PID: 3294 (code=exited, status=1/FAILURE) Sep 26 13:36:51 lcrowbar01 systemd[1]: Started DigitalRebar Provision Integrated DHCP and File Provisioner. Sep 26 13:36:51 lcrowbar01 dr-provision[3294]: invalid argument for flag `--static-port' (expected int): strconv.ParseInt: parsing "192.168.1.10": invalid syntax Sep 26 13:36:51 lcrowbar01 systemd[1]: dr-provision.service: Main process exited, code=exited, status=1/FAILURE Sep 26 13:36:51 lcrowbar01 systemd[1]: dr-provision.service: Unit entered failed state. Sep 26 13:36:51 lcrowbar01 systemd[1]: dr-provision.service: Failed with result 'exit-code'.

Didn't like that

feck it, brb...

that's better. testing now

Ok. Will stop the daemon and run by hand. Will have another go at this tomorrow, thanks for your help so far today gentlemen.

@nzsouthernman - we'll look for you on the channel - cheers !!

``` {{if .ParamExists "part-scheme"}} {{template "part-seed-"+(.Param "part-scheme")+".tmpl" .}} {{else}} {{template "part-seed-default.tmpl" .}} {{end}} ``` so this isn't valid but I hope it should be obvious what I'm trying to do here - how can I get this effect of picking a template given a particular parameter? (not too familiar with golang/`text/template` and I'm still lost after looking at documentation)

```{{template (printf "part-seed-%s.tmpl" (.Param "part-scheme)) .}}```

I think, @lae

@lae @greg Parameterizing the OS install disks and the partitioning thereof should totally be a thing?

the os-disk is parameterized. @lae is working on parameterizing the partitioning scheme in a kickstart/preseed.

Of course, learning ?d-i partman? syntax is half the battle :stuck_out_tongue:

I suspect that @lae is just coming up with way to inject templates by variable name.

Then lots of templates could be hanging around.

by OS - most likely.

@greg The only edge-case I might see if making a linux-raid out of two/more hard disks before using the md as the disk to install on?

@lae is trying to do sub-templates?

It would be a great thing to have community-provided disk partitioning templates for preseed/kickstart

sub-templates work.

@lae is trying to generate parameterized named templates

@greg Sorry for my ignorance, but is that an entire ?top-level? template (like the standard ?net_seed.tmpl?) or a sub-template that would nest inside the top-level template?

golang text templates allow for templates in templates.

DRP originally (3.0.1) didn?t. We then added it to like 3.0.3 or 3.0.4.

This allows templates to include templates (recursively).

v3.1.0 adds tasks and stages which an alternative to some templates in templates.

Kickstarts/preseeds (the non-script parts) are perfect places for sub-templates. The command parts can be do better with tasks/stages now.

Hopefully we can learn about the new tasks/stages during the community call today

@wdennis - we'll be talking about the new v3.1.0 features :slightly_smiling_face:

you can check out the full agenda: https://docs.google.com/document/d/1utYrt4x0VIvTTlZqMHA1Jm489ab4W-Nlma9EVvBTdDk

@shane shared a file: https://rackn.slack.com/files/U6QFVRJNB/F7847GB6H/v001_-_20170926_-_intro__naming__digital_rebar_provision_3.1_and_3

Yes, parameter-driven sub-templates would be great for partitioning schemes - a collection of such templates could be community (or DR) provided, and less wheels reinvented

@shane, Thanks! will check this stuff out before the call

@wdennis the content system is also a key feature for 3.1 to help those reuse thoughts

@shane Meetup soon, right?

Meetup underway at https://www.meetup.com/digitalrebar/events/243477494/

Nice work @shane :slightly_smiling_face:

Thx @wdennis - pleasure to "virtually meet" you :slightly_smiling_face:

- for those of you couldn't attend the meetup - we'll post the video shortly and provide links here ... keep an eye out for additional announcements on scheduling and agenda items for v002 meetup

@greg @wdennis that's exactly what I'm trying to do - have separate partitioning templates based on what workload is required the "edge-case" of making linux raid is well, an understatement. there are lots of possible disk layouts, including with LVM (such as making a layout that follows CIS benchmarks (https://www.cisecurity.org/cis-benchmarks/)) - I was originally doing the same thing back when I was using cobbler.

Anyway, so I tried what @greg provided but I get this: ``` [lae@yuzu provision-content]$ drpcli templates upload templates/labs-seed.tmpl as labs-seed.tmpl Error: Parse error for template labs-seed.tmpl: template: labs-seed.tmpl:65: unexpected "(" in template clause ```

@zehicle Can't seem to find your slack invite. My gmail or uber account?

@stanchan checking...

gmail

@lae Will have to let @greg respond, not a golang guy myself, so not going to try to t?shoot the syntax? But, powerful idea, and looking forward to the solution :slightly_smiling_face:

My idea would be to have a library of disk partitioning templates that one could choose from for a given deployment

Most of my use cases use a single disk (could be a RAID virtual disk al la Dell PERC) with partitions for swap and then LVM PV, as is the case with default net_seed template

Time to feed the :bear:!

But every once in a while, I get a more custom request, which I have to handle when I provision the ?n? systems of that type

@ use should be using the DR bear icon! :stuck_out_tongue:

@wdennis - we have a hardware RAID partitioning solution that will be released soon - that piece is a paid-for-content piece, as we've put a LOT of development and testing work in to it ... but it's designed to do a large number of flexible hw raid setups

@shane sounds very cool, but really thinking more about flexibility of Linux partitioning of one disk

are you using ubuntu or centos based distros ?

yes :stuck_out_tongue:

Actually mostly Ubuntu these days

well - as bad as 'debian installer' is - the 'partman' pieces do have some pretty flexible capabilities - albeit - not very intuitive .... and those could be pretty easily pushed in to a flexible set of templates

But have folks who occasionally spec CentOS

yeah, I'm pretty familiar with d-i's partman now

Yes, I wish d-i was as easy as kickstart?

@lae ... my condolences ... :slightly_smiling_face:

so like, I'd prefer to be able to use my own existing seeds with those layouts but at the moment I can only really do a copy of a seed per layout...which is gonna be a pain to manage if I brought every layout I used

hence, trying to parametrize the partitioning scheme template

I?d pay for a sane disk partitioning UI (like the Ubuntu installer?s for instance) that generates correct d-i partman lines?

disk partitioning UI sounds great

lot of work though, don't think that's what rackn worked on

(or maybe it is)

Once you have the correct d-i in a library of configs, it should be easy enough (?) to modify

Sounds like a community-contrib thing?

Would be great to have a way to have something like ?Ansible Galaxy? for DR community-provided resources

Exposed thru the DRP UX :slightly_smiling_face:

not UI pieces - but templates/profiles to and a plugin to drive the raid configs

What I was trying to say was if there is community-provided content (such as templates, content packs, plugins, etc.) There could be a way to browse them thru the UX (kind of like what?s available in GitHub Atom editor in the way of Packagaes / Themes)

It would be up to the end-user whether to trust them and utilize them or not (perhaps with a rating system?)

@wdennis - we have content browsing already in the UI ...

:slightly_smiling_face:

@shane Need to learn more about that :wink:

if you go to the Content page - the right side panel labeled "organization content" is a library of available content to pull in to your local Endpoint

Not too familiar yet about Content packs, Plugins, Stages, Tasks, Jobs

@shane uploaded a file: https://rackn.slack.com/files/U6QFVRJNB/F78JEAR4Y/screen_shot_2017-09-26_at_14.03.53.png

in that view, the middle panel is content I've added; the right panel is the available content

I added the content from the content library - but I did do it via the CLI - not UI ... could just as easily have clicked on the UI "Transfer" link

the content also is versioned - as you can see if something has an upgrade available - and you can choose to Upgrade your content pack for what you currently have installed

in that screen shot - the drp-community-content has an upgrade available

Yes, I see it in my DRP UI - all the available content right now is RackN-provided?

"drp-community-content" is the ... ahem ... Community Content :slightly_smiling_face: freely available

lol

some of that content you see is going to shuffle over to the "free-for-register" content

like I mentioned - UX is still Tech Preview ... and we're baking a few of the last bits

About registering - can self-create a RackN beta account, or how to do so?

@zehicle and team did an amazing job pulling all of that together

n.m - found ?sign up? link

Upper right - click on "RackN Login" and then sign up

OK, did signup

Went to Content page and tried to d/l update for community content, got error...

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F799Z242Y/image_uploaded_from_ios.jpg

looks like you tried to add content that already exists

But anyways, get the idea...

"ce-root-access" already exists - you'd have to destroy it first, then re-create

So, `explode_iso.sh` is failing for every ISO I have uploaded with an exit code of 255

any ideas?

@wdennis - you have an older version of "drp-community-content" (note the long winded name w/ spaces)

you probably tried adding the newer pack "drp-community-content" - they're same things - just renamed

that's a bug I ran in to - and we got cleaned up a bit

you should be able to nuke the "Digital Rebar Provision Community Content" pack then Transfer the "drp-community-content" pack

Ok

Remember the golden rule: UX is Tech Preview still

:slightly_smiling_face:

And a damn fine tech preview it is

@carl - can you check the tftpboot/isos/ directory on your DRP Endpoint - the ISO gets staged there with "drpcli bootenvs uploadiso FOO" - then explode_iso.sh runs on it

additionally - check the filesystem space of your tftpboot/isos/ directory

if you did "isolated" install - then it's going to be in $HOME/drp-data/ location - if you did "production" install - it should be in /var/lib/dr-provision/tftpboot (for 3.1) or /var/lib/tftpboot (for 3.0.x)

@lae - I?ll try and look at it tonight.

I probably messed something up.

also thanks for the stickers! (they arrived today)

Nice!

For anyone who missed the meetup today, here is the https://youtu.be/LpVHYY9NdYo

hmm

I guess I could use ansible as another way to manage my templates

@lae uploaded a file: https://rackn.slack.com/files/U54E4SD4G/F79A7LEFL/image.png and commented: oh right - the machine overview is a little unwieldy when a lot of profiles exist

@lae - thx for the feedback on that - we know that page can get a bit crowded, and working on some ideas/thoughts around filtering and presentation of that information. If you have any ideas - feel free to share. :slightly_smiling_face:

- the v002 meetup details have been posted: meetup link: https://www.meetup.com/digitalrebar/events/243490128/ agenda document: https://docs.google.com/document/d/1FRFI-vONJY9yje9UsBqCI8XhojJ0XARsFgs4jbm-VRk also - vote now - if you'd like to see us move to a Weekly format with more demo/content versus the current published every-other-week format: meetup poll: https://www.meetup.com/digitalrebar/polls/1255504/

@shane shared a file: https://rackn.slack.com/files/U6QFVRJNB/F78DZGSTS/v002_-_20171010_-_digital_rebar_provision_3

So, question about the RackN login on the DRP UI...

When I log in, it has my org as "Personal", no way to change that?

But bigger question - when I use the "sandwich" UI widget to open the left nav, and select Endpoints, I just see "127.0.0.1:8092" which of course is bogus, and there's no way to add actual DRP endpoints to that

"tech preview" :relaxed:

Multi-endpoint management not baked yet - simply change URL reference for now

Ah, got it

Maybe when I run into these problems, you can say I've been "TP'd" (tech previewed) :P

Hi Gents. Update on yesterday - the issue is definitely the embedded dhcp server in dr-provision. I switched dhcp out to my firewall and enabled the pxeboot bits to pull off drp's tftp server and the sledgehammer env. booted fine, and now I have a machine registered.

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7AE3QR4P/image_uploaded_from_ios.jpg

@wdennis that is exactly the feature set that I'm working on right now... adding an endpoint exposed a validation loop. Once it's working, you'll be able to add/remove endpoints from your org. You'll also be able to shared orgs for multiple users. In the meantime, we can manually add endpoints for you on the backend so they correct on your list if you ask 1x1

@zehicle Cool, will PM details

@nzsouthernman_twitter - we should talk about it some more. We need to fix that. My guess is that the subnet isn?t quite right.

Is there anything I can provide from my end to help? My working firewall is pfSense, and I just reloaded dr-provision with --disable-dhcp

The subnet will still be there

IT would be nice to see the subnet object and what the network specs are. Is DRP directly attached to the network in question?

I first had it working by disabling the subnet in the UX entirely, then enabled fw dhcp server, entered the right next server & lpxelinux.o file and hey presto!

Yes, drp is directly on the network. Network is an esxi 6.5 vswitch.

{ "ActiveEnd": "192.168.1.200", "ActiveLeaseTime": 60, "ActiveStart": "192.168.1.100", "Available": true, "Enabled": false, "Errors": [], "Name": "ens192", "NextServer": "192.168.1.10", "OnlyReservations": false, "Options": [ { "Code": 1, "Value": "255.255.255.0" }, { "Code": 3, "Value": "192.168.1.1" }, { "Code": 6, "Value": "192.168.1.1" }, { "Code": 15, "Value": "burnside.school.nz" }, { "Code": 28, "Value": "192.168.1.255" }, { "Code": 66, "Value": "192.168.1.10" }, { "Code": 67, "Value": "lpxelinux.0" } ], "Pickers": [ "hint", "nextFree", "mostExpired" ], "ReadOnly": false, "ReservedLeaseTime": 7200, "Strategy": "MAC", "Subnet": "192.168.1.0/24", "Validated": true }

Okay - so - it sounds like it might be subnet config, but it also could be the vswitch. I seem to recall that the vswitch can ?help? by not letting broadcasts through in some cases and some settings.

The test client is a vm with one vnic, on the drp vswitch. The drp host has one nic, on the drp vswitch. The pfsense fw has two nics, one on drp vswitch and one on our core lan.

The subnet looks okay - 66 isn?t required. the server fills it in.

(all three machines are client vm's on the same esxi host)

But you're probably right, it's most likely b/c drp is sitting on esxi. Ubuntu 16.04.3 with open-vm-tools installed, so the vnic should be 10gigE.

hmmm - A couple of other things you can try. If you want, the subnet looks okay. DRP has some preferences that control debugging. You can change the debugDhcp preference to 2 (through the cli) or HIGH in the global setup UI page. This will cause the DHCP server part of DRP to dump packet content.

That way we can see if DRP is getting and responding to DHCP messages.

Well, if we are in an Ubuntu vm, it should be okay.

Now, ubuntu firewall rules could get in the way. Not sure what you have there.

It is indeed dishing out dhcp leases - I had them coming through yesterday ok.

firewall should be disabled - that was one of the install instructions I followed yesterday when setting this all up.

okay - cool and good.

dr-provision2017/09/27 02:29:48.674105 Received DHCP packet: type Discover xid 0x4f61e751 ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 02:29:48.674494 Received DHCP packet: type Request xid 0x4f61e751 ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 02:29:48.674718 xid 0x4f61e751: 192.168.1.104 is no longer able to be leased: No lease for 192.168.1.104, covered by subnet 192.168.1.0 These are from before I --disable-dhcp

When I fired it up it cleaned out yesterday's leases

ooh, ubuntu firewall may have been running... give me a sec...

also apparmor can do funky things to DHCP servers (though mostly keep them from starting).

dr-provision2017/09/27 03:16:12.941340 Received DHCP packet: type Discover xid 0x579e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 03:16:15.028534 Received DHCP packet: type Discover xid 0x589e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 03:16:19.093171 Received DHCP packet: type Discover xid 0x599e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 03:16:27.167050 Received DHCP packet: type Discover xid 0x5a9e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dang.

No dice. However under Subnets/Leases I see a brand new lease listed.

Make sure you enable the subnet if you are testing it

dr-provision2017/09/27 03:17:45.724011 Received DHCP packet: type Discover xid 0x599e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 03:17:45.724446 Subnet ens192: handing out existing lease for 192.168.1.101 to MAC:00:50:56:9e:6e:cb dr-provision2017/09/27 03:17:53.798103 Received DHCP packet: type Discover xid 0x5a9e6ecb ciaddr 0.0.0.0 yiaddr 0.0.0.0 giaddr 0.0.0.0 chaddr 00:50:56:9e:6e:cb dr-provision2017/09/27 03:17:53.798614 Subnet ens192: handing out existing lease for 192.168.1.101 to MAC:00:50:56:9e:6e:cb Good catch Greg.

But no pxeboot. :(

Yeah - your node isn?t accepting the DHCP reply for some reason, it appears.

No matter - handing the pxeboot off to the firewall works for me while I'm fiddling around.

You should see an ACK and Request message.

Have to head home now, will have another crack getting further tomorrow.

Make sure that the otehr DHCP servers are off as well in the future (just in case).

cool - though DRP is designed to work in the pfSense case as well.

I?m using pfSense as the DHCP server, handing off PXE boot to DRP

@lae - it turns out that `template` doesn?t do variable expansion on its parameters.

So, I wrote my own.

This will be in the next release. I?m putting in tip now.

This can be dangerous.

We don?t prevent loops

yo - wow im on slack with you still

did any check show up or do i need to call the bank :slightly_smiling_face:

@monkey moving this to 1x1

@greg are you basically just defining your own implementation for `template` on rendering the template? (just so I understand what's going on in that commit)

Yeah - the golang txt template call doesn?t know about templates to recurse (it seems like it should, but it doesn?t). So, that is what implemented.

@lae - I was waiting for the build to finish.

but here we go.

Tip is updated to have CallTemplate , which works like template but can take parameters.

This is ?dangerous? and probably when text template doesn;t really support it. You can create infinite loops. I?m not opposed to this. One day I may get around to writing a loop detector, but it isn?t there now.

Here is an example content file.

@greg uploaded a file: https://rackn.slack.com/files/U02DGQYK1/F7B17T1AB/template-test.yml and commented: Example CallTemplate content

To import it, you will need tip installed.

```drpcli contents create - < template-test.yml```

With this in place, I do this:

```drpcli machines create '{ "Name": "fred", "Address": "1.1.1.1", "Stage": "template-viewer" }'```

This creates a fake machine with stage set to `template-viewer` which implicitly sets the bootenv to `template-viewer`

```drpcli machines list Name=fred```

Get the UUID and do this:

```curl http://127.0.0.1:8091/machines/f17a55fe-c415-4f38-88dd-6fdfb2ce0c8d/test```

should get this: ``` This is a test template. We are doing parameterized injection of templates. This is the empty template with no pills added. ```

```drpcli machines set f17a55fe-c415-4f38-88dd-6fdfb2ce0c8d param template-selector to blue-pill```

will get this:

``` This is a test template. We are doing parameterized injection of templates. This is the blue-pill. ```

setting `template-selector` to `red-pill` will get something different

You can probably guess.

Setting it to `green-pill` will show template in template expansion.

``` This is a test template. We are doing parameterized injection of templates. This is the green-pill. Is both pills at once. This is the blue-pill. This is the red-pill. ```

Setting a second variable `bonus-pill`

```drpcli machines set f17a55fe-c415-4f38-88dd-6fdfb2ce0c8d param template-bonus-pill to bonus-pill.tmpl```

Will import a template from within a template dynamically.

`bonus-pill` is a full template name.

setting `bonus-pill` to `centos-7.ks.tmpl` will embed the kickstart file in test template

This is how I often test my templates quickly.

@lae @shane hope that helps. :slightly_smiling_face:

@greg - that's awesome - nice job whipping this up so quickly ... hoping it solves @lae request

The pattern works for tasks too. Actually, it is really useful for tasks because tasks don?t have a direct render point like bootenvs do.

though I'm concerned .... you select "red" pill and "blue" pill and combined pill is "green". Shouldn't it be "purple" ??? :stuck_out_tongue_winking_eye:

That isn?t how light sabers work.

And yes the hate is flowing.

:slightly_smiling_face:

:slightly_smiling_face:

I had this earlier, but was waiting on the build to finish. Had to work around sourceforge being partially down.

lol

ok - just updated my deployment to tip (yay ansible) - will try `CallTemplate` in a bit

(it works)

\o/

:slightly_smiling_face:

if I'm reading this thread - we just added dynamic workflow to DRP. I thought v3.2 was more of a cleanup release :stuck_out_tongue:

very exciting stuff

has joined #community201709

@shane / @greg - would it be possible to create a video based on the already-published vid of Terraform/DRP integration, showing creation of the stages / workflows / parameters used to do the install at Packet?

hey @wdennis - you can see the drpcli calls I made in that video in the github content - it's all posted there

the "demo-run.sh" is just the front-end driver to the process which controls the bin/control.sh script - that script makes all of the direct calls to do the work

we have the "5min-drp" video posted in youtube - but it doesn't directly show the drpcli commands

It's not the 5-min DRP install one, but the Terraform integration one @greg did the demo on

https://youtu.be/5bxcpmxQXx4 - interested in how the stuff used at around 5:00 in the vid was created

stage maps ?

those are plumbed in via the 5min-drp cli calls too

Well, how to create the stages 1st, then tie them together in maps (workflows?)

sure - the JSON blob that's used to inject the stages ... sort of shows it ... we're working on getting better doc on the 3.1 features on the website around that

you can also piece stages together in the UI ... so for first time work - it's a good process to follow

then you can pull the JSON blob that gets built from the UI

I'll review the 5-min vid in the meantime

problem - restarted DRP (isolated mode) and am getting a TLS Exception error when trying UI...

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7AMFJ67K/image_uploaded_from_ios.jpg

reclick it- your certs should be generated.

could have been

No good

Still getting error

Hard reset the browser window? Not sure.

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7B8XQPNG/image_uploaded_from_ios.jpg

Did restart browser - still getting the error...

Ddi you click to https://192.168.1.148

Yes

How to "Accept Certificate"?

advanced button usually.

The browser should warn saying self-signed cert.

If I go to the DRP server's URL, it just flips me back to the RackN UX

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7B933A4U/image_uploaded_from_ios.jpg

Then you?ve accepted the cert

The DRP server software gens a SSL cert?

yes to do ssl

Where located?

in the directory where you started drp

server.crt and server.key

I see 'server.crt' and 'server.key' in my base DRP directory, but they have old dates...

we will reuse them.

It is most likely something else.

does the cli work?

Yes, drpcli works

@wdennis uploaded a file: https://rackn.slack.com/files/U416T0AAX/F7CA4C2J3/image_uploaded_from_ios.jpg

You closed the browser and repoened it.

Log out from the Saas Part (in the upper right corner) and refresh everything.

Yup, then even rebooted browser host system

Ok, signed out then back in to RackN UX site

Now will use sandwich menu and select my endpoint

Still getting TLS Exception

just use the DRP endpoint redirect. I?m worried that some cookies are something are getting in the way.

Because it may be that you need to enter https://

it is supposed to add it, but it may not be working.

Did that, no good still

That is why I want you to log out of the RackN Portal. Then use the DRP redirect to see if you can access the DRP pages.

I did

okay

Trying another client system now

Port 8092 right?

yes

Browser can't connect on this one...

firewall rules?

Oh goddamn it

YES

Stupid docker

Note to self: find way of perma-clearing docker-created iptables rules on my DRP host

So, that's weird that the UX was throwing a TLS error when really it couldn't connect to the endpoint...

apparently. It is the normal cause for failure.

I think it is a guess and points to the most likely failure.

'systemctl stop firewalld; systemctl mask firewalld' FTW

or just, `firewall-cmd --zone=public --addport=8092/tcp --permanent && firewall-cmd --reload`

anyway, I just happened to stumble onto this, happy birthday to rackn's twitter account

or well whoever's birthday the rackn's twitter account is set to

Thanks! Tomorrow is our official 3 year anniversary

we just updated the Ansible integration to make it easier. Demo is of Kubernetes deploy. https://youtu.be/b5himGQ1Zew