I think if you have the rpm already pre-built, you can probably bypass building  StarlingX and just try to run the patch-build command

You just need to make sure the tool can find the rpm.

Someone else on this channel might have info about the build problems and environment setup

Al

From: LiuYongfu <114769003@qq.com>
Sent: Wednesday, March 8, 2023 7:08 AM
To: Bailey, Henry Albert (Al) <Al.Bailey@windriver.com>; starlingx-discuss <starlingx-discuss@lists.starlingx.io>
Subject: Re：RE: [Starlingx-discuss] how to install my own rpm

CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi，

    thanks very much . so I need to make a pach  and install it .  my first step is to build the starlingx .when I build this.  It has some problems as below:

could you please help me ? thanks very much.

---

LiuYongfu 114769003@qq.com

Original

From:"Bailey, Henry Albert (Al)"< Al.Bailey@windriver.com >;

Date:2023/2/18 5:05

To:"LiuYongfu"< 114769003@qq.com >;"starlingx-discuss"< starlingx-discuss@lists.starlingx.io >;

Subject:RE: [Starlingx-discuss] how to install my own rpm

With StarlingX, the way to add an rpm to the platform is through a patch.

Within the StarlingX patching system it will install the rpm, and will prevent it from being removed by a reboot.

There is a video from the Open Infrastructure Foundation that shows an example of creating your own patch to allow an rpm to be installed.  It uses the patch_build.sh command

https://www.youtube.com/watch?v=vwqhxpgaxXE

For example, if I wanted to add an RPM for pyflame  my patch_build.sh   syntax would look something like this:

# Run this script from within $MY_WORKSPACE

PATH=$MY_REPO/stx/update/extras/scripts:$PATH

PATCH_ID=PYFLAME

DIR=std/rpmbuild/RPMS

PYFLAME=pyflame-1.6.6-4.el7.tis.1.x86_64.rpm

patch_build.sh \

--id ${PATCH_ID} \

--reboot-required=Y \

--summary "New pyflame rpm " \

--desc "Adds a new pyflame rpm "\

--controller ${DIR}/${PYFLAME} \

--controller-worker ${DIR}/${PYFLAME} \

--controller-worker-lowlatency ${DIR}/${PYFLAME}

Be aware that in this example,  I am ‘adding’ an rpm  rather than updating one, so I needed to include special fields

(--controller ,  --controller-worker ,  --controller-worker-lowlatency ) so that the platform knows that the rpm belongs on controller and AIO hosts (and in this example it will not install it on worker or storage nodes)

The video (and my example) assume you have a development environment setup with your variables, etc..  already defined.  If that is not the case, you may need to ensure you have the starlingx/update repo checked out and make the appropriate changes to the commands.  

Al

From: LiuYongfu <114769003@qq.com>
Sent: Saturday, February 11, 2023 5:52 AM
To: starlingx-discuss <starlingx-discuss@lists.starlingx.io>
Subject: [Starlingx-discuss] how to install my own rpm

CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.

Brief Description
-----------------

when I install my private(own) rpm to the compute node. when the node is restarted .my private rpm is deleted.
so I want to know : are there some ways to install my own rpm on compute nodes or is it just a bug?

Steps to Reproduce
------------------
I installed my GPU rpm packages on the compute node,and it worked right.

Expected Behavior
------------------

the rpm packages works correctly all the time.

Actual Behavior
when reboot the compute nodes .my gpu rpm packages are removed by the system.

thanks very much.

LiuYongfu 114769003@qq.com