ntp recipe didn't install ntpdate files - yocto

I have included ntp package.
When yocto OS image comes out, it has only ntp, there is no ntpdate.
I found someone has the same problem but no solution:
https://forums.xilinx.com/t5/Embedded-Linux/How-to-install-ntpdate-command/td-p/1052770
Why it has only ntp? The recipe did include ntpdate.
https://git.congatec.com/yocto/meta-openembedded/blob/ddd3ccd63949f9a24040eb39c136ef9730f2662a/meta-networking/recipes-support/ntp/ntp_4.2.8.bb
ex:
install -m 755 ${WORKDIR}/ntpdate ${D}${bindir}/ntpdate-sync
...
install -m 0644 ${WORKDIR}/ntpdate.service ${D}${systemd_unitdir}/system/

ntp and ntpdate are two packages of the ntp recipe.
If you want to add ntpdate, just add to your image:
IMAGE_INSTALL += "ntpdate"
Explanation:
Take a look at the recipe, and at the PACKAGES variable:
PACKAGES += "ntpdate sntp ${PN}-tickadj ${PN}-utils"
It means that the ntp recipe contains the packages: ntp (default ${PN}), ntpdate, sntp, ntp-tickadj, ntp-utils.
You can also take a look at the environment of the ntp recipe:
$ bitbake ntp -e |grep ^PACKAGES=
PACKAGES="ntp-src ntp-dbg ntp-staticdev ntp-dev ntp-doc ntp-locale ntp ntpdate sntp ntpq ntp-tickadj ntp-utils"

In the ntpd manual page where is explained the option -q is reported:
-q
Exit the ntpd just after the first time the clock is set. This behavior mimics that of the ntpdate program, which is to be retired. The -g and -x options can be used with this option. Note: The kernel time discipline is disabled with this option.
In my opinion this means that is preferible use the daemon ntpd -q -g -x instead of ntpdate.

Related

How to install/start using swtpm on Linux

I am trying to start using TPM on Linux, on my Raspberry Pi to be accurate, and the easiest would be to use swtpm to get used to commands and system, before using a TPM chip. I've tried the way presented on https://github.com/stefanberger/swtpm/wiki#compile-on-ubuntu-2104, but I got the error "Unable to locate libtpms-dev".
Then I looked for a way to install libtpms, and found this one solution. But after getting those both, I still couldn't run this command :
sudo swtpm socket --tpmstate dir=/home/ludovic.peyter/swtpm --tpm2 --server type=tcp,port=2321 --ctrl type=tcp,port=2322 --flags not-need-init,startup-clear
All I get is the following error :
swtpm: SWTPM_NVRAM_Lock_Dir: Could not open lockfile: No such file or directory
and
swtpm: Error: Could not initialize libtpms.
And here am I stuck, finding nothing to help me with this problem, or even an other way to avoid it.
Thanks for reading.
I have a complete solution, built with many different solutions and my own tests.
sudo apt -y install dpkg-dev debhelper libssl-dev libtool net-tools libfuse-dev libglib2.0-dev libgmp-dev expect libtasn1-dev socat python3-twisted gnutls-dev gnutls-bin libjson-glib-dev gawk git python3-setuptools softhsm2 libseccomp-dev automake autoconf libtool gcc build-essential libssl-dev dh-exec pkg-config dh-autoreconf libtool-bin tpm2-tools libtss0 libtss2-dev
Then make a new directory for more comfort, and step into it.
Clone git repository for libtpms :
git clone https://github.com/stefanberger/libtpms.git
Move to the generated libtpms directory and run these commands :
./autogen.sh -–with-openssl
make dist
dpkg-buildpackage -us -uc -j4
Then, as asked at the end of the last running command, run :
libtool --finish /usr/lib/aarch64-linux-gnu/
The directory could be different, so pay attention to the warning at the end of dpkg command.
And to finish the installation of libtpms :
sudo apt install ../libtpms*.deb
Now get back to the previous directory and clone swtpm git repository :
git clone https://github.com/stefanberger/swtpm.git
Now run this command :
dpkg-buildpackage -us -uc -j4
And the command asked by the previous running command :
libtool --finish /usr/lib/aarch64-linux-gnu/swtpm
The directory could be different, so pay attention to the warning at the end of dpkg command.
Now finish the installation with this :
sudo apt install ../swtpm*.deb
Everithing you need is installed. Now you need to modify the file ~/.profile or ~/.bash_profile to add this line :
export TPM2TOOLS_TCTI="swtpm:port=2321"
And now, everytime you need your swtpm, open two terminals, and in one of them run :
swtpm socket --tpmstate dir=<swtpm_path> --tpm2 --server type=tcp,port=2321 --ctrl type=tcp,port=2322 --flags not-need-init,startup-clear
In the other terminal, as long as the above command runs, you can run all your TPM commands.
Have you checked that the tpmstate directory exists? A mkdir /home/ludovic.peyter/swtpm2 could fix it for you.

Docker-compose: /usr/local/bin/docker-compose : line 1: Not: command not found

i'm trying to install Docker-compose on my Raspberry Pi 3+ which installed Raspbian buster.
I followed instruction on docker.com. After I entered command : sudo curl -L https://github.com/docker/compose/releases/download/1.20.0/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose.
It show a table for downloading
Result
It seems nothing downloaded, just have a file docker-compose saved in /usr/local/bin/docker-compose. When I opened it, it empty. Then I enter command docker-compose -v, it displayed error /usr/local/bin/docker-compose : line 1: Not: command not found.
Anyone have solution?
UPDATE:
Added the following command to my answer to download the LATEST version without specifying any version number at all so the download can be scripted.
curl -L "https://github.com/docker/compose/releases/download/$(curl https://github.com/docker/compose/releases | grep -m1 '<a href="/docker/compose/releases/download/' | grep -o 'v[0-9:].[0-9].[0-9]')/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
It's a bit untidy, but it works. If you have a more elegant way than mine, ping it to me in the comments and I'll update my answer.
Just need to set the perms on the file:
chmod +x /usr/local/bin/docker-compose
Use the file command to validate that you pulled the correct arch for your system.
Intro:
Although docker-compose can be installed from a repo per the accepted answer, apt-cache show docker-compose reveals that as of 20211201 the repo version is only v1.25; about 2 years behind the current v2.1.1 release. In order to take advantage of more modern docker file versions, I needed to get the Github download working.
Short Answer:
The Docker documentation for Docker-Compose is WRONG. They forgot to preface the version number in the command with a "v"; consequently the download fails. Apparently this has been wrong for ages...
Longer Answer:
I ran the below command from the Docker-Compose documentation, and substituted the version "2.1.1" for "1.29.1" per Docker's guidance:
To install a different version of Compose, substitute 1.29.2 with the
version of Compose you want to use.
sudo curl -L "https://github.com/docker/compose/releases/download/2.1.1/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
The resulting download was 9 KB for a 23 MB binary. Clearly the link was bogus. So I went to the root of the address used in the command "https://github.com/docker/compose/releases" and right-clicked on the version of Docker-Compose that I wanted and chose "Copy Link Address"
This revealed the link Docker was telling folks to use didn't have a "v" prefaced before the version number in the https:// address part of the command.
Solution:
Preface a "v" before the version number you want in the link as below and the command executes successfully:
sudo curl -L "https://github.com/docker/compose/releases/download/v2.1.1/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
BTW, I too was downloading docker-compose for a Raspberry Pi using the aarch64 binary for Ubuntu 20.04 LTS. However, the missing "v" fix for the broken download address should work for any platform.
This is because on a raspberry pi the url part of the command results in
https://github.com/docker/compose/releases/download/1.24.1/docker-compose-Linux-armv7l
Looking at the latest stable release at https://github.com/docker/compose/releases/tag/1.24.1 you can see there is no download for the armv7l architecture so the file is empty because there is nothing to download.
Will update answer once I figured out how to install docker-compose on Raspian.
Edit:
Via apt-get. Note: Currently (Nov. 8 2019) this installs version 1.21 which is not the latest available.
sudo apt-get install docker-compose
Via pip3. (Installs latest)
sudo apt-get install python3-pip
sudo pip3 install docker-compose
And then restart your system with
sudo shutdown -r

Install ack-grep on CentOS

I went through fair amount of google search to install ack-grep on CentOS but I didn't find anything help. I also looked for the source codes but couldn't find it neither. Does anyone know how to install it on the OS?
Thanks a lot.
Could be essentially the same as https://stackoverflow.com/a/23155007/35946 but on CentOS 6.7 the answer is:
# yum install epel-release
# yum install ack
if you don't have the root permission, you can do as follows:
$ curl https://beyondgrep.com/ack-2.22-single-file > ~/bin/ack && chmod 0755 !#:3
or you can change to root user:
$ sudo su
# curl https://beyondgrep.com/ack-2.22-single-file > /bin/ack && chmod 0755 !#:3
You can get it from the EPEL software repository.
From the EPEL FAQ:
For EL5:
su -c 'rpm -Uvh http://download.fedoraproject.org/pub/epel/5/i386/epel-release-5-4.noarch.rpm'
...
su -c 'yum install ack'
For EL6:
su -c 'rpm -Uvh http://download.fedoraproject.org/pub/epel/6/i386/epel-release-6-8.noarch.rpm'
...
su -c 'yum install ack'
Go to Beyond Grep and look at the section titled
Install The ack executeable
curl http://beyondgrep.com/ack-2.14-single-file > ~/bin/ack && chmod 0755 !#:3
And replace ack.2.14 with the current version of ack.
You may need to create the directory mkdir ~/bin/ first. You may
also need to modify ~/.bashrc to include this new path E.G.:
PATH=$PATH:$HOME/bin
Then reload ~/.bashrc
source ~/.bashrc
Test the installation by running ack:
rpm -qa | ack s
This should display any installed packages containing the letter s. (some linux distributions may use ack-grep as the command.
How did you try installing it? Are you using yum? The package is probably not called "ack-grep", but just "ack".
The name "ack-grep" is a Debian-specific thing because there was already a package called "ack", so they called it "ack-grep" instead. That was years ago and now they're dropping the original "ack" package and renaming "ack-grep" to "ack".
For RedHat Enterprise just do sudo yum install ack

How to run "dpkg-reconfigure ca-certificates" noninteractively?

I would like default to "accept" new certificate. I have tried the following.
$ dpkg-reconfigure -f noninteractive ca-certificates
It runs, but the CA is not added.
If not possible, is it OK to directly modify /etc/ca-certificates.conf?
Thank you for the help.
no this won't work out of the box.
Based on http://blog.sandipb.net/2009/08/08/adding-new-ca-certificates-in-ubuntu-jaunty/, you can do the following sequence that doesn't need interactive selection:
sudo cp my.cert /usr/local/share/ca-certificates/
sudo update-ca-certificates
tested and works on debian 5/6.
Actually, you can specify minimum priority question to show using param -p with valid priorities (low medium high critical). If you set it to critical, should work fine for your scenario.
sudo dpkg-reconfigure -p critical ca-certificates

How to set the rpmbuild destination folder

I noticed rpmbuild (-bb and --buildroot options) creates the .rpm in different locations depending of what OS are you using:
GNU/Linux Ubuntu <= 9.04: /usr/src/rpm/...
GNU/Linux Ubuntu >= 9.10: /home/rpmbuild/...
GNU/Linux Fedora: /usr/src/redhat/...
So how can I set manually the destination folder for all OS?
Replying myself, adding:
%define _rpmdir /outputdir
to .spec file.
You may want to use the command argument --define : For example, this will send the rpm files into the current directory.
rpmbuild anything.spec --bb --define "_rpmdir $(pwd)"
This will send the rpmsto output dir
rpmbuild anything.spec --bb --define "_rpmdir /outputdir"
Or perhaps something more complicated such as Custom gradle task for rpmbuild .
For me the solution was to set _topdir to a subdirectory called rpmbuild inside my project. I also set a few macros, to control the .rpm path:
rpm:
rpmbuild -bb package.spec -D "_topdir $(shell pwd)/rpmbuild" -D "SRC $(shell pwd)" -D "NAME $(NAME)" -D "ARCH $(ARCH)" -D "VERSION $(VERSION)" -D "RELEASE $(RELEASE)"
mv rpmbuild/RPMS/$(ARCH)/$(NAME)-$(VERSION)-$(RELEASE).$(ARCH).rpm ./
rm -rf rpmbuild
Then I use make to generate the .rpm file and move it. I prefer rpmbuild to operate inside its directory.
make rpm
package.spec is using these macros, e.g.
Name: %{NAME_PACKAGE}
Setting up the rpmbuild environment in /home/ http://www.linuxquestions.org/questions/linux-software-2/need-rpm-package-for-php-version-5-2-7-and-up-on-redhat-5-1-a-766486/#13
Actually it is not recommended to use /usr/src/**. I.e. by using /home/[name]/rpms/ you can work as unprivileged user. No su or sudo is required to build packages.
Just a tiny comment..
adding "%define _rpmdir /outputdir" to spec file is also support by rpmbuild in AIX OS