Crave is a command line interface for running commands against Crave platform clusters.
Crave looks for a crave.conf
configuration file in the current working directory or any of its parent directories. If it does not find the file there, it will search the user's ${HOME}
directory for the configuration file. You can specify the crave.conf to be used by setting it in the command line using the -c
(or --configFile
) parameter.
The latest stable version of crave
can be download from the crave GitHub page. Pick the binary for your platform. For Windows users, download and unzip the contents into a folder.
For Mac: brew install --cask accupara/crave/crave
Alternatively, your administrator approved version of crave can be downloaded from the "Downloads" page in the Crave UI.
When running crave, you might be asked to give permissions to a program called chisel
. Please allow this program to run and accept connections.
Crave follows the following syntax that users can use when running commands from a terminal window:
crave [crave options] [command] [command options]
where:
crave options
are options for crave itself. For example, --quiet
or the -c
flags.command
are the crave actions. Examples are run
, ssh
, list
, pull
, etc.command options
are the options for the crave command being run.The following is a list of the top-level crave options.
-h, --help show this help message and exit
-c CONFIGFILE, --configFile CONFIGFILE
Config file to use for the command. If not provided,
this file is searched in the parent directories and at
$HOME
-q, --quiet Silence HTTPS warnings. Also do not show progress bar
when downloading artifcts.
-v, --verbose print out verbose messages
Crave supports the following commands.
$ crave --help
usage: crave-0.2-6591 [-h] [-c CONFIGFILE] [-q] [-v] [-n] {version,list,set,stop,run,exec,ssh,pull,push,fullbuild,localrun,discard,lookbusy,getlog,fetchdiagnostics,devspace,clone} ...
Crave Build invoker
positional arguments:
{version,list,set,stop,run,exec,ssh,pull,push,fullbuild,localrun,discard,lookbusy,getlog,fetchdiagnostics,devspace,clone}
sub-command help
version Prints crave client version
list List my projects, running builds and ssh sessions
set Set config options
stop Stop a build or interactive session.
run Run build command on crave servers
exec Runs additional commands in current job
ssh Start an SSH session on the remote workspace
pull Pull job artifacts from build workspace to local disk
push push job artifacts to build workspace from local disk
fullbuild Start a new sandbox build. For use with CI (e.g., Jenkins)
localrun Run build commands on local machine
discard discard a user workspace
getlog Show logs of current (or previous) build
fetchdiagnostics Show system logs for build
devspace Start an SSH session on the remote devspace
clone Clone or destroy a project source into a devspace location
optional arguments:
-h, --help show this help message and exit
-c CONFIGFILE, --configFile CONFIGFILE
Config file to use for the command. If not provided, this file is searched in the parent directories and at $HOME/crave.conf
-q, --quiet Silence HTTPS warnings. Also do not show progress bar when downloading artifacts.
-v, --verbose print out verbose messages
-n, --noUpdate Do not check for crave updates
Prints the version of crave you are using.
$ crave version
crave 0.2-6591 darwin/x86_64
Lists all projects that you are a part of. It also lists running jobs (if any).
$ crave list
Configured Projects:
Id Name Source Url
---- ----------------- ------------------------------------------------------------------
4 Linux kernel https://github.com/torvalds/linux.git
16 docker-images https://github.com/accupara/docker-images.git
11 kubernetes git@github.com:kubernetes/kubernetes.git
12 linkerd git@github.com:linkerd/linkerd.git
2 postgreSQL https://github.com/postgres/postgres.git
15 protocolbuffers https://github.com/protocolbuffers/protobuf.git
Your jobs:
Job Id Project Name Job Status Local Workspace Job Url
-------- -------------- ------------ ---------------------------------------- --------------------------------------------------
5102 Linux kernel running /Volumes/case-sensitive/github.com-linux https://www.crave.io/app/#/build/info/5102?team=1
This mode runs the build commands provided on the command-line and preserves the workspace at the end of the build which enables users to pull the artifacts back to local disk after the build is complete. It also allows users to run incremental builds.
When executed without any further options, crave will run the default configured set of commands from the project configuration. If any commands are specified on the command-line, then crave will override the configured commands with these.
Crave will run this command on the workspace and stream back the stdout/stderr to the terminal until the command completes. The command can be terminated by hitting ctrl-c
. This will kill the command on the terminal as well as on the server.
$ crave run make clean
$ crave run -- make -j64
Picking up local changes (if any)...
Waiting for build to start...
using branch "master"
using base commit "67b9b3ca328392f9afc4e66fe03564f5fc87feff"
checking build system type... x86_64-pc-linux-gnu
checking host system type... x86_64-pc-linux-gnu
checking which template to use... linux
checking whether NLS is wanted... no
checking for default port number... 5432
checking for block size... 8kB
checking for segment size... 1GB
checking for WAL block size... 8kB
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
...
...
------------------------------------------------------------------------
Build Successful
Total time: 4m4.853300055s
------------------------------------------------------------------------
Build logs saved at: https://dev-server.crave.io/app/#/build/info/5757?team=1
$ crave run --help
usage: crave run [-h] [--projectID PROJECTID] [--no-artifacts] [--detached]
[--clean] [--no-patch] [--artifacts ARTIFACTS_LIST]
[--message MESSAGE] [--full-patch-stat]
[--disable-build-cache]
[cmd [cmd ...]]
positional arguments:
cmd Command to run
optional arguments:
-h, --help show this help message and exit
--projectID PROJECTID
Id of the project to build
--no-artifacts Do not download build artifacts when build completes
--detached Start build in background (do not stream
stdout/stderr)
--clean Start build on a clean workspace (causing a full
build)
--no-patch Run with remote commit id and branch without
generating a patch
--artifacts ARTIFACTS_LIST
Override the default artifacts downloaded with this
--message MESSAGE Set the given message as the message attribute for
this job
--full-patch-stat Show full patch stats when generating patches
--disable-build-cache
Disable using cache during build
--projectID
In normal cases, Crave tries to automatically guess the project configuration that should be used to run the command based on the Git URL of the project. However, when more than one project points to the same Git URL, crave cannot guess which project configuration should be used. In that case, it is required to provide a --projectID
option to crave. A default projectID option can be set using the crave set
command.
--clean
This option tells crave
to create a new clean workspace for running the build commands. This workspace is created from the snapshot of the sources.
--no-patch
This option tells crave to not generate a patch for the run. This will preserve the sources on the server as-is and will not modify any contents. Any patches that have already been uploaded and applied previously will stay applied.
--detached
This option forces crave to disconnect from the stream after starting the job. Crave will print the job URL where the job progress (stdout) can be viewed in the browser.
--no-artifacts
This option tells crave to not download any artifacts after the build completes.
--disable-build-cache
This option tells crave to not use the build cache during the build process. All commands will be run through the compiler and no acceleration is enabled.
This command provides an easy way for developers to log into the workspace console and inspect the contents of the workspace. By default, this command will drop the user into an ephemeral SSH session at the mount location of the sources.
Normally, this is a new SSH instance. If there already is a job running on this workspace, then the --current-job
parameter can be used to get access to the running instance of the job. This can be used to debug the running processes inside the build container.
This command tells crave to pull artifacts from the remote workspace. The changes are incrementally downloaded so that repeated downloads of the same file do not cause any additional downloads.
crave pull build/kernel/fs
Stops any running jobs on the current workspace.
$ crave stop --help
usage: crave stop [-h] [--projectID PROJECTID] [--force] [--ssh] [--all]
[ids [ids ...]]
positional arguments:
ids one or more job ids of the jobs to stop
optional arguments:
-h, --help show this help message and exit
--projectID PROJECTID
ID of the project for which jobs should be stopped
--force Force-stop all jobs queued or running on this
workspace
--ssh Stop all ssh sessions on this workspace
--all Stop all ssh sessions and jobs running on this
workspace
When a job ID is provided as a parameter, crave will only stop that job. To stop more than one jobs, the job IDs could be provided as a list. If a job ID is not specified, crave will stop any jobs running on the current workspace (for the current directory).
crave stop 5102
By default, crave stop
will only stop the running build jobs on this workspace. To stop SSH sessions, use the --ssh
option
crave stop --ssh
To stop all jobs and SSH sessions, use the --all
flag.
Crave allows you to proactively delete your workspace without waiting for the default workspace GC cycle to clean it up.
$ crave discard --help
usage: crave discard [-h] --current-workspace [--projectID PROJECTID] [-y]
optional arguments:
-h, --help show this help message and exit
--current-workspace, --current-ws
delete the current workspace
--projectID PROJECTID
Id of the project to build
-y, --yes Silence the confirmation question
crave discard
shows a confirmation question and success message.
$ crave discard --current-ws
This action will discard the current workspace. Continue? [Y/n] y
workspace deleted successfully
If the workspace does not exist, then crave discard
would show the following message
$ crave discard --current-ws
This action will discard the current workspace. Continue? [Y/n]
workspace has already been deleted
crave.yml
Crave supports yaml file crave.yaml
which allows users to
-- override certain project settings (such as docker image used for build and artifacts to be downloaded after build)
-- add user-specific files which are not a part of source repository
crave.yaml
does not need to be added to source repository for crave to use it.
Following are the fields for crave.yaml
image
This is used to override the Build Image
specified in Project Configuration
$ cat crave.yaml
Linux kernel:
image: "accupara/lkbuild@sha256:c31ec38936e30bce9ed7355bb428ab8173900c0c4e7b3f5ff626d195b0484d73"
artifacts
This can be used to override the Build Artifacts
specified in Project Configuration
$ cat crave.yaml
rsync:
artifacts: ["compat.o" ,"io.o"]
include_files
This is used to create patch for custom files in a user's workspace.
$ cat crave.yaml
protocolbuffers:
include_files:
- testFile
- Makefile.custom
workspace_per_branch
This is set to True
to ensure that same workspace is used across different branches.
If it is not set, unique workspaces are used for different branches.
$ cat crave.yaml
linkerd:
workspace_per_branch: True
env
This tag can be used to add custom environment variables to crave run
or to the SSH sessions.
$ cat crave.yaml
MyProject:
env:
key1: value1
key2: value2
Crave supports configuring multiple projects using the same crave.yaml
file
$ cat crave.yaml
Linux kernel:
image: "accupara/lkbuild@sha256:c31ec38936e30bce9ed7355bb428ab8173900c0c4e7b3f5ff626d195b0484d73"
rsync:
include_files:
- testfile1
- testfile2