spirals-team/docker-machine-driver-g5k

Docker Machine driver for the Grid'5000 testbed infrastructure.

docker-machine-driver-g5k's Issues

More options for reservation

It would be useful to specify more options when we create a machine, like the clusters we want to reserve.

Support auto-generation of SSH keys

The SSH key generation should be automated instead of requesting users to manage it.
This could increase the security by generating ephemeral keys, and simplify the use of the driver by removing frequent errors due to the passphrase or location of the public key.

Support standard environment reuse

The Grid'5000 nodes are already deployed with a standard environment (Debian 9).
Currently the driver deploy the OS at every machine creation.
Skipping the deployment of the image and provisionning this environment directly would save ~10 minutes per machine.
This enhancement depends on the #22 task completion.

Run docker from the frontend

Hello,

I'd like to run my experiments from the frontend. docker-machine is installed on the frontend but docker is not! I can provision machines but I cannot run docker containers on them. Users cannot install new softwares on the frontend, so, is there a solution?

Thanks,

Error during provisioning: Docker service fail to start

Due to a bug in Docker machine (docker/machine#4156), the newest version of Docker Engine (v17.06) fail to start after provisioning and you will get the following error message:

[...]
Setting Docker configuration on the remote daemon...
Error creating machine: Error running provisioning: ssh command error:
command : sudo systemctl -f start docker
err     : exit status 1
output  : Job for docker.service failed because the control process exited with error code. See "systemctl status docker.service" and "journalctl -xe" for details.

This bug is already fixed in Docker machine release v0.12.1, please update accordingly.

Support walltime extension

Grid'5000 updated to OAR 2.5.8 supporting walltime extension.
Currently this is not supported, the command sleep submitted to OAR wait for the initial walltime.

Support Debian 9 standard environment

The Grid'5000 team provides Debian 9 (stretch) images for kadeploy.
Currently the driver deploy Debian 8 (jessie) images.
The driver should deploy the latest environment image.
This enhancement requires Docker Machine with the netstat patch (> v0.13.0, currently unreleased).

EOF bug in mac

With the version 2.1 after the creation of the job and when i try to access to the machine it will generate an EOF error

x509: certificate signed by unknown authority

Hello,

I have been using this driver for "long" time and everything is fine ! But from time to time I got this error message "Post https://api.grid5000.fr/stable/sites/lyon/jobs: x509: certificate signed by unknown authority". If I repeated the action, it could disappear. This error could be really annoying if it happens inside a script.

I'm not sure if it is a problem related to the g5k-driver or g5k itself or my machine !

I'm running MacOSX, and I'm using Tunnelblick as VPN client.

Thanks in advance for any help.

Error with pre-create check: "Connection to frontend of 'luxembourg' site failed using docker-machine on Grid5000

Hi everyone,
I am trying to use docker-machine for the grid5000.
When I do "sudo openvpn Grid5000_VPN.ovpn" I got this :

Mon Mar 18 13:46:58 2019 OpenVPN 2.4.4 x86_64-pc-linux-gnu [SSL (OpenSSL)] [LZO] [LZ4] [EPOLL] [PKCS11] [MH/PKTINFO] [AEAD] built on Sep  5 2018
Mon Mar 18 13:46:58 2019 library versions: OpenSSL 1.1.0g  2 Nov 2017, LZO 2.08
Mon Mar 18 13:46:58 2019 WARNING: No server certificate verification method has been enabled.  See http://openvpn.net/howto.html#mitm for more info.
Mon Mar 18 13:46:58 2019 NOTE: the current --script-security setting may allow this configuration to call user-defined scripts
Mon Mar 18 13:46:58 2019 TCP/UDP: Preserving recently used remote address: [AF_INET]194.254.60.36:443
Mon Mar 18 13:46:58 2019 Attempting to establish TCP connection with [AF_INET]194.254.60.36:443 [nonblock]
Mon Mar 18 13:46:59 2019 TCP connection established with [AF_INET]194.254.60.36:443
Mon Mar 18 13:46:59 2019 TCP_CLIENT link local: (not bound)
Mon Mar 18 13:46:59 2019 TCP_CLIENT link remote: [AF_INET]194.254.60.36:443
Mon Mar 18 13:46:59 2019 [vpn.grid5000.fr] Peer Connection Initiated with [AF_INET]194.254.60.36:443
Mon Mar 18 13:47:00 2019 TUN/TAP device tun0 opened
Mon Mar 18 13:47:00 2019 do_ifconfig, tt->did_ifconfig_ipv6_setup=0
Mon Mar 18 13:47:00 2019 /sbin/ip link set dev tun0 up mtu 1500
Mon Mar 18 13:47:00 2019 /sbin/ip addr add dev tun0 172.20.64.6/18 broadcast 172.20.127.255
Mon Mar 18 13:47:00 2019 /etc/openvpn/update-resolv-conf tun0 1500 1554 172.20.64.6 255.255.192.0 init
RTNETLINK answers: File exists
Mon Mar 18 13:47:00 2019 ERROR: Linux route add command failed: external program exited with error status: 2
Mon Mar 18 13:47:00 2019 WARNING: this configuration may cache passwords in memory -- use the auth-nocache option to prevent this
Mon Mar 18 13:47:00 2019 Initialization Sequence Completed

**But when I used the first an example given in the section "Usage examples", of course I changed the username and the password to correspond to mine ** : https://github.com/Spirals-Team/docker-machine-driver-g5k
I got this :

Running pre-create checks...
Error with pre-create check: Connection to frontend of 'luxembourg' site failed. Please check if the site is not undergoing maintenance and your VPN client is connected and properly configured (see driver documentation for more information)".

I tried the frontend 'lille', 'lyon' and all the rest I still have the same error.
Thanks in advance for your help