For instance, I have a virtualised Linux-box running in my windows environment, and I have a shared-folder mapped to /vagrant and in that shared folder I have readme.txt. Is there a command I can run from within the virtualised OS to pop up readme.txt in my favourite text-editor in the windows host environment?
In a nutshell I'm just wondering if there is a standard way to invoke a host command from the guest.
I'm thinking an analogue of the cygstart command that cygwin provides, perhaps a part of the guest additions package.
This seems like it should be a well trodden path, since its highly likely a guest might want to communicate with its host in certain configurations.
Looks like somebody wrote a tool for this https://github.com/marcosdiez/openfile
Unfortunately it looks like there isn't a VirtualBox API for this so he has to run a server in the host:
it is necessary to run a C# server on Windows which accepts and runs
remote commands from the Unix box.
Should do the job though!
Related
I have a very strange problem with Cygwin sshd and changing the PATH environment variable through the Windows control panel. This is with Windows 10 64 running in a Parallels VM
After a reboot ssh sessions into the Windows machine will still use the old path. Local Cygwin sessions have no such problem: they will use the new path.
Note the bold text: I am aware that without the reboot this is expected to happen (because sshd got started with the old environment). But the After the reboot has got me baffled. If I ssh into the machine I see the old PATH. If I then reboot the VM again and ssh in again I finally see the new path.
Incidentally (fodder for future people googling this) I had the problem while using gitlab-runner to run CI/CD jobs on Windows using Parallels VMs on my Mac. So I would prepare the VM to have all the right tools installed and everything set correctly, then shut it down. gitlab-runner will then clone the VM and run the CI/CD jobs on it. Works now, as long as I reboot the original VM twice with an ssh session in between before having it cloned by gitlab-runner:-)
I never posted the answer I found: the problem was that hibernate was enabled on Windows. Actually the first "restart" was usually a shutdown followed by a startup.
But, by default Win10 has hibernate enabled, and if that is the case a shutdown and startup don't actually reboot the OS.
Disabling hibernate fixed it.
Double-check your Cygwin sshd installation, as described in "Installing Cygwin and Starting the SSH Daemon"
it makes sure the %PATH% does not reference other SSH, like W10 OpenSSH.
it stops any other SSH service.
it defines a Cygwin SSH Windows service, using a local account, which then should pick up the same account new path after a single reboot (or even with restarting the service, without reboot)
I am not sure why you need two reboots in your case, but see if that Cygwin sshd setup works better.
I'm trying to test my development environment and keep getting this error while going through the "Building Your First Network". I've installed the prerequisites on a Windows 10 Home edition so I'm using Docker Toolbox. The error is described as follows and happens just after a execute the ./byfn.sh up -l node command:
2019-11-14 17:06:26.982 UTC [msp] loadCertificateAt -> INFO 002 Failed loading ClientOU certificate at [/var/hyperledger/orderer/msp/cacerts\ca.example.com-cert.pem]: [could not read file /var/hyperledger/orderer/msp/cacerts\ca.example.com-cert.pem: open /var/hyperledger/orderer/msp/cacerts\ca.example.com-cert.pem: no such file or directory]
As you can see there is a backslash just before the ca.example.com-cer.pem which I think is causing the problem. Anyone have an idea of how I might solve this issue?
Docker Toolbox for windows doesn't provide a great experience for users. It's limited in how it can share windows directories with linux containers running in the linux VM it creates. I believe it can be made to work but you have to use very specific folders in your home directory and you would have to google it to find all the details but it's a lot more of a hassle.
You have other options of course
Upgrade to Windows Pro and use Docker for Windows (which doesn't have the same sharing limitations, but still not a great experience and you can still encounter problems sharing the directories)
Install a hypervisor such as virtualbox and run a linux desktop distribution such as ubuntu or mint and use that instead (which I would recommend and would suggest ubuntu 18.04 mate edition)
Wait until next year when hopefully microsoft will release WSL2 for everyone (including home edition) and then Docker won't need a linux VM running and it should be a much better experience trying to run linux containers on windows.
As mentioned by lindluni here
Backslash problem in Hyperledger 2.0.1 when orderer/peer look for .pem files
the problem arises when one creates crypto artifacts under Windows since "golang filepath.Join uses the current OS's filepath scheme". Hence, under Windows backslash is used in all the generated config.yaml files found in crypto-config/*. One can replace this backslashes before the deployment to fix the issue.
I've two Windows machines and both have bash installed. However, they differ in the way bash starts up:
Machine 1:
Typing "bash" starts the bash. Then I can type commands like ls.
Machine 2:
It seems like it starts bash and directly runs ls in it, by only typing "ls". After that it switches back to CMD automatically, like this:
Does anyone know which setting enables the behavior of machine 2? Everything looks the same for me. It's a nice feature and I want to enable it on machine 1 as well.
Bash on Ubuntu on Windows executables (binaries) cannot run from Windows applications such as cmd.exe or PowerShell.exe - Windows doesn't even see them as executable.
The likeliest explanation is that you've installed a separate Unix emulation environment such as GnuWin, which comes with native Windows binaries.
To see the location of your - by definition Windows-native - ls executable, run where ls, which will probably tell you what product it came with, such as
C:\Program Files (x86)\GnuWin32\bin\ls.EXE.
Note that the Ubuntu on Windows binaries are stored in a user-specific manner in
%LOCALAPPDATA%\lxss\rootfs\bin, but that is a moot point, given that you cannot invoke them from Windows.
In Windows 10, there is a built-in Linux subsystem (one of the greatest features of Windows 10). It gives you almost a complete Linux shell for various distributions, and you can almost do anything with it (user mode).
I bet it is installed on your first computer, and you are using that subsystem. In the latest version of this subsystem, you can run both Linux executables and Windows exe files.
We are having a linux build server and the build scripts are written in ruby. All the jars are deployed on linux servers only and are generally copied to the application servers using ssh.
There is a new service which needs to be deployed on a windows server. Is there a way to copy and configure the jar deployment from the linux machine to the windows server.
Using ruby / shell scripting?
If you are already using SSH in your build script I would think the easiest thing to do would be to install an SSH server on the Windows box. You can use sshd via Cygwin or opt for a commercial product like WinSSHD (I'm sure there are numerous other free or paid products out there).
Once you have that running you can more than likely reuse most of the existing script relating to moving files around and use a gem like Win32Utils to do any windows specific tasks.
What is the best way to edit remote files on a Windows box (XP) from Linux?
I am running emacs 23.3 locally under Linux.
I would like to keep the extra software on the windows-box minimal, but I can install some extra software if needed. Cygwin is already there.
If cygwin is already there, running a sshd in the windows machine and using Tramp (maybe using the scp method) is the best option to my knowledge.