I've been trying for quite a while now to get the following thing to work - without success. I know that my approach is a bit of a dirty hack, so I'm always open for suggestions about how to do this in a better way.
We're running VirtualBox on Linux machines in a school environment. There are a couple of applications that absolutely need to be run under Windows, which is why there's no way around VirtualBox (forget wine!). So we put a Virtual Machine image on every local hard drive which the students can run if they need to. This image really ought to be "read only" in one way or another. Obviously, even running the machine will make changes to that image, so we need a mechanism that automatically reverts those changes. We're somewhat flexible about when this is supposed to happen. To me, doing this at each reboot seems to be the best approach but I wouldn't mind it being done at, say, logoff.
Right now we're using rsync on the image every time the computer boots which works as long as the virtual machine isn't started until that process is finished. Despite the fact that it works, it's a pain for the administrator as it may lead to various kinds of errors that are hard to reproduce, so there must be a better way.
My idea was to use a snapshot which I automatically revert to. Since simply deleting the Snapshot VDIs won't work, I wanted to create a "template snapshot vdi" (around 180MB) which is copied to a world writeble location (e.g. /tmp) at boot time. I know there's still a race condition but copying 180MB on a local hard drive should be significantly faster and predictable than rsyncing some 15GB. I then configure the virtual machine to use that snapshot.
Doing this VirtualBox produces error
Parent UUID {00000000-0000-0000-0000-000000000000} doesn't match UUID {12345678-1234-1234-1234-123456789012} in the xml file.
A possible cause is this issue with VirtualBox, though I'm not sure about this.
What came to my mind right now while writing those lines is using a script to copy/rsync the snapshot VDI template to the "right" path within the home directory just before the machine is being started by that script.
Are there any other suggestions?
Related
One of my hosts runs on Mac OS Catalina, and it constantly runs out of disk space...
I have scheduled tasks running there and every day it uploads files into /Users/labuser/myfolder and removes older files from that folder.
After digging through folders I found that /System/Volumes/Data/Users/labuser/myfolder takes 90% of occupied space on my host.
Is there a way to disable this feature on Catalina and stop it from growing /System/Volumes/Data/... ?
/Users/labuser/myfolder is equivalent to the folder with /System/Volumes/Data/ prepended. macOS 10.15 Catalina added firmlinks (some more description here), but actually from a practical perspective (to the user) these are one and the same.
Thus, your problem has nothing to do with a "feature" on Catalina; rather it has to do with the amount of data you're storing and backing up from /Users/labuser/myfolder.
Whether you use ncdu or another disk usage manager that will solve you problem of finding out why you're consuming all of your disk space.
One other relevant point is that because these are "symlinked" (called firmlinks by Apple), some disk inventory apps don't know how to handle this and end up in a recursion scenario when trying to understand total disk usage. I've seen this behavior with ncdu also. That being said, if you run the disk inventory on a subfolder of /System/Volumes/Data/, e.g.:
cd /Users
ncdu
It should avoid these issues.
(i posted this question on VMware community forums, with no response)
All of my developers use VMware workstation 11 for running their Visual Studio development environments in Windows. I originally created a base VM, and then simply shared it around to all the users, unbeknown to all the issues it would cause with computer name conflicts, joining the domain, etc. Apparently I didn't 'sysprep' it. Someone with some VMware knowledge tried to help me out to set it up, but im still not sure if it is 100% correct.
I am now at a point where I need to make changes/updates to the base VM (new VS, run Windows Updates, etc) and distribute a new one around to all the developers again, and I need some help to make sure I do it properly this time:
I have 2 VM images currently: 1) a ‘base’ one that opens up to this screen http://screencast.com/t/tBO0IUvwtpbY and 2) one that opens up to this screen http://screencast.com/t/DLV71PeNygrk . I’m assuming it is the ‘base’ one that I make the ‘edits’ to?
If I do edit the ‘base’ VM, what option(s) should I choose in that sys prep dialog?
Once I make the necessary updates to the base VM, what do I need to do to 'prepare' it for distribution to my developers (do I need to run sys prep, do I need to make a copy of VM beforehand, etc)?
Currently the VM has 2 HD's: 1) used for the OS and applications (Persistent and non-Independent), and 2) used to store data files (Persistent and Independent) . Is this the best way to set it up and use the VM (and then use 'snapshots' to revert back to previous states), or should the OS disk be set up as 'non-persistent' after the user has set up the VM for first use (i.e. joined the domain, set up VS Workspace, etc) ? Basically we have had a few scenarios recently where the OS and VS have been corrupted and it took the devs a while to fix it, so I want them to have some sort of way to revert back to a known clean working state.
Thank you
In answer to your questions you are on the right track.
the first screen that opens is correct, customise everything you require in the OS then run the OOBE before distribution
Choose OOBE and tick Generalise, then choose shutdown.
Once the VM is shutdown export it and pass it round your developers. they will need to complete the wizard and then join to the domain if required, this will ensure all the sids are different on each server, remember they must still change the Servers name before joining the domain
Snapshots are best, if your dev needs to reboot etc he doesn't want to loose his changes, he then has the option to revert to a snapshot.
Hopefully this has answered your question, let me know if there is anything else you need :)
I'm currently working on trying to resolve a crash/exception on an unmanaged C++ application.
The application crashes with some predicatibility. The program basically
process a high volume of files combined with running a bunch of queries through
the access DB.
It's definitely occuring during a file access. The error message is:
"failed reading. Network name is no longer available."
It always seems to be crashing in the same lower level file access code.
It's doing a lower level library Seek(), then a Read(). The exception occurs
during the read.
To further complicate things, we can only get the errors to occur when
we're running an disk balancing utility. The utility essentially examines file
access history and moves more frequently/recently used files to faster storage retrieval
while files that are used less frequently are moved to a slower retrieval area. I don't fully
understand the architecture of the this particular storage device,
but essentially it's got an area for "fast" retrieval and one for "archived/slower."
The issues are more easily/predicably reproducible when the utility app is started and
stopped several times. According to the disk manufacturer, we should be able to run
the utility in the background without effecting the behaviour of the client's main application.
Any suggestions how to proceed here? There are theories floating around here that it's somehow related to latency on the storage device. Is there a way to prove/disprove that? We've written a small sample app that basically goes out accesses/reads a whole mess of files on the drive. We've (so far) been unable to reproduce the issue even running with SmartPools. My thought is to try push the latency theory is to have multiple apps basically reading volumes of files from disk while running the utility application.
The memory usage and CPU usage do not look out of line in the Task Manager.
Thoughts? This is turning into a bit of a hairball.
Thanks,
JohnB
Grab your debug binaries.
Setup Application Verifier and add your application to its list.
Hopefully wait for a crash dumb.
Put that through WinDBG.
Try command: !avrf
See what you get....
If I already have an image that exists, can I create an image based on the existing one, except I want to make changes to the exisiting one (mainly configurations).
I do this all the time. I actually keep each of my VMs in a separate directory and duplicate the entire directory to make a copy. All references within the VMX file (configuration) are relative to the current directory.
One thing you need to watch out for. The VMX file has a line with the MAC address of the virtual network card:
ethernet0.generatedAddress = "00:0c:29:ff:1f:c7"
You'll need to change that if you want to run both VMs at the same time - I usually just bump the final digit up by 1 (to c8).
I also change the displayName in that file so I can tell the difference between them when they're running.
Yes, you can just copy the image off to external storage. Just find the image file(s) on your drive and do the copy when the image is not running. You can then change the original all you want. Is this what you are after?
What I do is create a base "clean" VM which I then run Sysprep on before cloning. You can run into a few problems when you don't reset the 'unique' elements of a windows installation and you're trying to run them simultaneously.
I'm running ~20 VMs at the moment and if any one gets seriously messed up (they're used for testing) I've got clean base images of Windows 2000, Windows XP, Vista and Server 2003 at the ready so I can be back up and running in 20mins or less.
Depending on what your needs are, you might try the (free) VMWare Converter. It lets you change drive sizes and other image parameters.
As others have said, this is exactly how you implement full backups for your VMs.
When the VM is not running, merely copy the virtual disks into a different location, then restart the VM.
I'm trying to read variables from memory. Variables, that doesn't belong to my own program. For instance, let's say I have this Adobe Shockwave (.dcr) application running in browser and I want to read different variables from it. How it's being done? Do I need to hook the process? But it's running under virtual machine, so I don't know how to do it.
This task is pretty much trivial in normal w32 applications (as it is mainly just
CBT-hooking / subclassing), but as I mentioned before, I've got no idea how it's being
done with flash / shockwave.
I'm using C++ (VS9) as my development-environment, in case you wish to know.
Any hints would be highly appreciated, so thank you in advance.
Best regards,
nhaa123
If you're trying to do it manually just for one or two experiments, it's easy.
Try a tool like Cheat engine which is like a free and quick and simple process peeker. Basically it scans the process's memory space for given key values. You can then filter those initial search hits later as well. You can also change those values you do find, live. The link above shows a quick example of using it to find a score or money value in a game, and editing it live as the game runs.
without having debug Binaries/DLLs of the Apps, your only chance is asking some hackers.
Normally you can connect to a process with a debugger, but without the debugging symbols of the binaries you don't see any variable names - just memory addresses.
Further the Flash/Shockwave code runs inside a sandbox inside the browser to prevent security holes by manipulated Flash code. So you don't have a real chance to get access to the running Flash code / to the plugin executing the Flash code - except you have a manipulated version of such a plugin.
So your task is quite hard to solve without using less legal methods. The next hard thing is the virtual machine - this could be solved by implementing your app as a client/server solution, where the "inspector" / watchdog runs as server inside the virtual machine and the client requesting the variable status/content running on your normal host. The communication could be done as simple socket connection.
If you have the chance to write your own Flash/Shockwave plugin, you maybe could be able to see contents of variables.
Sorry, that I cannot help you any further.
ciao,
3DH