I'm working on a new feature for our product, a component of which has some fairly major security implications: it runs as a WCF service, and performs some highly-privileged actions. As such, I'd like that component to only be installed when the user requires it (and removed when it's no longer required), and for the installation to be accomplished programmatically.
I've looked into advertising (specifically, assigned installs), and it appears to be a good mechanism for accomplishing this. I'm not sure whether the installation of an advertised feature can be undone, however, which would be useful for removing the component when it's no longer required. I'm also not completely sure whether programmatic installation could be achieved with advertising: advertised COM classes seems to be discouraged.
I'm wondering, therefore, if a simple call out to msiexec to install/remove the component would be more sensible. The component could be packaged in a separate MSI, or as a separate feature of our main MSI.
What techniques have you used for installing features on demand?
I resolved this by installing all the necessary bits and pieces up front and then starting/disabling the WCF Service as required.
Related
we are developing an application with a Helper Tool - which is installed into the system using SMJobBless. This works as expected; but there is a caveat.
We do frequent automatic deployments - sometimes more than one per week. Everytime the Helper Tool version changes, we re-register it - causing a password prompt. These 2 factors would quickly become irritating to our users.
Is there a way to have the password prompt appear only once, during the initial Helper Tool installation? Could subsequent updates happen without a prompt? Perhaps there is a way to leverage the existing Helper Tool to install a newer version of itself?
Short answer: No. SMJobBless() always prompts for admin credentials. There's no way to stop it from prompting. If you call this API, it'll prompt (or fail).
Longer answer on workarounds:
If your helper tool is running with admin/root privileges, it could theoretically replace itself with a new version. Think very carefully before doing this. Getting this right and maintaining security is very difficult, and the fact that even the major OSes have had vulnerabilities in installer functionality is a strong indicator that the risks of going this route may outweigh the benefits.
If you must proceed, read up on:
Race Conditions, Secure File Operations, and Time of Check vs Time of Use
Apple's Security APIs, particularly SecRequirementCreateWithString and SecCodeCheckValidity.
macOS Code Signing In Depth and the Code Signing Requirement Language
You would have to ensure that your helper tool cannot be tricked into replacing itself with (or executing) malicious code, or you will have opened your software up to being a trivial root exploit vector.
Also note: Regardless of what Apple currently does to verify helper tools installed by SMJobBless, it is conceivable that they could tighten the requirements in the future and refuse to run helper tools that have been modified since they were installed via SMJobBless. The safest method (in multiple respects) is to just call SMJobBless whenever you need to install/update the helper.
I have two versions of the same MSI package: 'genuine' and 'developer'. The former is intended for shipping, whereas the latter is built on developers' computers for the sake of testing, etc. I want to prevent accidental leaking of the 'developer' version, so that the end user couldn't have it installed. What are the possible approaches?
The MSI is built with WiX, just in case.
I've seen Securing an msi to prevent unauthorized use, but the answers there are focused on preventing launch not installation. I don't need to obfuscate, encrypt or protect the package contents. I only need to protect unqualified users from accindentally screwing things on their machines, due to erroneous packages being installed.
Honestly, because MSI is an open book ( Edit with ORCA ), it's best to do this type of enforcement (DRM) in the application at runtime. I could suggest all sorts of techniques in the MSI but they could be removed by someone knowing what they are doing.
What is your "leaking" concern? Do you not trust your developers? Do you not trust your CM / Release practices? Do you have a strong NDA in place? Proper IT security?
If you are trying only to prevent accidental misuse - that is, someone unknowingly installing a developer copy who wanted a user copy - I would suggest a simple launch condition. You would document that "developers" must put something on their machine (perhaps a file, a registry key, or even an environment variable), and have the msi check for that. It would then refuse to install unless the "I am a developer" footprint was there.
If you want to avoid the footprint, you could require that "developers" launch the msi from the command line and pass DEVELOPER=1 and instead check for the property. Or more extreme you could require specific and changing values for the DEVELOPER property. It all depends on how annoying you want to make it to be a "developer" for your product.
Note that this is not secure in the slightest against malicious misuse. Among other limitations, it does nothing to prevent administrative installations that allow access to all the files.
Note as well that it's still a smart idea to follow the usual launch condition patterns to avoid preventing uninstallation of the package.
We have developed a software in vb.net using Visual Studio 2013. Now we want to build a custom installer with following steps/features:
User Start to install our software.
At 'Enter Serial Key' option, user enters the 16 digit Serial which have we provided.
When clicking 'OK' button, our software connect to our IP and save the Serial Key with some other user's information to our database.
A confirmation Key is returned back to our software.
Software writes a file and save it to the system folder.
It is almost like Adobe or Corel registration process.
We are open to other techniques also which must secure that our software must install on a single machine only.
Please be noted that we are a group of novice programmers(not so advance level), so; if the process is elaborated, it will be very helpful to us.
One-Shot Setups: "A setup is run once, an application can be started again - in order to resolve and debug problems interactively - with meaningful error messages show to the user."
Hence: avoid license validation in the setup.
Short version on licensing.
License Key: Preferring to deal with license keys in your application seems logical for several reasons: the one-shot nature of setups
yields poor reliability (no interactive debugging - poor ability
to resolve problems). The end result is lots of support calls for something very trivial.
Further, the risk of piracy and hacking is a major concern when
exposing a license validation DLL in the setup. And finally
communication over the Internet is difficult with today's setups (proxies, firewalls, etc...) - which is a modern way to validate license
keys (in the future setups might have full Internet access, but be careful assuming too much since corporate users may have lots of restrictions and poor deployment could hurt sales and acceptance of the software for corporate use).
Finally your application must usually support a trial version,
and then you need a license dialog in your application anyway.
Why complicate your setup too?
CAs: Custom actions are complex and vulnerable to failure in general - due to complex sequencing-, conditioning- and
impersonation issues and overall poor debugability. More information:
Why is it a good idea to limit the use of custom actions in my WiX / MSI setups?
Overall Complexity of Deployment: A short, attempted summary of the overall complexity of deployment:
Windows Installer and the creation of WiX
(section "The Complexity of Deployment").
I would remove all licensing features from the setup and add them to the application. Your setup can still write a license to disk or to the registry by passing it to msiexec.exe as a public property -
UPPERCASE properties (or you can "hide" things a little more by using a transform to apply the serial property - it has exactly the same effect as setting the property on the command line). You can also set the LICENSE property from a dialog in the setup when it is run interactively, but my favorite approach is to allow adding the license key unvalidated to the registry in silent deployment mode, and to instead enter the license key directly in the application, and not the setup, for interactive deployments (the above description is for silent deployment):
msiexec.exe /I "C:\Install.msi" /QN /L*V "C:\msilog.log" LICENSE="123-456-789"
This will allow the license to be easily added to each machine in a corporate deployment scenario. The license value is simply written to disk or registry without validation. The application will verify it (more secure than a validation dll in the setup).
There is no need to mess with any complex setup dialogs, but you will need a license dialog in your application as explained below.
As a setup developer you should offer to help implement the feature in the application instead of the setup so it doesn't seem like a case of "passing the buck". This is all for overall software reliability and foolproofness - and several reasons are listed below.
Almost all large corporations deploy MSI files silently, so the setup GUI will be ignored most of the time anyway. You are then simply adding risk and wasting resources if you deal with licenses in the setup.
One drawback: An application run as a non-admin user after installation can not write to HKLM to share a serial between all users on the computer (a setup running with elevated rights can). It must either write to HKCU or the setup must have prepared write access to a specific HKLM location in the registry for the application to write to. I prefer writing to HKCU for each user since the license is then less available for copying by others, and it is kept as user specific data (allows roaming, though that is a hated feature by most IT professionals). However, a HKLM license key written by the application or the setup during installation (as explained above with a public property set) allows all users to share a license when launching the application.
There are several more concrete reasons to keep license handling and validation out of your setup:
A significant number of support requests always result from people who have problems registering their license keys in the setup. A setup is run once, an application can be started again if there are problems. This is more important than you might think for inexperienced users. You also have better features available to handle exceptions and error conditions and whatever unexpected problems may occur in the application.
Serial validation in the setup exposes a validation dll / method that is easily cracked by pirates. You won't prevent piracy by eliminating it from your setup, but at least you make it more difficult. It is more secure in the application if you cloak things a bit (static linking, encryption, obfuscation, putting the validation process online, and / or whatever is done by security professionals that I am unfamiliar with).
Allow application trial version: If the setup needs to support a trial version of the application, you should allow the user to enter a license key if they end up buying the product - preferably without having to re-run the setup or uninstall / reinstall just to add the license key. In other words you will likely need to deal with licensing in your application anyway, why complicate your setup too? More risk, more QA, more potential support requests and potential for multiple required fixes in both setup and application. High total cost?
If your application runs with different editions, what if the user buys an upgraded license? They should just be able to enter it into the license dialog and unlock features if possible and not uninstall and reinstall with all the clunk that involves. For some upgrades this is hard to achieve though, and you often end up with separate setups for different editions.
If the network is using a proxy server for Internet access, you will have problems registering the license over the Internet during the setup (often asked for by marketing). You have more features to check and deal with this in the application - it can try again and wait for access (generally you hook up to IE for automagic proxy configuration if possible). For corporate deployment you need a silent install option too which doesn't validate the key but just writes it to the registry. Trying to access the Internet from a silent install of an MSI is in my opinion a rather extreme deployment anti-pattern. I find it dubious in the setup GUI as well. Do the registration in the application - much less controversial, and you can set up firewall rules to allow it to access the Internet (msiexec.exe is likely blocked - and for good reason). There could also be hardware firewalls and / or security software to deal with that makes Internet access difficult or even impossible without some clunky admin server configuration. This could kill your software from consideration is my experience: "Just get this off our network and application estate - there must be better options - far too clunky and error prone".
UPDATE: As deployment technology matures and becomes more "Internet based" this "truth" may change, and we could end up doing everything "online" with deployment designed specifically to run via online "repositories" for example. We will have to wait and see. For now my opinion is that any setup Internet access requirements are erroneous and undesirable.
Setups that mess with licensing may sometimes cause license data to be deleted during upgrades, patching and migration scenarios due to bugs in the setup. This is a lot more serious at times than you would think - the package might hit thousands of workstations in large companies and be cumbersome to fix.
There is a rather bad "anti-pattern" in the MSI technology itself whereby self-repair or manually triggered repair will reset values in the registry that has been changed by the application. This can wipe out license keys. We see this all the time, and it is the technology's fault. It is just not logical in this area.
There are some fixes - or rather workarounds - for this (use a permanent component, write license from a custom action instead of from a component, etc...), but I find them quite clunky and you have to have a lot of experience to know all the pitfalls - and even experienced users mess this up.
Licensing is a huge corporate headache - often what is desired by a company or corporation is that licensing is centrally managed on a server, and not based on text serial numbers at all (for example concurrent or floating licenses acquired on application launch via the network). Just mentioning this though it is sort of out of scope for the question. In these cases what you specify during installation is generally an IP address pointing to the license server, or just a regular host name to be resolved by WINS or DNS.
As you might have already guessed Windows Installer doesn't provide any out the box feature to handle licencing. But there are commercial licencing solutions which you can go for if affordable.
LogicNP
DESAWARE
Since this is very broad question difficult to explain low level implementation details. I can give you a direction.
First of all you will need a custom UI where user can type in the License/Activation key. There are ways to incorporate a custom UI into windows installer, I have already explained few approaches in SO, refer to the following threads.
Show custom Form during installation
How to add additional custom window to VS setup projects UI flow
By following above approaches you should be able to add a UI where user will type in a key. Once user added the key, he will press Activate button on the custom UI, Button click event handler will invoke the necessary logic to Insert/Validate the activation key entered by the user.
Maybe you could try Inno Setup which is free (even open source) installation system.
It is script based which allows you to tune your installer and perform in it everything!
Creating custom page with entering serial number is really easy, see this example: CustomPage for Serial Number in Inno Setup
and there is also integration for Visual Studio.
I have an application which consists of 2 web sites, 3 Windows services (2 of them hosts WCF services) and 3 SQL databases.
It's likely that all system will be installed on 1 or 2 servers, but also possible that some services will be installed on separate machines if performance will be seriously affected by them.
I need to create WIX installer for all this staff. The thing is I'm not sure if it would be better to create single complicated installer which will be able to install whatever component set (some services may be installe), or few separate installers.
Single-installer approach has important advantage if the whole product is installed on the same machine - we will need only few dialogs shown to user and we will be able to setup connection configurations between components automatically. But of course development of such installer is very complicated: installer wizard pages should be shown based on features, installing currently on machine. There will be multiple conditions and testing all them will be a hell.
Multiple-installer approach is easier to implement and test. But of course it will be harder for user to install the product - he needs to set same service installation URLs few times (once per installer) - to setup connection configurations to WCF services.
I searched for best-practices in this area but looks like there is not a lot of information. Can anybody suggest which way is more appropriate?
If you design it correctly you can switch back and forth between the two approaches. Personally I've used InstallShield which has a feature called RELEASE flags and Product Configurations. It's basically a way of defining multiple builds that pull in different features and apply different branding (ProductName, ProductVersion, ProductCode, UpgradeCode, INSTALLDIR and so on.) This is how you'd make say a multiple single product installations and then build it again as a product family suite. This pattern applies to SOA very much. The same thing could be done in WiX using preprocessor statements to control compilation.
I once worked for a company that made an SOA based server product. The system had 26 installers of 5 catagories ( Win Client, Mobile Client, Service Layer, Web UI layer and SSRS Reports) I did each as it's own installer and it really helped cut down on the complexity. Each service had it's own database which made writing the SQL scripts to create and update the DB a lot easier. It also cut down on the amount of configuration each installation had to do since it only had to know how to set itself up and the other piece it communicated with.
The bootstrapper to bring all these pieces together may or may not be needed. Part of the reason we choose SOA is to give us complete control on how to scale the system out. This type of flexibility (26 pieces on 1 server, 1 piece on 26 servers, 2 tiers, 3 tiers, n tiers ) makes it difficult to encapsulate through a simple wizard UI.
I would go with multiple installers, bundled into a big single installer, this allows you to easily separate and test them. The bundle installer has the purpose of capturing the data from the user a single time and pass it to all installers selected for installation.
You can do this using Burn support from Wix or other setup authoring tools (that provide you with a GUI, allowing you create the projects and dialogs much easily). Wix is a very good tool, but the time you need to create the projects is too much, compared with the costs of a license for a good setup authoring tool that provides and drag and drop IDE to design dialogs and chain the installation of the separate installers.
EDIT
To pass values from the bootstrapper to the background installers is very easy, you don't need to write any custom code. This is handled directly from the setup authoring tool. All you need to do is to make sure the tool is capable to create Windows Installer compliant packages. This way, you can pass the selections made from the user to the background installers as property values, in the command lines of the installers.
The bootstrapper installer should not register with Windows Installer, so it does not appear in Control Panel, as a separate applications, there you should see only the separate background installers.
I'm providing a second answer to give a different perspective from my first. Microsoft Team Foundation Server installation is 1 single installer for different roles. You can scale out the various layers of the system but the install is all the same: install it all.
What they have chosen instead is to invest in a role configuration wizard that gets executed post installation to activate a given installed feature and configure it's settings and that of the next layer it needs to communicate with.
This is a very good way to go also.
There was a challenging situation happened when i was working with install to provide product key validation. I had to use C++ unmanaged code to validate the key. Actually we had the main validation logic written in C# and I had to create a mixed project. Problem was not stopped only with these, it continued. Since I used VC++ code, it expected atleast the VC++ runtime redistributable to be installed in the client machine. I thought of dropping the plan to migrate our install to Wix because of these kind of problems.
But I came to know that there is a nice and very cool feature that DTF is available in Wix to integrate any kind of actions in C#. I used it and could integrate the key validation in couple of hours and till now it is working fine in all client machine I implemented before 6 month.
Do you have any interesting moment or nice experience with DTF?
Search my blog at http://blog.deploymentengineering.com for DTF and you'll find a lot of useful content. I love DTF but I still believe that the best solution is to avoid a CA whenever possible in the first place. C#, like VBScript before it, is so luring that it tends to suck imperative thinking developers into writing CAs when not needed. I believe this is the reason DTF wasn't released for so long.
At my day job my approval is required for anyone who believes they need a CA. I instruct the developers on basic MSI philosphy, how to use DTF, how to attach a debugger and I make it clear that they are on the hook if it ever has any issues. The result is very few but well written CAs in our product line.
I have written several .NET CAs to support our WiX based installs:
Managed Wrapper around HTTPAPI.DLL - supports creating IP/Port SSL bindings and HTTP Url ACLs for use in deploying WCF services. I plan to turn this one into a Wix Extension. It was very interesting learning how to properly handle rollbacks, etc.
SSL Picker dialog that displays all the SSL certificates on the system and allows you to pick one.
SQL Server browser dialog - lets you browse your network for SQL Servers and then browse SQL Servers for Databases. Optionally uses impersonation. This is for crafting a connection string.
I am in the process of writing a set of CAs that will use the Microsoft.Web.Administration assembly to do native installs of web applications on IIS 7 (without requiring the IIS 6 Metabase Compatibilty feature be installed).
First off, the C#/DTF custom actions are still custom actions (no magic here :-)), so you should follow all the various CA guidelines working with this kind as well. It simplifies most of MSI tasks by abstracting low-level API behind the high level well-designed classes. Also, keep in mind that you can use managed code CA only in case the target machine has .NET installed (or install it as a prerequisite). Finally, the dtf.chm documentation which is distributed along with WiX toolset has some simple, but self-explanatory examples.
Hope this helps.