Thunderbird 78: How to add security exception? - ssl

In previous versions of Thunderbird, when connecting to a server with a self-signed certificate, a warning was displayed, but the option was offered to create a security exception for that server with this dialog:
I could click on "Confirm Security Exception" and thereafter, TB could communicate with the server.
But in this version of TB, I only see a message:
Sending of message failed.
The certificate is not trusted because it is self-signed.
The configuration related to [server name] must be corrected.
There is no indication of what "must be corrected" for TB to trust the certificate.
The Certificate Manager in the Privacy settings section does open the dialog above. But once the correct IP:port is entered, the "Get Certificate" button does nothing apart from displaying "No Information Available". This renders the dialog pretty useless.
I'm not concerned about man in the middle attacks due to the lack of CA signing because the mail server is on an intranet.
Is there any other way to get TB 78 to work with self-signed certificates?

I also suffer from this issue for hours. Finally, I figure it out to solve it.
At first, I tried to download the certificate from Firefox. When I try to access https://bad-mail-admin.com:465 and it tells me:
This address uses a network port which is normally used for purposes
other than Web browsing. Firefox has canceled the request for your
protection.
Then I googled it and found the solution: https://support.mozilla.org/en-US/questions/1083282
Just go to about:config, right-click anywhere on the page and choose New > String. Create an option named network.security.ports.banned.override, with the value: 465
Then firefox show https://bad-mail-admin.com:465 successfully.
Immediately I realized that Firefox and TB use the same browser engine. So I repeat do it in "Config Editor" of TB.
Then go to Certificate Manager of TB, enter https://bad-mail-admin.com:465. Press "Get Certificate" and finally I can get the certificate and add it to the security exception.

only problem is that when the user has multiple email accounts that use different ports. That port override doesn't take multiple ports. I tried comma space colon semi-colon separated list with no success.

Related

Is the Plugin Registration Tool not recognizing SSL certificate?

Notice: this exact question can be found on the dynamics community forum which as usual isn't exactly responsive...
I can't figure out what's wrong with this environment...
CRM and ADFS are on the same server, different ports:
By browser, navigating to https://myorg.mydomain:444 redirects to https://sts1.mydomain:442 adfs login screen shows up, I input credentials, then I'm redirected back to CRM, everything works perfectly no matter which organization I navigate to. The SSL certificate is a wildcard one, covering *.mydomain (again, no issues whatsoever). Outlook client also works without a hitch.
My issue is, the registration tool (I'm using the one from the 2016 SDK, but this also happens with the 2013 SDK's one ) doesn't seem to be able to connect.
The exception message showing up in the log is (I'm translating from my native language to english, messages might not be 100% accurate)
[Top] Unable to establish a trust relationship for the SSL/TLS secure channel with authority 'sts1.mydomain'
[Inner level 1] Underlying connection closed: <same as above>
[Inner level 2] The remote certificate wasn't deemed valid from the validation procedure
Nothing in particular stands out in the Event Viewer... What's wrong ?
Just before writing this, I also tried the 2011 Registration Tool and it spits out a different error: it attempts to login to ADFS through HTTP instead of HTTPS (it complains about not finding http://sts1.mydomain:442 which doesn't exist).
I also tried importing the aforementioned SSL cert into my trusted root cert authorities, it doesn't seem to matter (everything stays the same).
Update: I forgot to show the connection settings:
(o) On-Premises ( ) Office 365
Server: myorg.mydomain
PORT: 444 [X] Use SSL
Authentication Source: IFD
Username: DOMAIN\USERNAME
Password: PASSWORD
Domain: <BLANK>
[X] Display list
I haven't started fiddling with plugins in 2016 so I'm not sure how that works and which endpoint it's using but I'd try the 2011 plugin registration tool too. It's what I've been using up until now and I think it is a better one than the 2013 and later since you can have multiple servers setup in it.
Regards

IdentityServer: The remote certificate is invalid according to the validation procedure

I'm trying to setup SSO using OWin and Thinktecture Identity Server but I am not having any luck getting the SSL certificate to work. At least I think that's the problem. All works fine when I'm in visual studio, but if I try to use IIS on my machine it gives me the error "The remote certificate is invalid according to the validation procedure". I've also tried using IIS as the client treat the instance running in visual studio as the token authority but I still get the same error. Anyone have any ideas on what I'm doing wrong?
In my case I was just trying to work through the samples (for ID3v2) and getting the cert errors running locally. Since some samples even do self hosting via owin I'm not even sure where it's getting the certs for host side??
Anyway my fix was to copy the cert to the Trusted Root:
Windows => Start => run MMC.EXE
File=> "Add/Remove Snap-In..." => Certificates
Use Computer Account => Local computer => Finish => Ok
Go under Personal / Certificates
Right click "localhost", Select Copy
Paste to "Trusted Root Certificate Authorities"
Done. Enjoy.
After spending a lot of time for me the solution was pretty simple
I just opened the Certmgr.msc ---> deleted the localhost certificate from the Trusted Root certification authorities.
Then opened my solution (after I had run the identity sever)
clicked run the visual studio asked fro me if I want generate new certificate to iis express (ssl),
I had clicked yes and then it started to work properly:)
You need to add whatever certificate IIS is using to your Trusted Root Certification Authorities store on your local computer.
That can be caused by bad configuration on a previous certifications (sometimes can happen when you disagree to install a certification) :
Windows Start and open Certmgr.msc
Under Personnel/Certificats, find all localhost certificats and delete them
Same thing to do under Trusted Root Certification, and then close.
Start your application, you will get an exception.
Open the Package Manager Console and excute: dotnet dev-certs https --trust
Restart your application, normally you have a valid certificat now.
Adding certificate to Trusted People store should be enough according to readme file in examples provided by the authors.
In a production scenario it should be better because Root store is for CAs and when you add something there that authority is not only trusted, but any certificate signed by it is automatically trusted.
You can check this an further details from the microsoft reference. An extract of the 2 store short description:
Root: Certificate store for trusted root certification authorities (CAs).
TrustedPeople: Certificate store for directly trusted people and resources.
P.S: I tested it an it works. In my scenario I have IS on machine A and a set of web applications using IS on machine A and B. IIS certificate on machine B is different from the one used in A and by IS, but I just added it on machine B Trusted People store and the "certificate error" disappeard.
Some times it doesn't work though the above settings were done and you have given the URL as "https://localhost", instead give the URL as "https://MachineName".
i.e machine name should match certificate's "issue to" value
For .Net Core change TrustServerCertificate=False to TrustServerCertificate=True and that will solve your problem like I have it below.
"DataConnect": "Server=tcp:127.0.0.1,1433;Initial Catalog=dbName;Persist Security Info=False;User ID=username;Password=password;MultipleActiveResultSets=False;Encrypt=True;TrustServerCertificate=True;Connection Timeout=30;"

Signtool error: No certificates were found that met all given criteria with a Windows Store App?

I'm trying to sign a Windows 8 appx package with a pfx file I have. I'm using a command like so:
signtool.exe sign /fd sha256 /f "key.pfx" "app.appx"
And from this, I get:
SignTool Error: No certificates were found that met all the given criteria.
What "criteria" am I not meeting? This is only for testing so these are self-signed certificates. I've tried importing the key and then signing it, but it always results in the same error. How do I fix this?
When getting this error through Visual Studio it was because there was a signing certificate setup to match the computer it was originally developed on.
You can check this by going to the project properties > signing tab and checking the certificate details.
You can uncheck "Sign the ClickOnce manifests" to disable signing.
If you don't want to turn this option off you will have to install the certificate.
Try with /debug.1,2 As in :
signtool sign /debug /f mypfxfile.pfx /p <password> (mydllexectuable).exe
It will help you find out what is going on. You should get output like this:
The following certificates were considered:
Issued to: <issuer>
Issued by: <certificate authority> Class 2 Primary Intermediate Server CA
Expires: Sun Mar 01 14:18:23 2015
SHA1 hash: DD0000000000000000000000000000000000D93E
Issued to: <certificate authority> Certification Authority
Issued by: <certificate authority> Certification Authority
Expires: Wed Sep 17 12:46:36 2036
SHA1 hash: 3E0000000000000000000000000000000000000F
After EKU filter, 2 certs were left.
After expiry filter, 2 certs were left.
After Private Key filter, 0 certs were left.
SignTool Error: No certificates were found that met all the given criteria.
You can see what filter is causing your certificate to not work, or if no certificates were considered.
I changed the hashes and other info, but you should get the idea.
1 Please note: signtool is particular about where the /debug option is placed. It needs to go after the sign statement.
2 Also note: the /debug option only works with some versions of signtool. The WDK version has the option, whereas the Windows SDK version does not.
I got the same problem in my console application development and as a quick workaround,
go to project properties then,
click on signing tab and uncheck "Sign the ClickOnce Manifest".
Image Description:
FYI You can also see this less one minute video solution. The above picture is taken form the video.
Please always check your certificate expiry date first because most of the certificates have an expiry date. In my case certificate has expired and I was trying to build project.
If you do not have to sign the app, right click on your project
Project Properties -> Signing -> uncheck "Sign the ClickOnce Manifest"
Also as this MS article suggests,
If you are using Visual Studio 2008 and are targeting .NET 3.5 and using automatic updates, you can just change the certificate and deploy a new version,
In my case I have the wrong type of certificate that I am trying to associate. I had "Server Authentication" rather than "Code signing".
You should be able to see this in Certificate snap in the Intended Purpose section. After that, it just work fine.
Got the same issue, turned out that the private key to the certificate had no permission.
To fix - open the certifacte management, find your certificate, right click -> Manage Private Keys and then in security on top be sure that your user is added and given permissions, that fixed it for me.
In case anyone else runs into this: My problem ended up being that I needed to run the command prompt as administrator before using the signtool.exe app. Then everything works wonderfully.
just uncheck the 'Sign the click once manifests' from the signing tab in project properties,it will remove the error and you can create a new one as from there.
I had this problem and I'm not entirely sure which step below made it work, but hope this helps somebody else...this is what I did:
Install the downloaded certificate (.crt) into certificates (I put it into “personal” store) - right click on .crt file and click Install Certificate.
Run certmgr.msc and export the certificate (found in whichever store you used in the 1st step) as a pfx file including private key and extended properties
Use the exported .pfx file when signing your project
Example signtool: signtool sign /f "c:\mycert.pfx" /p mypassword /d "description" /t http://timestamp.verisign.com/scripts/timstamp.dll $(TargetPath)
where the password is the same as provided during Export
I solved this by using the /sm flag to specify to look in the machine store instead of the default, which is My (Local User) store. Also, it can help to turn on debug for signtool by using /debug.
I'm having the same problem, reading some answers (posted here), I saw my certificate expired.
Just create a new one from my start project. Then at certificates manager deleted the expired certificate.
Now everything compiles fine.
The criteria include account name (whose private key it is associated with), domain, company, expiration date, intended purposes, among other things.
There are many different possible reasons for this error to occur, some have been listed already. Here is another tip: When importing a certificate, be sure you work with the original file received from the certificate authority (CA), or else some of the properties might be lost.
Example: recently I tried to import a certificate exported from a different account on the same machine. The certificate became visible to my account but was not associated with my account, and as a result signtool refused to recognize it without explicitly providing the file name and a password. Which, when done as part of the build process and written out explicitly in a batch file or source file, may not be sufficiently secure. (Importing the original CA-issued certificate solved it.)
I had the same "After Private Key filter, 0 certs were left" message and spent too much of my life trying to figure out what the message meant.
The problem was that I had installed the certificate incorrectly in the Windows Certificate store so there was no private key associated with the code signing certificate.
What I should have done was this:
Using either Firefox or Internet Explorer, submit the
request to the issuer. This generates a PRIVATE KEY which is stored silently by the browser (a dialog appears for a fraction of a second in Firefox). Note that other browsers may not work: your life is too short to find out if they do.
Submit the request, jump through the issuer's validation hoops and loops, sacrifice a goat, pray to the gods, submit a signed statement from your great grandparents, etc.
Download the certificate (.crt) and import it into the same browser. The browser now has both the private key and the certificate.
Export the certificate from the browser as a Personal Information Exchange (.p12) file. You will be asked to supply a password to protect this file.
Keep a backup copy of the .p12 file.
Run the Certificate Manager (certmgr.msc), right click on the Personal certificate store, select All Tasks/Import... and import the .p12 file into Windows. You will be asked for the password you used to protect the file. At this point, depending upon your security requirements, you can mark the key as exportable so you can restore a copy from the Windows store. You can also mark that a password is required before use if you want to break batch scripts.
Run signtool successfully, breathe a sigh of relief, and ponder how much of your life you have wasted due to bad error messages and poor or missing documentation.
My problem ended up being that I did not understand the signtool options. I had provided the /n option with something that did not match my certificate. When I removed that it stopped complaining.
I have had this issue too, tried a lot. Used SDK as well as Visual Studio signing, but everywhere I got "No certificates were found that met all the given criteria".
Solution:
Be aware that, if "after private key filter": '0 left' shows up with option signtool sign /debug..., the cause is your PC doesn't has the CA itself in the store. To solve this, install the CA first (in my case a .crt file), then run the sign again. It should work right now!
Signtool only can be used with a CA which is requested ánd owned by the same PC.
I had a similar problem my computer name had change and the certificate had expired. I was able to resolve this issue by creating a new test certificate.
In Visual Studio, right click on project in solution explorer. Select properties. Select Signing in properties window. Click "Create Test Certificate....". Enter password information for test certificate and click ok.
With /debug, when you get this message "After Private Key filter, 0 certs were left.", one reason could be that the pfx file doesn't have the private key.
When you export the installed certificate to pfx file ensure to enable the check box to also include the private key.
Go to project properties and uncheck all fields from the Firm before init the compilation
The digicert Token I use, must be recognized as "Microsoft Usbccid-Smartcard-Leser(WUDF)".
In case not, I get this error message 'No certificates were found that met all given criteria ...'.
That kept me searching in SignTool options and the properties of the certificates quite long with no effort at all. So I hope it helps someone :-)
I got this error when using Git Bash.
Using PowerShell succeeded.
If it helps anyone.

Quickbooks webconnector Integration 1048 error

I am using QuickBooks 2013 pro on server A, I have a self signed cert on server B. I had server A navigate to the services on SERVER B install the self signed cert into trusted root. Then had the web connector on SERVER A try to add an application. It is coming back with this error
qbwc1048: QuickBooks Web Connector Could not verify the web
application server certificate.
Any help in solving this would be great. Thank you very much
For anyone else having a similar problem, it may not be the certificate at all. It could be the server is closing the connection or returning an error code.
For more details, first click file -> clear log in the web connector. Then, try adding your application again. After it fails, click the view log button at the top right. Scroll down for more details on what happened.
You will see something like, remote server returned an error (401) unauthorized, or a similar message that is far more useful than what appears in the dialogue box.
#consolibyte Closed all IE windows and re-imported the SSL works now. Thank you

netsh error on windows 2008 R2

We are upgrading the server from Windows 2003 to 2008. As part of the process, I need to configure a port with a SSL certificate. When I ran the following command:
netsh http add sslcert ipport=1.2.3.4:8000 certhash=certificatehash appid={someGUID}
I got the following error:
SSL Certificate add failed, Error:
1312 A specified logon session does
not exist. It may already have been
terminated.
When running the command prompt with an administrator does not resolve the issue. Notice that I did not run into this issue on Windows 2003 (using httpcfg) and that things work well there.
Has anyone encountered this issue? Thanks.
the 1312 shows up at several occasions. aside from typos on the commandline, the other most common are:
- the certificate is not in your certificate store at all (check with MMC and 'certificates' snap-in)
- the certificate is in the wrong store: it should be in the store of 'local computer' not 'local user' (remember to choose proper account when activating the snapin inside the MMC)
- the certificate doesnt contain a Private Key (open the certificate and check whether it contains only public, or both keys)
I've spent over 1.5 hours before I found out that I've had generated a certificate succesfully, but misplaced some switches and it got written to a file without the private key:)
Are you running from an elevated command prompt (not just as an admin)?. There's an open source GUI tool that drives the HTTP config APIs directly- I use it on 2008 R2 with no trouble (it auto-requests elevation via UAC). I've had mixed results with netsh/httpcfg. This one always works for me (and it behaves the same everywhere).