LDAPS is correctly set up using docker-openldap.
Certificates correctly generated and deployed (also tried self generated certificates)
ldapwhoami and ldapsearch both works
Presto complains authenticator not loaded, with DEBUG level logging it says "simple bind failed: outbound connection failed"
It turns out by default osixia/openldap container will run with TLS client certificate verification on (demand). This requires a local certificate and a private key, i.e. the ldap server will try to validate client owns a valid certificate.
ldapwhoami works because the client certificate was imported to local machine, but on Presto node only a root certificate was imported during initialization.
The error message was not clear, showing UncheckedException, authenticator not loaded, simple bind failed, etc.
The solution is to disable client certificate verification. When launching the docker, add --env LDAP_TLS_VERIFY_CLIENT=try or never.
Importing the private key to default Java keystore might also work if client verification is desired.
Related
I had Configure the Domino Credential Store.
I had modified the Domino Proton Server settings that enable client authentication.
I created the Vault ID.
I created the IAM-store.nsf from template with error message.
Error executing agent 'DeleteExpiredDocs' in 'iam-store.nsf'. Agent signer 'Domino Template Development/Domino': You are not authorized to perform that operation
I gave the IAM's functional ID access to the database.
I installed the IAM services for domino with the following message.
result screen of install domino-iam-service-2.2.0.tgz
Since I would like to config the iam-services for my testing server.
I select to setup the pilot mode.
According to the tutorial, https://doc.cwpcollaboration.com/appdevpack/docs/en/iam_landing_page.html
I could access the demo database, with anonymous setting of proton server.
C:\src\domino-db\package>npm run ptest -- read serv.org.com:3003/App\node-demo.nsf -q "Form = 'Contact' and LastName = 'Moody'"
read the content of demo database
Config the pilot mode successfully.
What is doing wrong?
Error, when try to startup pilot mode of IAM Service
I have put all the certificates to the folder config/certs,
in which the certificates are created by create_certs.cmd from the tutorial.
And I have convert the ca.crt into ca.pem.
Besides, I also put the keys created by ProtonCA into the config/certs.
Keys created by ProtonMicroCA
According to the tutorial, I modified the make_certs.cmd as the following:
make_certs.cmd
the certificates are posted to the config/certs directory
I'm not sure about your complete setup, a support ticket would help us diagnose this better. There should be a ca folder in the config/certs directory that contains any root certs you're using (like the ca.pem you have)
A little background:
I have a Tesla Powerwall which has it's own built in web server that can be accessed on the local network. It only allows SSL connections and uses a self signed certificate. I have setup port forwarding that allows me to connect to the web server remotely. For a while, i've had working node.js apps both on a local Pi and also a remote AWS instance that made requests to the Powerwall web server to retrieve bits of information.
Since yesterday, Tesla updated my Powerwall and now everything has stopped working. I can only assume they have changed something regarding how the web server handles it's self signed SSL certificate.
Firstly, my Pi running on the local network would not make successful node.js requests to the local server. I managed to get this working by adding an entry to my /etc/hosts file like this:
192.168.1.42 powerwall
and now my node.js app can successfully connect again using https://powerwall
When using Safari or Chrome to connect remotely, I can connect if I use my IP address (After trusting the self signed cert) but cannot connect when using my DDNS address that points to home. (I have confirmed the DDNS is working). It gives me the error:
Safari can’t open the page “https://home.xxxxxx.com:4444” because Safari can’t establish a secure connection to the server “ home.xxxxxx.com”.
My AWS node.js app will not connect regardless of me using the IP address or DDNS address giving me the error:
Client network socket disconnected before secure TLS connection was established
This is how I am trying to connect:
request({
url: 'https://xx.xx.xx.xx:xxxx/api/system_status/soe',
method: 'GET',
rejectUnauthorized: false,
requestCert: true,
agent: false,
headers: headers
}
I have tried adding:
secureProtocol: 'TLSv1_method'
and attempted with the methods TLSv1_method TLSv1_1_method TLSv1_2_method in case it needed a specific method, with no luck.
Does the above sound like the SSL settings on the server have been screwed down?
What can I do to:
a) access the site remotely through a browser using the DDNS address
b) force node.js to not be interested in the SSL certificate at all and just connect
----- EDIT
Certificate:
Data:
Version: 3 (0x2)
Serial Number:
46:.....
Signature Algorithm: ecdsa-with-SHA256
Issuer: C=US, ST=California, L=Palo Alto, O=Tesla, OU=Tesla Energy Products, CN=335cbec3e3d8baee7742f095bd4f8f17
Validity
Not Before: Mar 29 22:17:28 2019 GMT
Not After : Mar 22 22:17:28 2044 GMT
Subject: C=US, ST=California, L=Palo Alto, O=Tesla, OU=Tesla Energy Products, CN=335cbec3e3d8baee7742f095bd4f8f17
Subject Public Key Info:
Public Key Algorithm: id-ecPublicKey
Public-Key: (256 bit)
pub:
04:ca...
ASN1 OID: prime256v1
NIST CURVE: P-256
X509v3 extensions:
X509v3 Key Usage: critical
Digital Signature, Key Encipherment
X509v3 Extended Key Usage:
TLS Web Server Authentication
X509v3 Basic Constraints: critical
CA:FALSE
X509v3 Subject Alternative Name:
DNS:teg, DNS:powerwall, DNS:powerpack, IP Address:192.168.90.1, IP Address:192.168.90.2, IP Address:192.168.91.1
With HTTPS, the domain needs to match what’s signed in the cert; it’s usually the public domain.
It’s not supposed to be the IP, and it certainly won't be the DDNS hostname (if I understood correctly) you’re pointing at it.
There are 3 possible approaches;
Add the certificate from the powerwall as a ‘known’ rootCA (as already suggested),
Tell node.js to skip checking the validity of the certificate, or
Try with HTTP 😬
Proper operation of the HTTPS connection process will also depend on you accessing the powerwall using the domain name registered in the certificate (which may require your DNS server to respond with the appropriate IP when the lookup is made ~> like DNS spoofing proof-of-concept for a CTF).
Also, to your musings in comments, while some browsers may allow you to override an expired or self-signed cert (or when connecting via IP), but it’s very sketchy to connect with a domain and get a cert that specifies and entirely different domain (which is why the browser might not even present you the option).
HTH
Post-resolution update:
How to get the DNS name to match what's on the certificate:
add an entry in the client system's /etc/hosts or equivalent
connect using the hostname (not the IP)
When connecting over public Internet:
How to get public-internet connections through to the local host:
get a public-facing HTTPS cert (e.g.) that matches your DDNS domain or /etc/hosts entry
Host a HTTP-proxy
relay requests from Internet (hopefully with filtering/validation) to the powerwall
(you will have 2 HTTPS connections: one from AWS -> proxy, one from proxy->powerwall)
Host a custom API that will return exactly the [minimum] info needed by the AWS service
How to trust a self-signed certificate? (this wasn't the blocking factor)
Try this for debugging:
openssl s_client \
-connect 192.168.1.42:4444 \
-CAfile /path/to/self-signed-cert \
-verify_hostname powerwall \
-debug
Can find more options in openssl s_client -help
Do you have any servers running on your home network (apache, nginx, etc)? You're probably trying to connect to https://my.ddns.com and you're passing it directly to powerwall, which has a certificate for powerwall.
Connecting to a host that returns a certificate which does not contain that hostname will cause a TLS error. You probably want to run a forward proxy, where your server hosts my.ddns.com, sets up the TLS connection and then forwards the traffic (without TLS) to 192.168.1.44.
I'm trying to send a GET request using the Insomnia app to a NodeJS server app -- I didn't write the app but have joined the team.
Although I get a reasonable JSON response when I hit the URL -- https://127.0.0.1:9999 -- from the browser, I get the error "Error: Peer certificate cannot be authenticated with given CA certificates" when I'm using Insomnia. Using a Mac, MacOS 10.12.4. Node v6.3.1.
The Insomnia timeline says:
* Preparing request to https://127.0.0.1:9999/
* Enable automatic URL encoding
* Enable SSL validation
* Enable cookie sending with jar of 2 cookies
* Hostname in DNS cache was stale, zapped
* Trying 127.0.0.1...
* TCP_NODELAY set
* Connected to 127.0.0.1 (127.0.0.1) port 9999 (#8)
* WARNING: using IP address, SNI is being disabled by the OS.
* SSL certificate problem: Invalid certificate chain
* Curl_http_done: called premature == 1
* Closing connection 8
Thanks for any help!
There is little documentation on how Insomnia handles certificates. As long as they are normal certificates that are signed by a typical CA authority there is typically no problem. Since you also refer to your loopback address (127.0.0.1) I assume that you're also testing with a self-signed certificate.
I noticed that Insomnia uses the Mozilla list of certificate authorities. It does not use the list of your operating system.
The list is stored in a text-file in a directory like C:\Temp\insomnia_5.12.4. In my case it was for example 2017-01-18.pem. You can add your own signing authority certificate to this file.
I didn't look into how stable this file is or how it is created.
You can also workaround the certificate errors by disabling them in your settings (settings > Validate SSL Certificates).
I've seen both this and this — same problem, different question.
I'm trying to connect my Windows 8.1 Store app to an ASP.NET Web API web service, secured over HTTPS using a self-signed certificate. It's a proof-of-concept application that will end up on < 5 different machines and seen only internally, so I was planning to just install the certificate as trusted on each of the target machines.
When I try this on my development setup, both HttpClient APIs fail to establish the trust relationship when calling the service.
Windows.Web.Http.HttpClient exception: "The certificate authority is invalid or incorrect"
System.Net.Http.HttpClient exception: "The remote certificate is invalid according to the validation procedure."
My self-signed certificate (public-key-only .cer version) is installed in both the "User" and "Local Machine" Trusted Root Certification Authorities on the client. I'm really surprised that this isn't enough to get WinRT to trust it. Is there something I'm missing, or is there just no way to set up the trust relationship for a self-signed SSL certificate that will make HttpClient happy?
Details on my setup:
ASP.NET Web API
Azure web role running in Azure emulator
Cert issuer: 127.0.0.1
Cert subject: 127.0.0.1
Cert key: 2048-bit
Windows 8.1 Store application
Certificate (.cer file with public key only) installed in User\Trusted Root Certification Authorities
Certificate (.cer file with public key only) installed in Local Machine\Trusted Root Certification Authorities
Certificate (.cer file with public key only) added to Windows Store app manifest under "CA"
I am not asking for a workaround to configure HttpClient to accept self-signed or invalid certificates in general — I just want to configure a trust relationship with THIS one. Is this possible?
You should be able to find out what is the problem with the certificate by doing a request like this:
// using Windows.Web.Http;
private async void Foo()
{
HttpRequestMessage request = null;
try
{
request = new HttpRequestMessage(
HttpMethod.Get,
new Uri("https://localhost"));
HttpClient client = new HttpClient();
HttpResponseMessage response = await client.SendRequestAsync(request);
}
catch (Exception ex)
{
// Something like: 'Untrusted, InvalidName, RevocationFailure'
Debug.WriteLine(String.Join(
", ",
request.TransportInformation.ServerCertificateErrors));
}
}
Using a HttpBaseProtocolFilter you can ignore certificate errors:
// using Windows.Web.Http;
// using Windows.Web.Http.Filters;
// using Windows.Security.Cryptography.Certificates;;
HttpBaseProtocolFilter filter = new HttpBaseProtocolFilter();
filter.IgnorableServerCertificateErrors.Add(ChainValidationResult.Untrusted);
filter.IgnorableServerCertificateErrors.Add(ChainValidationResult.InvalidName);
filter.IgnorableServerCertificateErrors.Add(ChainValidationResult.RevocationFailure);
HttpClient client = new HttpClient(filter);
HttpResponseMessage response = await client.SendRequestAsync(request);
The piece I was missing turned out to be that the certificate wasn't in the list of of IIS Server Certificates on my local machine!
Opening IIS Manager and checking out the Server Certificates section, I did find a 127.0.0.1 SSL certificate already set up by the Azure emulator:
CN = 127.0.0.1
O = TESTING ONLY
OU = Windows Azure DevFabric
However, my own self-signed certificate that I made outside of IIS, also with CN=127.0.0.1, was not in the list. I imported it, and now my Windows Store app's HttpClient connects happily (certificate warnings went away in Chrome and IE as well!)
If anyone can firm up the technical details on this, please comment — this fix feels a bit magical and I'm not sure I can pinpoint precisely why this worked. Possibly some confusion on my part between the two certs for 127.0.0.1, even though the thumbprint I had configured in my Azure project was always the one I was intending to use?
I am trying to setup a website (local testing atm), to connect to azure rest api to see our settings. I created a cert locally (W7 machine):
makecert -sky exchange -r -n "CN=azureConnectionNew" -pe -a sha1 -len 2048 -ss My "azureConnectionNew.cer"
I can see the cert in the certs MMC snap in. (do not have a right click edit permissions option when I view the cert in here).
I have a class library that setups up the connection, the cert is passed in by getting the cert (via the thumb string), this works great for the console app, but when I try and do this in a web app it all goes wrong. I get 403 errors.
I first thought that this was due to the fact that the website is running as the ApplicationPoolIdentity so doesn't have access to the cert. So I tried passing in the cert (to the same code as the console app), by loading the actual file:
var path = #"C:\temp\azureconnection\azureConnectionNew.cer";
var cert = new X509Certificate2();
cert.Import(path);
I still get 403 errors.
I tried exporting the cer file from MMC certificates snap in as a pfx file, (with private keys included). I set the local IIS set to use this cert and navigated to the https version of my local site but still got 403.
I am not sure how to include / setup / reference the cert so that IIS can send a HttpWebRequest from the server side to Azure and get a valid response.
It is always better to use Thumbprint of the certificate to get the certificate. Please make sure you have created the certificate correctly. Also please check you have placed the certificate in Personal certificate section in Local Machine. You can check this using MMC snap in. please try below code..
var store = new X509Store(StoreName.My, StoreLocation.LocalMachine);
store.Open(OpenFlags.OpenExistingOnly | OpenFlags.ReadOnly);
var certificate = store.Certificates
.Cast<X509Certificate2>()
.SingleOrDefault(c => string.Equals(c.Thumbprint, “CertificateThumbprint”, StringComparison.OrdinalIgnoreCase)); // please replace CertificateThumbprint with original Thumbprint
This isn't the right way to use the certificate - it needs to be stored in the personal/certificates store of the user running the code (you should update the App Pool identity to be a user who can login and into whose certificates you import the cert. Here's sample code showing you how to use the service API: http://code.msdn.microsoft.com/windowsazure/CSAzureManagementAPI-609fc31a/