I have Virtual Box with Gitlab instance and I'm trying to register on the same machine gitlab-runner, during that I'm getting issue about IP Sans
VM: https://bitnami.com/stack/gitlab/virtual-machine
Process
I think verifying certificate is successful (please correct)
Also what I have done also
added "subjectAltName=IP:192.168.8.6" to /etc/ssl/openssl.cnf
Generated cert and key in /etc/gitlab-runner
Copied these 2 to: /etc/gitlab/trusted-certs/
Doing also solution from below also doesn't help
Gitlab-CI runner: ignore self-signed certificate
Any ideas how I can further debug? Any help appreciated
From this post
Step1 edit ssl configuration on the GitLab server (not the runner)
+sudo nano /etc/pki/tls/openssl.cnf
# find line
[ v3_ca ]
subjectAltName=IP:192.168.1.1 <---- Add this line. 192.168.1.1 is your GitLab server IP.
Step2 Re-generate self-signed certificate on the GitLab server (not the runner)
cd /etc/gitlab/ssl
sudo openssl req -x509 -nodes -days 3650 -newkey rsa:2048 -keyout /etc/gitlab/ssl/192.168.1.1.key -out /etc/gitlab/ssl/192.168.1.1.crt
sudo openssl dhparam -out /etc/gitlab/ssl/dhparam.pem 2048
sudo gitlab-ctl restart
Step3 Copy the new CA to the GitLab CI runner in /etc/gitlab-runner/certs/
Step4 Register your runner
gitlab-runner register --tls-ca-file="$CERTIFICATE"
this work for me.
For those errors like:
x509: certificate is not valid for any names, but wanted to match gitlab.example.com
x509: certificate relies on legacy Common Name field, use SANs instead
...
I am running gitlab server 15.7.1 docker container from my laptop (following The Official Install Guide - With docker-compose), and installed a gitlab runner at that laptop host too.
In my case, the self-signed certificate should be re-requested manually according with the following steps:
Entering the running gitlab container:
docker compose exec web bash
In container, copy the openssl.cnf to /etc/gitlab/ssl so that I can edit it from host machine:
cp /opt/gitlab/embedded/ssl/openssl.cnf /etc/gitlab/ssl/
At host, Modify openssl.cnf to add new line into v3_ca section:
subjectAltName=DNS:gitlab.example.com
NOTE that a DNS name needed instead of IP
In container, copy back the file:
cp /etc/gitlab/ssl/openssl.cnf /opt/gitlab/embedded/ssl/
In container, recreate x509 req and restart gitlab services to sign the gitlab server certificate again:
cd /etc/gitlab/ssl
openssl req -x509 -nodes -days 3650 -newkey rsa:2048 -keyout /etc/gitlab/ssl/gitlab.local.key -out /etc/gitlab/ssl/gitlab.local.crt
openssl dhparam -out /etc/gitlab/ssl/dhparam.pem 2048
gitlab-ctl restart
Now, gitlab-runner register should be ok.
gitlab-runner register --tls-ca-file="$CERTIFICATE"
Lucky to anyone.
Related
I have created a socketcluster nodejs app. I followed their official docs to deploy the service to Google K8s Engine. However the ingress service is not running up and complains about :
Error:googleapi: Error 400: The SSL key is too large., sslCertificateKeyTooLarge
I tried following certificates:
4048 Key size certificate from Let'sEncrypt
2048 Key size using cert created using Open SSL.
Both of them result the the same error.
Do any one know how do I resolve this? And where do I get proper certificate for enabling TLS?
IIRC, only RSA-2048 and ECDSA P256 keys are supported:
openssl genrsa -out PRIVATE_KEY_FILE 2048
openssl ecparam -name prime256v1 -genkey -noout -out PRIVATE_KEY_FILE
I also struggled due to this error on using Letsencrypt certs with 4096bit private key to a GKE ingress - even creating the secret worked fine for [1].
Finally overcame with editing "/etc/letsencrypt/cli.ini"
rsa-key-size = 2048
issued new certificate, keyfile and put those into secret.
[1] https://cloud.google.com/kubernetes-engine/docs/how-to/ingress-multi-ssl
On Cloud Shell, GCP with "openssl" and "gcloud", I tried to create a self-managed SSL certificate first running this command below to create "myCert.crt" and "myKey.key":
openssl req -new -newkey rsa:4096 -x509 -days 365 -nodes -out myCert.crt -keyout myKey.key
Then, ran this command below to create the self-managed SSL certificate "mycert" using "myCert.crt" and "myKey.key":
gcloud compute ssl-certificates create mycert --certificate=myCert.crt --private-key=myKey.key
But I got a similar error to yours:
ERROR: (gcloud.compute.ssl-certificates.create) Could not fetch
resource:
The SSL key is too large.
So I changed "rsa:4096" to "rsa:2048" then ran the first command again:
// "4096" is changed to "2048"
openssl req -new -newkey rsa:2048 -x509 -days 365 -nodes -out myCert.crt -keyout myKey.key
Then, ran the second command again:
gcloud compute ssl-certificates create mycert --certificate=myCert.crt --private-key=myKey.key
Finally, I could create the self-managed SSL certificate "mycert":
Created
[https://www.googleapis.com/compute/v1/projects/myproject-923743/global/sslCertificates/mycert].
NAME: mycert TYPE: SELF_MANAGED CREATION_TIMESTAMP:
2022-01-22T07:22:26.058-08:00 EXPIRE_TIME:
2023-01-22T07:22:08.000-08:00 MANAGED_STATUS:
I have an Outlook Web Add-In, running on node server, locally I need it to run with a valid ssl cert to debug in IE.
In Chrome it looks like this
I have built a cert using the following command line
C:\OpenSSL-Win32\bin\openssl req -newkey rsa:2048 -new -nodes -x509 -days 3650 -keyout key.pem -out cert.pem
I run the application using VS Code and the command line npm start
My env file is .env.development.local
it contains one line HTTPS = 'true'
I have looked at
https://www.npmjs.com/package/serve-https
It appears to look like that is not used for development
So inside scripts\start.js file what do I need to get it to respect the ssl cert.
Trying to achieve a sort of self-signed PKI setup utilizing openssl on RHEL, with a few caveats. I will attempt to provide as much information as possible here.
Versions: RHEL 6.7 | OpenSSL 1.0.1e-fips 11 Feb 2013
Caveats/constraints on the script: this script will be utilized to create multiple key sets - by default, one root keypair and cert and two client keypairs and certs per run. User input is asked for for file location, filename and passphrase on the client keys. All this was fairly straightforward, and I had a script that would run these commands on user request, and utilized the openssl.cnf file to point back to the root key to sign. I had used sed to change the location pointers in openssl.cnf based on filename originally, and was successfully able to sign the client cert.
However, there are two major caveats to this:
I was asked to change the script so that it is not dynamically the script or other files per each run, meaning openssl.cnf should not be edited on the fly if possinble. If this is needed to function however, then it should be fine.
The user needs to be able to run multiple sets of this script ad hoc, especially with regards to the client keypairs (I have the root script and the client generation script separate in a user selectable menu). That is to say, generating the client keypair is requisite on having a root key to associate with, but can be done multiple times, and the client key script should ask the user with which root key to associate and sign from?
Because of these constraints, it didn't seem that editing openssl.cnf was a prudent option, and not very scalable. So, given this info my question which I've been unable to figure out is simply:
Is there a way to point a client key to a variable which would be the root key cert to sign? (Rather than utilizing openssl.cnf for the 'certificate' and 'private_key' entries?)
As of now, I have:
root key & cert:
openssl req -config $dir/openssl.cnf -new -x509 -days 3652 -nodes -sha384 -newkey ec:ec-secp384r1.pem -keyout $userdir/${rootName}_private.key -out $userdir/${rootName}.crt -subj "stuff_here"
...
export rootName
client keys & certs:
read -p "Which root key do you want to associate this client keypair with? Please type absolute filepath and filename (ending in .key); rkAssoc #STILL NEED TO USE THIS VARIABLE
##KEY GENERATION
openssl req -newkey ec:ec-secp384rp1.pem -keyout $userdir/{$clientName}_privat.key -out $userdir/client/${clientName}.csr -subj "Stuff_here"
##SIGN CSR
openssl ca -config $dir/openssl.cnf -policy policy_anything -extensions usr_cert -days 730 -notext -md sha384 -in $userdir/client/${clientName}.csr -out $userdir/client/${clientName}_signedprivatekey.pem && echo "Client key created."
So I guess,
1) Did i do the client signing correctly (something seems off about it but not sure)
2) instead of referencing -req openssl.cnf I presume there is probably some kind of flag where you could do something more like
openssl ca ... -cert ${rkAssoc}
is this remotely correct or am I way off?
Thanks in advance for anyone who lends a hand.
OpenSSL has multiple ways of doing the same thing. You found one way of signing a CSR, with openssl ca. openssl x509 can act as a mini CA, so to eliminate the need for a config file, you could do something like:
openssl x509 -req -in /tmp/mykey.csr.pem -CA /path/to/ca/mycacert.pem -CAkey /path/to/ca/cakey engine -CAserial /path/to/ca/myca.srl -days 3600 -out /tmp/mykeypub.cert.pem
Where -CA points to your root CA cert, and -CAkey to your root CA key.
I'm trying to automate my infrastructure using puppet with Azure.
I've created my puppet master using puppet enterprise VM from gallery on Azure. After that, I've generated a certificate running the following commands:
sudo openssl req -x509 -nodes -days 365 -newkey rsa:1024 -keyout cert.pem -out cert.pem
sudo openssl x509 -inform PEM -in cert.pem -outform DER -out cert.cer
Then, I've uploaded my cert file to Azure Management Portal and finally, I've create a VM with the following command:
puppet azure_vm create --management-certificate /tmp/cert.pem --azure-subscription-id=xxxxxx --image 0b11de9248dd4d87b18621318e037d37__RightImage-Ubuntu-12.04-x64-v13.4 --location 'west us' --vm-name puppetslave --vm-user xxxx --password xxxxx --puppet-master-ip 23.102.xxx.xxx
Finally, I've exported my private key using PuttyGen and I'm using it to connect to my new linux vm. Unfortunately, I'm getting the following error:
server refused out key
And I don't know why. I'm stuck at this point and don't know how to proceed.
I am trying to use push notifications through my nodejs server deployed in openshift rehat server. I could not find any documentation/discussion to generate apns certificate in openshift.
I tried generating a certificate, but getting an error while generating CSR using this command
openssl genrsa -out sz.key 2048
Error: unable to write 'random state'
e is 65537 (0x10001)
After some googling, I see one option is to use aerogear. But I want to use my own server to handle the notifications.
Is there any other option to generate a certificate in openshift server?
Thanks developercorey for your help. But i still see the error.
Finally generated certificate for my openshift server using this script
openssl x509 -in aps_production.cer -inform DER -out myProdCert.pem
The private key associated with the SSL certificate can be exported from the Keychain Access application on your Mac computer. - mykey.p12
openssl pkcs12 -in mykey.p12 -out myProdKey.pem -nodes -clcerts
openssl s_client -connect gateway.push.apple.com:2195 -cert myProdCert.pem -key myProdKey.pem
Hopefully this will help somebody.
Try this command:
HOME=~/app-root/data/ openssl genrsa -out sz.key 2048
Basically it is trying to write some temp data into your $HOME, but you can't write to that directory, so it won't work, but it does give somewhat of a cryptic error message.