You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Kerberos connections immediately terminate when connecting to OMI server, despite OMI authorizing the given Kerberos ticket, returning a 200 OK and an encrypted body.
Upon decryption, the message is corrupted on the client side... which would seemingly indicate a problem like:
A difference in expectations around payload structure
Incorrect key negotiation?
The OMI server Kerberos configuration appears to be configured correctly as two other clients can use Kerberos to talk to the server:
omicli included with OMI itself
pwsh (6.2.1)
I have a full reproduction environment that's containerized at puppetlabs/bolt#1101
Follow the description in the PR and you'll end up with 3 containers:
Samba serving as Active Directory
OMI Server listening for winrm connections (domain joined)
An Ubuntu dev environment that has Bolt / winrm gem (domain joined)
I've spent a little bit of time looking at the messages, poking around in pry and augmenting the omi source in an effort to hone in on where the negotiation problem is, but haven't had the a-ha moment just yet. Before I get too far into the weeds, figured I'd post this issue to see if there are any known issues around the Kerberos handling.
Thanks!
The text was updated successfully, but these errors were encountered:
Edit: unfortunately this might not be related, I tried using PowerShell Core on Linux to connect to the endpoint and it is also failing to continue on past the first encrypted message received from the host. Seems like something special that Windows is doing that I am not privy to.
Kerberos connections immediately terminate when connecting to OMI server, despite OMI authorizing the given Kerberos ticket, returning a 200 OK and an encrypted body.
Upon decryption, the message is corrupted on the client side... which would seemingly indicate a problem like:
The OMI server Kerberos configuration appears to be configured correctly as two other clients can use Kerberos to talk to the server:
omicli
included with OMI itselfpwsh
(6.2.1)I have a full reproduction environment that's containerized at puppetlabs/bolt#1101
Follow the description in the PR and you'll end up with 3 containers:
I've spent a little bit of time looking at the messages, poking around in
pry
and augmenting theomi
source in an effort to hone in on where the negotiation problem is, but haven't had the a-ha moment just yet. Before I get too far into the weeds, figured I'd post this issue to see if there are any known issues around the Kerberos handling.Thanks!
The text was updated successfully, but these errors were encountered: