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
Currently I'm not able to use our internal TSA with cosign attest (I'm testing the version from #3889) - the error is below. The TLS connection to the TSA server requires a custom CA roots, but cosign attest - unlike cosign sign - doesn't have the --timestamp certificate fields, other than --timestamp-server-url. cosign sign does have the following fields:
$ cosign sign --help
[...]
--timestamp-client-cacert='':
path to the X.509 CA certificate file in PEM format to be used for the connection to the TSA Server
--timestamp-client-cert='':
path to the X.509 certificate file in PEM format to be used for the connection to the TSA Server
--timestamp-client-key='':
path to the X.509 private key file in PEM format to be used, together with the 'timestamp-client-cert' value,
for the connection to the TSA Server
--timestamp-server-name='':
SAN name to use as the 'ServerName' tls.Config field to verify the mTLS connection to the TSA Server
--timestamp-server-url='':
url to the Timestamp RFC3161 server, default none. Must be the path to the API to request timestamp responses,
e.g. https://freetsa.org/tsr
For parity with cosign sign and to enable using a non-public TSA which is crucial for a private infrastructure / BYOPKI use case, I suggest to add the command-line flags above. I will submit a PR equivalent to the patch that we internally have to apply for several month now - it would be great not to have to patch cosign code! 😄
Initially we hesitated whether we should try to submit that change since there were discussions regarding use of the trusted root. But now having looked into this more closely, I believe the trusted root is more relevant to the verification (cosign verify, cosign verify-attestation) than to producing signatures and attestation. Even though I'm aware and understand the reservations regarding the "command-line flag proliferation", I think in this case it would be justified. In particular, it would help up to test and get practical experience with the attestations in the "new bundle format" once #3889 is merged. Please let me know if you have any questions - will attach a PR with the suggested change shortly.
Expand the list of commands that support the mTLS and custom CA
TSA parameters to include `cosign attest` and `cosign blob-attest`.
Related to sigstore/cosign#4079
and its issue
sigstore/cosign#4078.
Signed-off-by: Dmitry Savintsev <dsavints@gmail.com>
Expand the list of commands that support the mTLS and custom CA
TSA parameters to include `cosign attest` and `cosign blob-attest`.
Related to sigstore/cosign#4079
and its issue
sigstore/cosign#4078.
Signed-off-by: Dmitry Savintsev <dsavints@gmail.com>
dmitris
changed the title
cosign attest: add TSA certificate command-line options and AttestOptions fields
cosign attest[-blob]: add TSA certificate command-line options and AttestOptions fields
Feb 24, 2025
Description
Currently I'm not able to use our internal TSA with
cosign attest
(I'm testing the version from #3889) - the error is below. The TLS connection to the TSA server requires a custom CA roots, butcosign attest
- unlikecosign sign
- doesn't have the--timestamp
certificate fields, other than--timestamp-server-url
.cosign sign
does have the following fields:For parity with
cosign sign
and to enable using a non-public TSA which is crucial for a private infrastructure / BYOPKI use case, I suggest to add the command-line flags above. I will submit a PR equivalent to the patch that we internally have to apply for several month now - it would be great not to have to patch cosign code! 😄Initially we hesitated whether we should try to submit that change since there were discussions regarding use of the trusted root. But now having looked into this more closely, I believe the trusted root is more relevant to the verification (
cosign verify
,cosign verify-attestation
) than to producing signatures and attestation. Even though I'm aware and understand the reservations regarding the "command-line flag proliferation", I think in this case it would be justified. In particular, it would help up to test and get practical experience with the attestations in the "new bundle format" once #3889 is merged. Please let me know if you have any questions - will attach a PR with the suggested change shortly.Previous related issues / PRs:
The current
cosign attest
error:The text was updated successfully, but these errors were encountered: