[release/5.0] Set isFullMatrix to true for release/5.0

[josalem]

As discussed on #41217 and in the hopes of avoiding something like #41217 happening again, these changes set the Azure Pipeline variable isFullMatrix to true. This should cause all platform variants to run for PRs on the release/5.0 branch. Specifically, this change will cause all the jobs in runtime.yaml to evaluate their trigger condition to true.

Customer Impact

N/A

Testing

N/A

Risk

This is a purely infrastructure change that should mitigate further risk.

CC - @dotnet/runtime-infrastructure @tommcdon

[ghost]

Tagging subscribers to this area: @ViktorHofer
See info in area-owners.md if you want to be subscribed.

[danmoseley]

If someone knowledgeable signs off on this, it's approved to merge as it reduces 5.0 risk eg of Windows 7 issues.

[jeffschwMSFT]

Approved. Please get a cr. I will follow-up on when to merge.

[ericstj]

@safern can you please have a look at this?

[safern]

I think we should do a more surgical change here. This enables also more helix queues for libraries tests which could slow down builds and clog queues, specially since those are intentionally only set for full matrix either due to limited resources or because they have a lower priority.

Why not just add Windows 7 to the libraries tests runs against a Checked runtime instead?

[josalem]

The intention was to mirror what we have set up for release/3.1. This change only sets isFullMatrix for release/5.0, which I would hope has much, much lower PR traffic than master at this point. If we want to do something more surgical till a later RC release, though, that is understandable. I don't think we'd do something as heavy handed as this in master, though, for the reasons you mentioned @safern.

[safern]

I think 5.0 is still having enough traffic to not do it yet, but I don't know what other folks think.

[safern]

Btw I agree we should do this, however I don't know if we should do it yet.

[josalem]

Then it seems like we're in consensus on this patch being merged, but we need to agree on the timing. I'm content to wait till traffic dies down a little bit if validation and CI time is a concern.

Let me look into what the surgical version would look like. If we are specifically looking to enable Windows 7, we may need to modify platform-matrix.yml and then plumb the change back up to the pipelines.

[safern]

I'll approve as the change makes sense, but we should wait on timing consensus.

cc: @jkotas @danmosemsft

[ericstj]

I agree. At least wait until after we are in tactics ask mode.

[akoeplinger]

The test failures seem valid/real though.

[ericstj]

Presumably any failures here are also showing up in CI builds — cc @Anipik

[ericstj]

So every test is failing on Windows7 with HRESULT: 0x8007007F, ERROR_PROC_NOT_FOUND. This appears to be #41217. @josalem perhaps you just need to rerun to ensure this was built with #41305.

Looks like most Linux builds are seeing a failure of System.Net.Security.Tests.ServerAsyncAuthenticateTest.ServerAsyncAuthenticate_SniSetVersion_Success(version: Tls11). @wfurt looks similar to #39445 is something missing from release/3.0?

System.AggregateException : One or more errors occurred. (One or more errors occurred. (Authentication failed, see inner exception.)) (One or more errors occurred. (Authentication failed, see inner exception.))\n---- System.AggregateException : One or more errors occurred. (Authentication failed, see inner exception.)\n-------- System.Security.Authentication.AuthenticationException : Authentication failed, see inner exception.\n------------ Interop+OpenSsl+SslException : SSL Handshake failed with OpenSSL error - SSL_ERROR_SSL.\n---------------- Interop+Crypto+OpenSslCryptographicException : error:1409442E:SSL routines:ssl3_read_bytes:tlsv1 alert protocol version\n---- System.AggregateException : One or more errors occurred. (Authentication failed, see inner exception.)\n-------- System.Security.Authentication.AuthenticationException : Authentication failed, see inner exception.\n------------ Interop+OpenSsl+SslException : SSL Handshake failed with OpenSSL error - SSL_ERROR_SSL.\n---------------- Interop+Crypto+OpenSslCryptographicException : error:14209102:SSL routines:tls_early_post_process_client_hello:unsupported protocol


Stack trace
   at System.Threading.Tasks.TaskTimeoutExtensions.WhenAllOrAnyFailed(Task[] tasks) in /_/src/libraries/Common/tests/System/Threading/Tasks/TaskTimeoutExtensions.cs:line 122
   at System.Threading.Tasks.TaskTimeoutExtensions.WhenAllOrAnyFailed(Task[] tasks, Int32 millisecondsTimeout) in /_/src/libraries/Common/tests/System/Threading/Tasks/TaskTimeoutExtensions.cs:line 84
   at System.Net.Security.Tests.ServerAsyncAuthenticateTest.ServerAsyncAuthenticate_SniSetVersion_Success(SslProtocols version) in /_/src/libraries/System.Net.Security/tests/FunctionalTests/ServerAsyncAuthenticateTest.cs:line 123
--- End of stack trace from previous location ---
----- Inner Stack Trace #1 (System.AggregateException) -----

----- Inner Stack Trace -----
   at System.Net.Security.SslStream.ForceAuthenticationAsync[TIOAdapter](TIOAdapter adapter, Boolean receiveFirst, Byte[] reAuthenticationData, Boolean isApm) in /_/src/libraries/System.Net.Security/src/System/Net/Security/SslStream.Implementation.cs:line 393
   at System.Threading.Tasks.TaskTimeoutExtensions.WhenAllOrAnyFailed(Task[] tasks) in /_/src/libraries/Common/tests/System/Threading/Tasks/TaskTimeoutExtensions.cs:line 96
----- Inner Stack Trace -----
   at Interop.OpenSsl.DoSslHandshake(SafeSslHandle context, ReadOnlySpan`1 input, Byte[]& sendBuf, Int32& sendCount) in /_/src/libraries/Common/src/Interop/Unix/System.Security.Cryptography.Native/Interop.OpenSsl.cs:line 282
   at System.Net.Security.SslStreamPal.HandshakeInternal(SafeFreeCredentials credential, SafeDeleteSslContext& context, ReadOnlySpan`1 inputBuffer, Byte[]& outputBuffer, SslAuthenticationOptions sslAuthenticationOptions) in /_/src/libraries/System.Net.Security/src/System/Net/Security/SslStreamPal.Unix.cs:line 114
----- Inner Stack Trace -----

----- Inner Stack Trace #2 (System.AggregateException) -----

----- Inner Stack Trace -----
   at System.Net.Security.SslStream.ForceAuthenticationAsync[TIOAdapter](TIOAdapter adapter, Boolean receiveFirst, Byte[] reAuthenticationData, Boolean isApm) in /_/src/libraries/System.Net.Security/src/System/Net/Security/SslStream.Implementation.cs:line 393
----- Inner Stack Trace -----
   at Interop.OpenSsl.DoSslHandshake(SafeSslHandle context, ReadOnlySpan`1 input, Byte[]& sendBuf, Int32& sendCount) in /_/src/libraries/Common/src/Interop/Unix/System.Security.Cryptography.Native/Interop.OpenSsl.cs:line 282
   at System.Net.Security.SslStreamPal.HandshakeInternal(SafeFreeCredentials credential, SafeDeleteSslContext& context, ReadOnlySpan`1 inputBuffer, Byte[]& outputBuffer, SslAuthenticationOptions sslAuthenticationOptions) in /_/src/libraries/System.Net.Security/src/System/Net/Security/SslStreamPal.Unix.cs:line 114
----- Inner Stack Trace -----

[josalem]

@ericstj rebasing to make sure it has all the newest changes including the revert of that PR.

[wfurt]

The ssl should be fixed with #41372 for 5.0. If we still see failures I can investigate.

[wfurt]

osx failure is #36466

[Anipik]

@ericstj @safern is this a good time to merge this one ?

[ericstj]

Let's make sure to rerun CI and get fresh results.

[Anipik]

rebased to get the lastest runtime & running the ci

[Anipik]

Failures are known #43166