[Bug]: build index hang if trying to build scann index on a milvus running with simd sse4_2

[yanliang567]

Is there an existing issue for this?

• I have searched the existing issues

Environment

- Milvus version:master-20231021-d2dbbbc1
- Deployment mode(standalone or cluster):

Current Behavior

build index hang if trying to build scann index on a milvus running with simd sse4_2

Expected Behavior

As sse4_2 does not support building scann index, milvus shall return a meaningful error immediately, instead of hang

Steps To Reproduce

1. run milvus with simd type see4_2
2. create a collection and insert some data
3. try to build scann index

Milvus Log

Milvus shall handel this error and return a meanful error to clients
[2023/10/24 02:46:53.103 +00:00] [ERROR] [indexnode/task.go:356] ["failed to build index"] [error="failed to create index, C Runtime Exception: => failed to build index, the current index is not supported on the current CPU model\n: internal code=2001: segcore error"] [errorVerbose="failed to create index, C Runtime Exception: => failed to build index, the current index is not supported on the current CPU model: internal code=2001: segcore error\n(1) attached stack trace\n -- stack trace:\n | github.com/milvus-io/milvus/pkg/util/merr.WrapErrSegcore\n | \t/go/src/github.com/milvus-io/milvus/pkg/util/merr/utils.go:737\n | [...repeated from below...]\nWraps: (2) failed to create index, C Runtime Exception: => failed to build index, the current index is not supported on the current CPU model\nWraps: (3) attached stack trace\n -- stack trace:\n | github.com/milvus-io/milvus/pkg/util/merr.WrapErrSegcore\n

Anything else?

No response

[yanliang567]

/assign @czs007
/unassign

[yanliang567]

pods on devops:

yanliang-sse4-2-182769-etcd-0                                  1/1     Running       0                 15h     10.102.7.63     devops-node11   <none>           <none>
yanliang-sse4-2-182769-milvus-standalone-ff74cc996-ph9fm       1/1     Running       0                 15h     10.102.10.38    devops-node20   <none>           <none>
yanliang-sse4-2-182769-minio-97cfc5445-9h4w6                   1/1     Running       0                 15h     10.102.9.87     devops-node13   <none>           <none>

[czs007]

The process of index construction is asynchronous, and it is not possible to directly return this error through the "create_index" function. Additionally, it seems inappropriate to set the index status as "Failed" at the moment. In the upcoming version 2.3.3, a new feature will be added. If there are multiple retry attempts, the index status will be set as "disabled." After resolving the issue through subsequent operational means, the index construction can be triggered again using the "enable" command.

[yanliang567]

change the milestone to 2.4 as comments above