Fix assert failure triggered by "CREATE INDEX CONCURRENTLY"

[fairyfar]

What does this PR do?

When resource groups are enabled, StartTransaction() may have taken a catalog snapshot
within GetResGroupIdForRole, and xmin is not InvalidTransactionId. Therefore, it does
not meet the status requirements for "CREATE INDEX CONCURRENTLY".

Reproduction method:

1. Compile Debug version and enable assert.

./configure --enable-debug --enable-cassert

2. Enable resource group management.

gp_resource_manager='group'

3. Test Case

create table t(a text, b text);
-- crash:
create index concurrently t_idx on t(a, b);
-- "ERROR:  relation "t_idx" already exists". Indicates that the transaction has been successfully committed.
create index concurrently t_idx on t(a, b);

4. Test Results

[fairyfar@bogon cloudberry]$ psql postgres
psql (16.9, server 16.9)
Type "help" for help.

postgres=# create table t(a text, b text);
NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Apache Cloudberry data distribution key for this table.
HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
CREATE TABLE
postgres=# create index concurrently t_idx on t(a, b);
FATAL:  Unexpected internal error (assert.c:46)
DETAIL:  ("MyProc->xid == InvalidTransactionId && MyProc->xmin == InvalidTransactionId", File: "indexcmds.c", Line: 5459)
server closed the connection unexpectedly
        This probably means the server terminated abnormally
        before or while processing the request.
The connection to the server was lost. Attempting reset: Failed.
The connection to the server was lost. Attempting reset: Failed.
!?> \q

[fairyfar@bogon cloudberry]$ psql postgres
psql (16.9, server 16.9)
Type "help" for help.

postgres=# create index concurrently t_idx on t(a, b);
ERROR:  relation "t_idx" already exists

Type of Change

• Bug fix (non-breaking change)
• New feature (non-breaking change)
• Breaking change (fix or feature with breaking changes)
• Documentation update

Breaking Changes

Test Plan

• Unit tests added/updated
• Integration tests added/updated
• Passed make installcheck
• Passed make -C src/test installcheck-cbdb-parallel

Impact

Performance:

User-facing changes:

Dependencies:

Checklist

• Followed contribution guide
• Added/updated documentation
• Reviewed code for security implications
• This PR contains AI-assisted code generation
• Requested review from cloudberry committers

Additional Context

CI Skip Instructions

---

[yjhjstz]

Can you add related test create index concurrently t_idx on t(a, b); ?

[fairyfar]

Yes, test cases have been added to resgroup_transaction.sql. However, compilation requires enabling enable-cassert.

[fairyfar]

I'm very sorry. The test case of ic-resgroup-v2 failed. It's related to this commit. Let me analyze the reasons.

[fairyfar]

I'm very sorry. The test case of ic-resgroup-v2 failed. It's related to this commit. Let me analyze the reasons.

The previous commit modified the GetResGroupIdForRole function to read user information from SysCache.
In concurrent scenarios, if one session drops a user while another session’s SysCache has not yet been refreshed,
the latter may still retrieve information for the dropped user. This behavior is incompatible with the design of
resource group.

Therefore, I re-evaluated the revisions and implemented a minimal fix. This revision releases the CatalogSnapshot
(setting xmin to InvalidTransactionId) immediately after reading the pg_authid catalog.