Inconsistency in Galera cluster while running add drop column

The load was running on a single node in a 3 node cluster. 3 threads were executing DML with only concurrent add and drop column causes inconsistency between the nodes.

2023-11-08 7:16:13 0 [Note] WSREP: Votes over a52d165a-7e06-11ee-bede-afb796da4be7:33835:
f99cf43909114ded: 1/3
Waiting for more votes.
2023-11-08 7:16:13 0 [Note] WSREP: Member 1(node2) initiates vote on a52d165a-7e06-11ee-bede-afb796da4be7:33835,f99cf43909114ded:
2023-11-08 7:16:13 0 [Note] WSREP: Votes over a52d165a-7e06-11ee-bede-afb796da4be7:33835:
f99cf43909114ded: 2/3
Winner: f99cf43909114ded
2023-11-08 7:16:13 7 [Note] WSREP: Got vote request for seqno a52d165a-7e06-11ee-bede-afb796da4be7:33835
2023-11-08 7:16:13 0 [Note] WSREP: Recovering vote result from history: a52d165a-7e06-11ee-bede-afb796da4be7:33835,f99cf43909114ded

To reproduce run pstress with below option. for compling refer pstress

pstress-md --address=127.0.0.1 --user=test --password=testpass --database=pstress  --pk-prob=100 --no-enc --threads=3 --no-tbs --undo-tbs-count=0 --no-virtual --no-table-compression --row-format=none --grammar-sql=0 --no-column-compression --only-cl-ddl --add-column=10 --drop-column=10 --no-temp-tables --no-partition-tables --port=$PORT_NODE1 --seconds=20 --seed=08112023