Summary:
To debug table OID collision, I added a debugging log so that each time
GetNewObjectId is called the newly allocated OID is logged.
When a table OID collision happens, if we see two identical OIDs allocated
to PG from the same database, then the bug is narrowed to the OID allocator's
algorithm.
Also made a few lint fixes.
Jira: DB-13220
Test Plan:
./bin/yb-ctl create --rf 1 --tserver_flags TEST_ysql_log_perdb_allocated_new_objectid=true
ysqlsh (15.2-YB-2.25.0.0-b0)
Type "help" for help.
yugabyte=# create table foo(id int);
CREATE TABLE
yugabyte=# \q
Look at postgres log:
I1122 18:43:48.581127 982937 varsup.c:714] allocated new object id 16384 in database 13515
I1122 18:43:48.603286 982937 varsup.c:714] allocated new object id 16385 in database 13515
I1122 18:43:48.607249 982937 varsup.c:714] allocated new object id 16386 in database 13515
Reviewers: fizaa
Reviewed By: fizaa
Subscribers: yql
Differential Revision: https://phorge.dev.yugabyte.com/D40220