FATAL: operator class "xxxx" does not exist for access method "btree" - Mailing list pgsql-hackers

From mao zhang
Subject FATAL: operator class "xxxx" does not exist for access method "btree"
Date
Msg-id CAHK7YHi5e=ngnX1DCWmv76cYNzN=UmrPqV29a=8D2oJXRi+MTg@mail.gmail.com
Whole thread Raw
List pgsql-hackers
Dear pgsql:
  When we adding a custom system table and defining an index for it at the same time, the code compilation is possible, but the following errors may occur when starting the database:
   ----------------------------------------------------------------------------------------------------------
    fixing permissions on existing directory /usr/local/pgsql/data ... ok
    creating subdirectories ... ok
    selecting dynamic shared memory implementation ... posix
    selecting default max_connections ... 100
    selecting default shared_buffers ... 128MB
    selecting default time zone ... Asia/Shanghai
    creating configuration files ... ok
    running bootstrap script ... 2023-07-19 09:40:47.083 CST [2808392] FATAL:  operator class "key_ops" does not exist for access method "btree"
    2023-07-19 09:40:47.083 CST [2808392] PANIC:  cannot abort transaction 1, it was already committed
    Aborted (core dumped)
    child process exited with exit code 134
    initdb: removing contents of data directory "/usr/local/pgsql/data"
---------------------------------------------------------------------------------------------------------------
 There are my steps as follows:
  1. add a new header file (pg_bm_client_global_keys_args.h) for the custom system table, the file path is : src/include/catalog/pg_bm_client_global_keys_args.h;
  2. the modified Makefile is: src/backend/catalog/Makefile, add the new header file at the label CATALOG_HEADERS of the file;
  3. make; make install
  4. when run the cmd : /usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data, the error will occur. 
This problem has been bothering me for a long time. Can you help me solve it?


Best Wishes.

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: pg_recvlogical prints bogus error when interrupted
Next
From: Michael Paquier
Date:
Subject: Re: ALTER TABLE SET ACCESS METHOD on partitioned tables