Horton
Menu

Re: Refactoring postmaster's code to cleanup after child exit

Heikki Linnakangas <hlinnaka@iki.fi>

View thread
From: Heikki Linnakangas <hlinnaka@iki.fi>
To: Andres Freund <andres@anarazel.de>
Cc: Thomas Munro <thomas.munro@gmail.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2024-11-14T14:13:17Z
Lists: pgsql-hackers 
On 09/10/2024 23:40, Heikki Linnakangas wrote:
> I pushed the first three patches, with the new test and one of the small
> refactoring patches. Thanks for all the comments so far! Here is a new
> version of the remaining patches.
> 
> Lots of little cleanups and changes here and there since the last
> versions, but the notable bigger changes are:
> 
> - There is now a BackendTypeMask datatype, so that if you try to mix up
> bitmasks and plain BackendType values, the compiler will complain.
> 
> - pmchild.c has been rewritten per feedback, so that the "pools" of
> PMChild structs are more explicit. The size of each pool is only stated
> once, whereas before the same logic was duplicated in
> MaxLivePostmasterChildren() which calculates the number of slots and in
> InitPostmasterChildSlots() which allocates them.
> 
> - In PostmasterStateMachine(), I combined the code to handle
> PM_STOP_BACKENDS and PM_WAIT_BACKENDS. They are essentially the same
> state, except that PM_STOP_BACKENDS first sends the signal to all the
> child processes that it will then wait for. They both needed to build
> the same bitmask of processes to signal or wait for; this eliminates the
> duplication.

Made a few more changes since last patch version:

- Fixed initialization in pmchild.c in single-user and bootstrapping mode
- inlined assign_backendlist_entry() into its only caller; it wasn't 
doing much anymore
- cleaned up some leftovers in canAcceptConnections()
- Renamed some functions for clarity, fixed some leftover comments that 
still talked about Backend structs and BackendList

With those changes, committed. Thanks for the review!

-- 
Heikki Linnakangas
Neon (https://neon.tech)

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix test name and username used in failed connection attempts

  2. tests: Don't fail due to high default timeout in postmaster/003_start_stop

  3. tests: Fix race condition in postmaster/002_connection_limits

  4. Pass MyPMChildSlot as an explicit argument to child process

  5. Assign a child slot to every postmaster child process

  6. Kill dead-end children when there's nothing else left

  7. Replace postmaster.c's own backend type codes with BackendType

  8. Use an shmem_exit callback to remove backend from PMChildFlags on exit

  9. Add test for dead-end backends

  10. Add test for connection limits

  11. Fix garbled process name on backend crash

  12. Consolidate postmaster code to launch background processes

  13. Fix comment on processes being kept over a restart

  14. Refactor code to handle death of a backend or bgworker in postmaster

  15. Make BackgroundWorkerList doubly-linked

  16. Minor refactoring of assign_backendlist_entry()

  17. Fix outdated comment; all running bgworkers are in BackendList

  18. Allow building with MSVC and Strawberry perl

  19. Fix rare failure in LDAP tests.

  20. Allow notifications to bgworkers without database connections.