Horton

Re: backend type in log_line_prefix?

Mike Palmiotto <mike.palmiotto@crunchydata.com>

View thread
From: Mike Palmiotto <mike.palmiotto@crunchydata.com>
To: Fujii Masao <masao.fujii@oss.nttdata.com>
Cc: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>, Alvaro Herrera <alvherre@2ndquadrant.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2020-03-17T19:03:22Z
Lists: pgsql-hackers

Commits

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

  1. Add backend type to csvlog and optionally log_line_prefix

  2. Remove am_syslogger global variable

  3. Unify several ways to tracking backend type

  4. Refactor ps_status.c API

  5. Add background worker type

> On 2020/03/15 19:32, Peter Eisentraut wrote:
> > On 2020-03-13 22:24, Peter Eisentraut wrote:
> >> On 2020-03-10 19:07, Alvaro Herrera wrote:
> >>> I like these patches; the first two are nice cleanup.
> >>>
> >>> My only gripe is that pgstat_get_backend_desc() is not really a pgstat
> >>> function; I think it should have a different name with a prototype in
> >>> miscadmin.h (next to the enum's new location, which I would put
> >>> someplace near the "pmod.h" comment rather than where you put it;
> >>> perhaps just above the AuxProcType definition), and implementation
> >>> probably in miscinit.c.
> >>
> >> I have committed the refactoring patches with adjustments along these
> >> lines.  The patch with the log_line_prefix and csvlog enhancements is
> >> still under discussion.
> >
> > I have committed that last one also, after some corrections.

Sorry for being late to this thread, but was wondering if anyone had
taken a look at the Process Centralization patchset that I submitted
to this CF:
https://www.postgresql.org/message-id/CAMN686HgTVRJBAw6hqFE4Lj8bgPLQqfp1c-%2BWBGUtEmg6wPVhg%40mail.gmail.com

There's quite a bit of that code that is in the same vein as the
MyBackendType changes proposed/merged in this thread.

I think we could reduce a large portion of redundant code (including
the pgstat_get_backend_desc code) while also
centralizing/standardizing process startup. A few useful features
(outside of code reduction) include the ability to identify backends
prior to their Main functions, cleaner control of SubPostmasterMain
logic (including implicit handling of shmem timing considerations).

If others think it's worthwhile, I will work on rebasing those changes
on the changes proposed/merged in this thread (re: MyBackendType).

Thanks,
-- 
Mike Palmiotto
https://crunchydata.com