Re: Patch to implement pg_current_logfile() function

On Wed, 23 Nov 2016 03:21:05 -0600
"Karl O. Pinc" <kop@meme.com> wrote:

> On Sat, 19 Nov 2016 12:58:47 +0100
> Gilles Darold <gilles.darold@dalibo.com> wrote:
>
> > ... attached v14 of the patch.

> patch_pg_current_logfile-v14.diff.retry_current_logfiles-part3

> Re-try the write of current_logfiles should it fail because the
> system is too busy.

Attached are 2 more doc patchs.

---

patch_pg_current_logfile-v14.diff.pg_ctl_and_more_docs

Applies on top of
patch_pg_current_logfile-v14.diff.retry_current_logfiles-part3

Mentions pg_ctl and how using it might affect whether logfile
paths are captured in the current_logfiles file.  I think
the default RedHat packaging captures logs this way.  The
Debian default is to ship with logging_collector=off
and to rely on pg_ctrl via pg_ctlcluster to manage
log writing.  Both RH and Debian then pass stderr to systemd
and that does log management.  I don't
recall other distros' practice.  If the default distro
packaging means that current_logfiles/pg_current_logfile()
"don't work" this could be an issue to address in the
documentation.  Certainly the systemd reliance on stderr
capture and it's subsuming of logging functionality could
also be an issue.

Cross reference the current_logfiles file in the pg_current_logfile()
docs.

Marks up stderr appropriately.

Adds index entries.

---

patch_pg_current_logfile-v14.diff.doc_linux_default

Applies on top of
patch_pg_current_logfile-v14.diff.pg_ctl_and_more_docs

Mentions, in the body of the docs, defaults and their impact on
current_logfiles and pg_current_logfiles.  It seems appropriate
to mention this in the main documentation and in the overall context
of logging.

Adds index entries.

Regards,

Karl <kop@meme.com>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein