Horton

Re: Is Recovery actually paused?

Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>

View thread View on postgr.es
From: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
To: Dilip Kumar <dilipbalaut@gmail.com>
Cc: Kyotaro Horiguchi <horikyota.ntt@gmail.com>, Yugo Nagata <nagata@sraoss.co.jp>, Masahiko Sawada <sawada.mshk@gmail.com>, Robert Haas <robertmhaas@gmail.com>, Simon Riggs <simon@2ndquadrant.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-02-07T13:14:31Z
Lists: pgsql-hackers 
On Fri, Feb 5, 2021 at 10:14 AM Bharath Rupireddy
<bharath.rupireddyforpostgres@gmail.com> wrote:
> > We can not do that, basically, under one lock we need to check the
> > state and set it to pause.  Because by the time you release the lock
> > someone might set it to RECOVERY_NOT_PAUSED then you don't want to set
> > it to RECOVERY_PAUSED.
>
> Got it. Thanks.

Hi Dilip, I have one more question:

+        /* test for recovery pause, if user has requested the pause */
+        if (((volatile XLogCtlData *) XLogCtl)->recoveryPauseState ==
+            RECOVERY_PAUSE_REQUESTED)
+            recoveryPausesHere(false);
+
+        now = GetCurrentTimestamp();
+

Do we need  now = GetCurrentTimestamp(); here? Because, I see that
whenever the variable now is used within the for loop in
WaitForWALToBecomeAvailable, it's re-calculated anyways. It's being
used within case XLOG_FROM_STREAM:

Am I missing something?

With Regards,
Bharath Rupireddy.
EnterpriseDB: http://www.enterprisedb.com

Commits

  1. Be clear about whether a recovery pause has taken effect.

  2. Pause recovery for insufficient parameter settings

  3. Remove most volatile qualifiers from xlog.c