+ * if track_activities is disabled, st_queryid should already have been
+ * reset
+ */
+if (!pgstat_track_activities)
+return;
The above two conditions can be clubbed together in a single condition.
2.
+/* ----------
+ * pgstat_get_my_queryid() -
+ *
+ * Return current backend's query identifier.
+ */
+uint64
+pgstat_get_my_queryid(void)
+{
+if (!MyBEEntry)
+return 0;
+
+return MyBEEntry->st_queryid;
+}
Is it safe to directly read the data from MyBEEntry without calling pgstat_begin_read_activity() and pgstat_end_read_activity(). Kindly ref pgstat_get_backend_current_activity() for more information. Kindly let me know if I am wrong.
Thanks and Regards,
Nitin Jadhav
On Mon, Apr 5, 2021 at 10:46 PM Bruce Momjian <bruce@momjian.us> wrote:
On Sun, Apr 4, 2021 at 10:18:50PM +0800, Julien Rouhaud wrote: > On Fri, Apr 02, 2021 at 01:33:28PM +0800, Julien Rouhaud wrote: > > On Thu, Apr 01, 2021 at 03:27:11PM -0400, Bruce Momjian wrote: > > > > > > OK, I am happy with your design decisions, thanks. > > > > Thanks! While double checking I noticed that I failed to remove a (now) > > useless include of pgstat.h in nodeGatherMerge.c in last version. I'm > > attaching v22 to fix that, no other change. > > There was a conflict since e1025044c (Split backend status and progress related > functionality out of pgstat.c). > > Attached v23 is a rebase against current HEAD, and I also added a few > UINT64CONST() macro usage for consistency.
Thanks. I struggled with merging the statistics collection changes into my cluster file encryption branches because my patch made changes to code that moved to another C file.
