Sergey Shinderuk

> could you make necessary changes - remove LockRelease() moving in receive_array() and pg_wait_sampling_reset_profile() functions? This moving is redundant as it doesn't protect collector_hdr->request shared variable but just is used...

``` +extern const LOCKTAG queueTag; +extern const LOCKTAG collectorTag; ... -extern void init_lock_tag(LOCKTAG *tag, uint32 lock); ``` Instead of init_lock_tag that clashes with the like named symbol in aqo, we...

@maksm90 I dropped the changes related to init_lock_tag from the patch and rewrote the commit message. Please take a look. Are you satisfied with the commit message?

My main concern is that these tests depend on timing of events. If the machine is temporarily lagging, they could fail. And this is totally possible in our buildfarm. In...

> Another option might be using some framework that allows controllable running of processes... Thanks for the links, good to know. > ...collector at least would take one sample for...

Regarding isolation tests with PostgreSQL < 14. Previously we discussed an option of building PostgreSQL from source and manually installing pg_isolation_regress and isolationtester in a CI build. After thinking some...

We use postgres:13-alpine image from docker hub for building and testing in Travis. In this image, there are no pg_isolation_regress and isolationtester binaries. To run isolation tests we need to...

Why do we need to maintain our own array of queryIds? Why can't we just read PgBackendStatus.st_query_id as pg_stat_activity does? It turns out st_query_id is zero during execution of a...

It turns out that queryId is always zero for prepared statements (extended protocol). This can be reproduced with `pgbench -Mprepared`. Maybe that was the real reason of complaints. We'll try...

Not sure this is relevant, but macOS supports so called diagnostic calls [1], and one of them, namely dgEnaPMC [2], enables RDPMC from user mode [3]. It can be called...