Re: Inconsistency in startup process's MyBackendId and procsignal array registration with ProcSignalInit() - Mailing list pgsql-hackers

From Kyotaro Horiguchi
Subject Re: Inconsistency in startup process's MyBackendId and procsignal array registration with ProcSignalInit()
Date
Msg-id 20211012.170944.1956442434012935707.horikyota.ntt@gmail.com
Whole thread Raw
In response to Re: Inconsistency in startup process's MyBackendId and procsignal array registration with ProcSignalInit()  (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>)
List pgsql-hackers
At Mon, 11 Oct 2021 16:03:57 +0530, Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com> wrote in 
> On Mon, Oct 11, 2021 at 12:41 PM Kyotaro Horiguchi
> <horikyota.ntt@gmail.com> wrote:
> > (I'm not sure how the trouble happens.)
...
>  If some process wants to send the startup process SIGUSR1 with the
> PGPROC->backendId and SendProcSignal(PGPROC->pid, XXXXX,
> PGPROC->backendId) after getting the PGPROC entry from
> AuxiliaryPidGetProc(), then the signal isn't sent. To understand this
> issue, please use a sample patch at [1], have a standby setup, call
> pg_log_backend_memory_contexts with startup process pid on the
> standby, the error "could not send signal to process" is shown, see
> [2].

Thanks, I understand that this doesn't happen on vanilla PG.

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center



pgsql-hackers by date:

Previous
From: Bharath Rupireddy
Date:
Subject: Re: Inconsistency in startup process's MyBackendId and procsignal array registration with ProcSignalInit()
Next
From: "Andrey V. Lepikhov"
Date:
Subject: Make query ID more portable