|From:||Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr>|
|To:||Yugo NAGATA <nagata(at)sraoss(dot)co(dot)jp>|
|Subject:||Re: Avoid stuck of pbgench due to skipped transactions|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
>>> I attached a patch for this fix.
>> The patch mostly works for me, and I agree that the bench should not be in
>> a loop on any parameters, even when "crazy" parameters are given…
>> However I'm not sure this is the right way to handle this issue.
>> The catch-up loop can be dropped and the automaton can loop over itself to
>> reschedule. Doing that as the attached fixes this issue and also makes
>> progress reporting work proprely in more cases, and reduces the number of
>> lines of code. I did not add a test case because time sensitive tests have
>> been removed (which is too bad, IMHO).
> I agree with your way to fix. However, the progress reporting didn't work
> because we cannot return from advanceConnectionState to threadRun and just
> break the loop.
> + /* otherwise loop over PREPARE_THROTTLE */
> I attached the fixed patch that uses return instead of break, and I confirmed
> that this made the progress reporting work property.
I'm hesitating to do such a strictural change for a degenerate case linked
to "insane" parameters, as pg is unlikely to reach 100 million tps, ever.
It seems to me enough that the command is not blocked in such cases.
|Next Message||Michael Paquier||2021-06-14 06:48:13||Re: Fix dropped object handling in pg_event_trigger_ddl_commands|
|Previous Message||Tom Lane||2021-06-14 06:40:53||Re: [bug?] Missed parallel safety checks, and wrong parallel safety|