Zdenek Kotala napsal(a):
> Alvaro Herrera napsal(a):
>> Zdenek Kotala wrote:
>>> I performed review and I prepared own patch which contains only
>>> probes without any issue. I suggest commit this patch because the
>>> rest of patch is independent and it can be committed next commit fest
>>> after rework.
>>>
>>> I found following issues:
>>
>> I noticed that CLOG, Subtrans and Multixact probes are added during a
>> regular Checkpoint, but not during a shutdown flush. I think the probes
>> should count that too (probably with the same counter).
>
> Yeah, good catch.
When I'm thinking about it, It seems to me better idea to have
TRACE_POSTGRESQL_CLOG_SHUTDOWN_START();
TRACE_POSTGRESQL_XLOG_SHUTDOWN_START();
Because you will able to determine what was a reason for flush and how log take
shutting down.
By the way why the shutdown order is following:
05617 CreateCheckPoint(CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_IMMEDIATE);
05618 ShutdownCLOG();
05619 ShutdownSUBTRANS();
05620 ShutdownMultiXact();
What does happen when kill -9/power lost comes between lines 5617 and 5618?
Zdenek