Hello,
the problem has reached the development support at IBM.
As soon as I have a solution for the problem I will setup a build farm
member (hopefully).
I can probably smoke HEAD on AIX 7.1 and 6.1.
Bye
Rainer
On 27.02.2014 08:21, Rainer Tammer wrote:
> Hello,
> I have checked this:
>
> >From prepared_xacts:
>
> LOG: Send kill to PID: 11731118
> LOG: Retrun value of kill PID: ret= 0
>
> From: plpgsql
>
> LOG: Send kill to PID: 11731120
> LOG: Retrun value of kill PID: ret= 0
>
> A kill -2 on the shell does not stop the backend process.
> A kill -15 does stop the backend process.
>
> So the kill does not return a failure.
>
> The code in /src/backend/storage/lmgr/proc.c - CheckStatementTimeout()
> of 8.4.20
> does not look much different. The code in pqsignal.c is also the not
> really different.
>
> Bye
> Rainer
>
> On 26.02.2014 18:00, Tom Lane wrote:
>> Rainer Tammer <pgsql@spg.schulergroup.com> writes:
>>> So we can be pretty sure that the SIGINT does not interrupt semop() on
>>> AIX 71.
>> Looks that way :-(. It would be interesting to add a check to see if the
>> kill() is returning a failure indication or not. The existing code
>> doesn't bother because there's not really anything it can do about it,
>> but for this purpose it would be good to know.
>>
>> regards, tom lane
>>
>>
>
>