Re: pg_restore out of memory

On 07/13/2016 01:51 PM, Miguel Ramos wrote:
>
>
> Finally, here are the log messages at the moment of the error.
> It is clearly not while building indices.
>
> The table in question is a big one, 111GB.
> Fields latitude, longitude and height are arrays of length around 500-
> 700 on each row (double and real).
>
> So, what does this mean?
> Was it the client that aborted? I think I saw that "unexpected message
> type 0x58" on other types of interruptions.

So where are you running the pg_restore from, manually from command line
or from within a script?

>
>
>
> Jul 13 20:10:10 ema postgres[97889]: [867-1] LOG:  statement: COPY positioned_scan (id_dataset, id_acquired_set,
sequence_number,id_scan_dataset, latitude, longitude, height, srid, srid_vertical) FROM stdin; 
> Jul 13 20:10:10 ema postgres[97889]: [867-2]
> Jul 13 21:08:06 ema postgres[97889]: [868-1] ERROR:  unexpected message type 0x58 during COPY from stdin
> Jul 13 21:08:06 ema postgres[97889]: [868-2] CONTEXT:  COPY positioned_scan, line 2779323
> Jul 13 21:08:06 ema postgres[97889]: [868-3] STATEMENT:  COPY positioned_scan (id_dataset, id_acquired_set,
sequence_number,id_scan_dataset, latitude, longitude, height, srid, srid_vertical) FROM stdin; 
> Jul 13 21:08:06 ema postgres[97889]: [868-4]
> Jul 13 21:08:06 ema postgres[97889]: [869-1] LOG:  could not send data to client: Broken pipe
> Jul 13 21:08:06 ema postgres[97889]: [869-2] STATEMENT:  COPY positioned_scan (id_dataset, id_acquired_set,
sequence_number,id_scan_dataset, latitude, longitude, height, srid, srid_vertical) FROM stdin; 
> Jul 13 21:08:06 ema postgres[97889]: [869-3]
> Jul 13 21:08:06 ema postgres[97889]: [870-1] LOG:  unexpected EOF on client connection
>
> Thanks,
>
> -- Miguel
>
>
> A Ter, 12-07-2016 às 15:10 -0400, Tom Lane escreveu:
>> Miguel Ramos <org.postgresql@miguel.ramos.name> writes:
>>> This because I have the impression that it is during index
>>> creation,
>>> where I think client role would be minimal.
>>
>> Hard to believe really, given the spelling of the message.  But
>> anyway,
>> be sure you do the run with log_statement = all so that it's clear
>> what
>> is being worked on when the error happens.
>>
>>             regards, tom lane
>
>
>
>


--
Adrian Klaver
adrian.klaver@aklaver.com