Il 05/07/2017 16:33, Adrian Klaver ha scritto:
> On 07/05/2017 01:05 AM, Moreno Andreo wrote:
>> Il 04/07/2017 20:51, Daniel Verite ha scritto:
>>> Tom Lane wrote:
>>>
>>>> Moreno Andreo <moreno(dot)andreo(at)evolu-s(dot)it> writes:
>>>>> So the hint is to abandon manual COPY and let pg_dump do the hard
>>>>> work?
>>>> If it is a newline-conversion problem, compressed pg_dump archives
>>>> would
>>>> be just as subject to corruption as your binary COPY file is.
>>> It's mentioned in [1] that the signature at the beginning of these
>>> files
>>> embed a CRLF to detect this newline-conversion problem early on,
>>> so I would expect COPY IN to stumble on a corrupted signature
>>> and abort earlier in the process, if that conversion occurred.
>>> Instead the report says it fails after a number of tuples:
>> Given what you said, can I assume it's a file transfer or an
>> hardware-driven (pendrive) problem?
>
> Daniel also mentioned the harddrive as a possible source of error. I
> would say monitoring where and when the issues appear may help with
> determining the source.
Yeah, trying to restore the same file on another machine should help
determine the possible culprit.