gzip failing to unzip

[william drescher]

On 3/22/2017 2:30 PM, Xen wrote:
> william drescher schreef op 22-03-2017 19:06:
>> To find a low frequency failure in a production system I have been
>> taking periodic dumps (using phymyadmin-export) specifying that
>> the
>> file be gzipped.  It appears to have worked just fine.
>>
>> Today I went to unzip  one of the files to use it as a mysql
>> source on
>> the development system so I could dig into the problem and..
>>
>> gunzip reported:
>> gzip: IS-2016-12-31.sql.gz: decompression OK, trailing garbage
>> ignored
>>
>> The "trailing garbage" is the last 3/4 of the file.  The
>> gzipped file is 17MB.
>>
>> Anybody have any ideas what is going wrong and or suggestions
>> that I
>> can try to recover the data.
>
> I have only sometimes 'got' that when the downloaded file (over
> http) was bigger than the actual file, ie. it was some download
> error through erroneous size reporting of the server, I believe.
>
> But you seem to suggest that the file is actually _meant_ to be
> 17MB big, right? Well not actually. Oh yeah, I'm mistaken. If 75%
> of the time is garbage that is a tad much :p. But it sounds to me
> that the file ended correctly, there was just more data that was
> not actually data. Are you completely sure the resulting file is
> corrupted?
>
> Regards.
>
The original file is 67 Mb.
When I gunzip the file, the part that is unzipped appears correct 
but the rest of the file is missing.
It may be relevant that the original file has a large section 
that encrypted.  Perhaps there is a sequence that gunzip 
interprets as the end of file.

-bill