оЕПЕЯШКЙЮ: ONUNLOAD ?

["Shulzhenko Vasyl" <vasilis@softline.kiev.ua>]

-----хЯУНДМНЕ ЯННАЫЕМХЕ-----
нР: Helmut Leininger <h.leininger@bull.at>
цПСООШ: comp.databases.informix
дЮРЮ: 8 ЛЮПРЮ 1999 Ц. 16:55
рЕЛЮ: Re: ONUNLOAD ?


>Neil Truby wrote:
>
>> Bob Davis wrote in message
>> >Have you tried an onunload > 2Gb?  The limitation should be OS based and
>> not
>> >enforced by Informix.  The onunload command has you reference a device with
>> >the -t option. It does not know or care whether you are sending the data
>> >stream to a file, tape, or the bit bucket
>>
>> One might have thought this was good sense.  But my experience of HP-UX
>> (which of course may not apply to AIX) is that onunload DOES know.  So, even
>> in a file system with large files enabled the onunload will still bomb out
>> at 2GBytes, but has no such scruples if the output file is directed to the
>> underlying raw logical volume, or of course tape.  All of which leads me to
>> the empirical observation that this limit is being enforced by onunload, not
>> the OS.
>>
>> Anyway, as posted previously, onunload to raw disk is the way to go if you
>> can do it.
>>
>> Neil Truby
>> Londis Stores
>> Hampton Hill, UK
>
>I think the limit is enforced by both.
>1) to be able to write a file > 2GB on AIX 4.2 (and up) the file system must be
>created as "Large Files Enabled File System" (AIX restriction)
>2) the application writing the file must be compiled with the "LARGE_FILES"
>option to have a 64-bit size_t (for off_t). AFAIK, onunload does not write to a
>pipe but opens a file. And there this restriction applies (Informix is still
>using standard 32-bit size_t --> Max OFFSET + SIZE for a chunk < 2GB).
>
>Regards
>Helmut
>--
>Helmut Leininger
>
>Bull AG / Vienna
>Unix Support
>Email: h.leininger@bull.at
>       helmut.leininger@bull.net
>
>This opinion is mine and not necessarily that of my employer.
>No guarantees whatsoever.
>
>