Stefan M. a écrit :
> On Wed, 27 Jun 2007 01:56:14 GMT, "El Maximo" <nobody@nowhere.org>
> wrote:
>
>> "Frédéric STOCK" <f.stock@jmontblanc.fr> wrote in message
>> news:4680d9a4$0$21149$7a628cd7@news.club-internet.fr...
>>> Hello,
>>>
>>> on new server: running SCO Unix OSR 5.06a, P4 xeon dual-core
>>> 2 Gb ddram
>>> 4 73Gb scsi hdd in raid-5
>>> Progress V8.3C 11
>>>
>>> multi-volume database
>>> 15 512mb vol + 1 free vol.
>>>
>>> I made a online probkup of this database from our actual server (bi P3
>>> xeon, SCO Unix OSR5.06a, Progress V8.3C11).
>>> I created the multi-volume database on the new server using the prostrct
>>> utility and restore it.
>>> When trying to rebuild indexes using "proutil database_name -TM 31 -TB
>>> 31 -G 0 -C idxbuild all" (same error in all indexes or some), during phase
>>> 1 block 3799040 of 4672721 it fails
>>>
>>> "SYSTEM ERROR: erreur E/S 0 sur writeto,ret 1536,fich. 6(srt),addr
>>> 2147481600 (290)"
>>>
>>> What's this 1536 error code ?
>>> How could I resolv it ?
>>> Thank's.
>>>
>>> F. STOCK
>> The 2147481600 makes me think you hit a 2GB limit somewhere.
>>
>> Post on the peg and you'll have an answer by morning. (www.peg.com)
>>
>
> and you have not or a incorrect srt file ..

could it be a >2Gb temporary file ?

So today I made a dump data (one of the .d table file is near the 2Gb
file size limit (2049244868 bytes)
create multi-vol database
load data by bulkload
rebuild all indexes and I got the same error with the same addr 2147481600.

Where are then temporary srt files, so I could have a look of size
during indexes rebuild ?

Thank's.
F. STOCK