Same version of DB for dump & restore? If not, was the dump done via the pg_dump from the newer version. If not, please do that.
-- Dan Langille http://langille.org/ On Aug 28, 2013, at 2:56 AM, Torello Querci <tque...@gmail.com> wrote: > Interesting ..... > > while trying to restore the database on the same machine as different > database I get this error message: > > ERROR: date/time field value out of range: "20016009:50:37.927936" > > Since I get this data from a database dump obtained with "pg_dump" on the > same hardware I suppose that can to be two possibility: > > - postgresql bug somewhere > - hardware problem that caused data corruption > > Since the dump file is 11G is not so easy to handle .... > I think that this is not related with create index problem since this field > is not used by this index and increase maintenance memory had worked. > > I'll fix it and go ahead in maintenance_work_mem test for index creating. > > > Best Regards > > > 2013/8/27 Torello Querci <tque...@gmail.com> >> >> >> >> 2013/8/26 Jeff Janes <jeff.ja...@gmail.com> >>> On Mon, Aug 26, 2013 at 10:01 AM, Torello Querci <tque...@gmail.com> wrote: >>> > Ok, >>> > >>> > now create index is finished using maintenance_work_mem=100MB. >>> > >>> > Thanks to all. >>> > >>> > I suppose that an error message more clear can help. >>> >>> Unfortunately, since no one knows what the real problem is, we can't >>> make the message more clear. Something that is never supposed to >>> happen has happened. >>> >>> One thing you could do is set log_error_verbosity to verbose. >>> >>> It seems like the most likely cause is flaky hardware, either memory >>> or hard-drive. In which case, your database is in serious danger of >>> irrecoverable corruption. >>> >>> Is it reproducible that if you lower the maintenance_work_mem you get >>> the error again, and if you raise it the error does not occur? >>> >> >> I'll try to restore the database on the same hw but different DB using >> differente maintenance_work_mem end verbosity and I'll posted the result >> here, if can help to improve the error message. >> >> >> Cheers, Torello >> >