Subject Re: Again: Error on database backup
Author fabiano_bonin
--- In, "Alan McDonald" <alan@m...>
> > I'm posting this message again, since i had no answers to the
> > previous post and i couldn't solve this problem using the tools
> > supplied for this (gbak/gfix) nor following the IBPhoenix docs on
> > how to restore corrupt databases.
> >
> > I'm feeling a little bit unconfortable because it's the first
> > problem i had with FB/IB that i couldn't solve making a backup
and a
> > restore or using gfix, and because it makes 15 days that my
> > is working without a backup of his data.
> >
> > When i try to backup this database, i have the following error:
> >
> > gbak: writing data for table REM1
> > gbak: ERROR: message length error (encountered 58, expected 54)
> > gbak: ERROR: gds_$receive failed
> > gbak: Exiting before completion due to errors
> >
> > If i run gfix it doesn't report nothing.
> > I tryed to follow all steps of IBPhoenix doc, without success.
> > The server is a IBM eServer xSeries 200, working with no-breaks,
> > it's not supposed to have problems caused by abnormaly server
> > shutdowns.
> > It's running Linux RH 9.0, was running FB RC6 but now i installed
> > RC7, and the problem persists.
> >
> > Apart this problem, the database seems to be ok, the users are
> > working on it without problems.
> >
> > I'd like to have a clue on what can cause this type of erros and
> > what can i do to solve it.
> >
> > Thanks again.
> I would recommend creating another database, and pumping metadata
and data
> to the new shell. This will mean some downtime (maybe only short
if you
> practice on an old backup).
> Are you backing up the database regularly? Do you test the restore
> regularly?
> This sort of corruption is rare - maybe there is something you are
> like taking tape backups of the file?
> Alan

The server makes the backup all night.

After the backup, the same batch restore it with another name
(test.gdb), just to see if it's ok, and generate a log in case of
error messages.

The tape backup is made after that, just on the gbk file, not on the
gdb file.

I have almost 15 servers running basically the same configuration
and batches for some years and it's the first time i saw this error.

I solved the problem removing the table REM1 (after removing its
dependencies), so i could make the backup, and then i restored it
and recreated REM1 and its dependencies and data.

But i expected to have this problem solved by gbak or gfix, like all
my previous corruption problems, not manually.