Subject Re: Recurring database corruption
Author Zoran Zivkovic
--- In, "Alan McDonald" <alan@...>
> > We also didn't have these problems untill recently. Now we are
> > backup of the database 6 times per day, :). I wrote about this
> > problem few days a go, but only response I got was "go to
> > interbase is buggy". It seems to me that we sometimes get
> > replys here. For me, even if I know IB6.0 has some limitations (or
> > bugs), I think if it worked ok for 5 years, pass on new version is
> > not "automatic solution" for some new problem.
> >
> > Anyway, I would appreciate if you find solution for this if you
> > send me private email with info what caused probelm to you.
> >
> > Regards,
> >
> > Zile
> FWIW - I agree. The old "upgrade to the latest version" is used too
often as
> a get out clause. I still have IB6 servers running (even IB5.6 and
BDE) and
> the issues you relate has nothing to do with the bugs in these old
> I have seen once before someone finally realise that someone, an
admin, had
> fooled around with forced writes and left them off after a restore.
> there been a restore recently with your systems?
> even with FW off, there shouldn't be corruptions multipke times a
day. I
> would next suspect an ethernet card. or some other hardware. Have
you tested
> them? Do the logs show any 10054 errors?
> Alan

Alan, my interbase.log is full of:

INET/inet_error: read errno = 10054

but to be honest, it was also case before these problems now have
with coruption of database.
Forced writes are ON, I am sure:

Database header page information:
Flags 0
Checksum 12345
Generation 12233
Page size 4096
ODS version 10.0
Oldest transaction 2475
Oldest active 7283
Oldest snapshot 6991
Next transaction 12227
Bumped transaction 1
Sequence number 0
Next attachment ID 0
Implementation ID 16
Shadow count 0
Page buffers 0
Next header page 0
Database dialect 1
Creation date Aug 10, 2006 9:52:44
Attributes force write

Variable header data:
Sweep interval: 20000

I suspect that it could be some hardare problem, becasue we had once
similar problem after upgrading server with new machine. I remember
than we had problem with memory. I will check this issue during the
weekend, and if I find something, I will advice after weekend.