Subject | Re: [firebird-support] Recurring database corruption |
---|---|
Author | Helen Borrie |
Post date | 2006-08-11T06:32:32Z |
At 04:21 PM 11/08/2006, you wrote:
database was inherited with Forced Writes off. Given the kind of
conditions people usually write PHP applications for, this risky
configuration has now become distinctly life-threatening....
To check, run gstat -h and look at the Attributes field. If "Force
Write" isn't there, then it's off.
./heLen
>I'm hoping that someone can give me ideas on the cause of recurring databaseAn educated guess (from the history and description) is that the IB
>corruption on a Firebird 1.5.2 database. We've used Interbase, and later,
>Firebird for over 10 years, and it has been a stable, dependable system.
>
>Within the last three months, I've had more problems with database
>corruption than in the previous 10 years. "What's different?" I ask myself.
>There are two new programmers hitting the database via PHP. That's the only
>significant difference that I'm aware of. The errors after validation
>tonight are:
>
>Summary of validation errors
>
> Number of index page errors : 6
> Number of database page errors : 3
>
>reported from IBExpert. The database validated with no errors less than 5
>days ago.
>
>I've checked the disk at a software level (this happens to be on a WinNT
>system, but we also use Linux) and everything checks out, so I'm assuming
>that there is some other problem, perhaps related tto the PHP stuff, which I
>don't control.
>
>I've never had these kinds of problems before, and now I'm having to rebuild
>the database weekly. The hardware is relatively modern, even if the O/S is
>not. Any ideas would be greatly appreciated.
database was inherited with Forced Writes off. Given the kind of
conditions people usually write PHP applications for, this risky
configuration has now become distinctly life-threatening....
To check, run gstat -h and look at the Attributes field. If "Force
Write" isn't there, then it's off.
./heLen