In case somebody's just reading by and has a clue, don't hold back :-)

For the first time since working with Firebird I'm probably about
losing data.

Here's what I did to successfully break my precious:
- shutdown -f while 1 one user who told he's not working was in fact
still working with the database
- starting a backup which was expected to run some 7 minutes
- terminated the process after some 30 minutes (no CPU usage of
gBak.exe for more than 20 minutes)

Now an ODBC connection test works fine.
Connecting via Jiri's .net provider times out (CPU 50% for 2++ minutes)
Connecting via IBExpert times out (CPU 50% for 2++ minutes)

From 6 databases on that server, only 1 is making trouble.
I made a file copy of the offending file to C:\Current.fdb.

gfix -v -n C:\Current.fdb
results in "unavailable database" (user&pw supplied...)

gfix -v -n localhost:C:\Current.fdb
results in a new child process fb_inet_server.exe at 50% for a short
time (~2 seconds), a new blank line in the command window.

gbak -b -v C:\Current.fdb C:\Current.fbk
results in "unavailable database"

gbak -b -v localhost:C:\Current.fdb C:\Current.fbk
a new process gbak.exe with constantly 0%, a new file "C:\Current.fbk"
with 0 bytes, a new child process fb_inet_server.exe at 50% for a
loooooong time, some update to the database, because c:\current.fdb
gets a new file-timestamp.

ouput says:
"readied database... "
"creating file ... "
"starting transaction"

nothing more happens.

How would I best go about trying to fix this?

I have a good backup which is 4 hours old. There are worse things in
the world than having to revert back to that, but if possible I'd try
to restore this one.

FB 1.5 Classic Server on W2K3


TIA André