Strange db-files in working directory

Lou Goddard lgoddard at
Thu Dec 2 15:39:31 UTC 2004

I've seen these files after forgetting to set the
permissions correctly on the working directory of bind.
You could make sure that Bind has write access to its own


-----Original Message-----
From: bind-users-bounce at [mailto:bind-users-bounce at]On
Behalf Of Walkenhorst, Benjamin
Sent: Tuesday, November 30, 2004 10:15 AM
To: jim at
Cc: bind-users at
Subject: AW: Strange db-files in working directory=20


> They were created by the name server, It uses temporary files to store
> in-progress inbound zone transfers. When the zone transfer completes
> successfully, the temporary files get renamed to whatever was
> specified as the filename for the zone in the slave name server's
> configuration. These temporary files might be left behind when a
> zone transfer doesn't complete properly: for instance because the TCP
> connection breaks.

I was thinking of something like that... I can't really tell what might =
happened, there are no truncated lines in those files. Then again, they =
a little smaller than the actual zone file.
It might, however, be related to some permission-weirdness we had on =
machine, or maybe a timeout from the master...=20

>     Benjamin> 2. Is it safe for me to delete them?
> Yes, provided nothing has the files open. lsof is your friend.

I was able to remove both files without problems. =3D)

> BTW, these are Frequently Asked Questions. Please consult the list
> archives before posting. You may find that your question has been
> asked and answered many times already.

Sorry, I should have thought of that.

>     Benjamin> (Yes, I should upgrade; No, 9.3 isn't available for our
>     Benjamin> OS, yet...)
> Why? Doesn't the 9.3 source compile on your chosen OS?

To be honest, I haven't tried. =3D(
The machine runs z/OS, not a standard unix. I am not sure if there is a=20
C compiler installed; I am pretty sure I won't get permission to build =
BIND from
source. Also, IBM seems to have modified its version of BIND a =
little(1), so I am not sure
if - and how - it would work at all. Not to speak of nasty EBCDIC-issues =
I might encounter...
At least IBM has announced to port 9.3 to z/OS...

Thank you very much,

(1) Their version of BIND4 allowed you to include load-balancing into =
and their BIND9 allows for a BIND4-compatibility mode.

More information about the bind-users mailing list