On Mon, Mar 31, 2014 at 11:37:42AM -0400, Deeztek Support wrote:
> I got the following error in one of our postfix servers this morning:
>
> fatal: open database /var/lib/postfix/smtpd_scache.db: Invalid argument
Why on earth do people routinely truncate log entries to leave out
the name of the daemon that is logging the message???
What Postfix service (daemon) logged this message? Was it after
a Postfix reload, restart, or some other time?
> This was preventing sending and receiving email. I ended up deleting the
> /var/lib/postfix/smtpd_scache.db file, restarted postfix and it started
> working again. This is the second time this has hapenned within the last 6
> months or so. Even though I know how to fix, I'm wondering why this is
> happening to begin with and to prevent it from hapenning in the future. Is
> there a bug of some sort? A quick google search didn't reveal anything of
> that sort. My environment is Ubuntu 10.04 and postfix 2.8.5.
Have you changed the master.cf entry for tlsmgr(8) from its default
configuration? It must have a process limit of 1.
tlsmgr unix - - n 1000? 1 tlsmgr
The database in question is not safe for multiple writers. Beyond
that you might have a buggy Berkeley DB, or SELinux blocking access.
Postfix tlsmgr(8) truncates the TLS session cache database whe it
restarts, so there should not be any errors even if the database
was previously corrupted.
--
Viktor.
