> On 23 Mar 2017, at 21:59, Noel Jones <njo...@megan.vbhcs.org> wrote: > > >> >> maybe up the loglevel, or use tcpdump to capture some packets and >> see if the postfix logs are correct. >> > > Increasing the postfix log level is unlikely to give any further > useful information -- the other end dropped the connection. > > Check the amavisd logs at this same time. If that doesn't provide > anything useful, maybe increasing the amavisd debug level will help.
Well, thanks. That at least turns up something: Mar 22 21:31:14 Dumbledore.local /Applications/Server.app/Contents/ServerRoot/usr/bin/amavisd[279]: sd_notify (no socket): STATUS=Starting child process(es), ready for work. Mar 22 21:31:14 Dumbledore.local /Applications/Server.app/Contents/ServerRoot/usr/bin/amavisd[13425]: (!!)TROUBLE in child_init_hook: BDB no dbS: Lock table is out of available locker entries, . at (eval 94) line 31. Mar 22 21:31:14 Dumbledore.local /Applications/Server.app/Contents/ServerRoot/usr/bin/amavisd[13425]: (!)_DIE: Suicide in child_init_hook: BDB no dbS: Lock table is out of available locker entries, . at (eval 94) line 31. I looked at amavisd, but there is no code at line 31 or 94, so I’m kind of stumped. And also interesting: after a logfile turnover it went on operating normally, it seems. > Is this message still in the postfix queue or did it eventually get > delivered? II haven’t been able to establish this yet. It’s hard to debug with Apple’s logging issues. Th spool directory is good as empty (only one entry in defer/deferred. G
