Am 30. September 2022 15:50:38 MESZ schrieb Wietse Venema <wie...@porcupine.org>: >Wietse Venema: >> Steffen Nurpmeso: >> > Steffen Nurpmeso wrote in >> > <20220929213725.gar4l%stef...@sdaoden.eu>: >> > |Viktor Dukhovni wrote in >> > | <yzxedjcbxxob5...@straasha.imrryr.org>: >> > ||On Thu, Sep 29, 2022 at 07:25:48PM +0200, Steffen Nurpmeso wrote: >> > ||> But on FreeBSD (only VM here for some years) on fresh install >> > ||> i always have sendmail hang minutes upon startup (i interrupt to >> > ||> come to login, which thankfully works), because it is of the >> > ||> opinion that the hostname is not valid on the network, or >> > ||> something. I do not have the same problem with postfix though. >> > || >> > ||By design. As Wietse has explained from time to time, Postfix starts up >> > ||and delivers local mail even when the network is down. Delivering mail >> > ||to remote systems of course requires some form of connectivity (dialup >> > ||uucp could still be used in principle). >> > | >> > |I think it has something to do with uname(2) that is used in order >> > |to query the nodename of the box. For my little mailer (MUA) >> > >> > That is, i have seen such hangs with my MUA -- whether sendmail >> > hangs due to this, i do not know. >> >> Sendmail hangs because it uses the nsswitch mechanism which may do >> DNS lookups, as it does in your example. > >I hate to post incorrect info, so here is a correction. When the >nsswitch lookup result (typically from /etc/hosts) has no name >with '.', then Sendmail will query the DNS directly, pausing 60s >after a DNS lookup failure. > > Wietse
First of all: THANKS to all who helped!!! Tried to check at first the queues, as Wietse suggested, but as this my own more or less privately (family) used mail server, all queues have been empty. Checked then system default for timeout of systemd units, was 180 secs. Set it to 10 mins, and postfix started also as systemd service unit again. Removed all -vvvv verbose flags, as suggested. Wanted to check network related startup timings at this weekend, but: Problem disappeared as suddenly as it popped up: postfix service now needs less than 5 secs to start, as before the issue. No idea why, probably something on the VPS's host system or the provider's network setup, as nothing changed on my VPS. So mystic, esp. as no other service like Dovecot, MariaDB, Apache,.... have been affected, thanks again to all! Michael
