control: tags -1 - moreinfo On Thu, 2025-05-22 at 19:17 +0000, Helge Kreutzmann wrote: > Am Wed, May 21, 2025 at 05:48:43PM +0100 schrieb Richard Lewis: > > cant hurt - you can also do this locally: > > > > systemctl edit logcheck > > # in the editor add: > > [Service] > > ExecStart=sleep 180 > > > > # im suggestibg a very long sleep, to give more chance to let exim > > do > > something - or to write its paniclog > > With this change (only) logcheck delivers only one mail and the exim > error message is gone. > > Maybe this can be shipped by default for Trixie?
Glad to hear that adding a sleep after the call to logcheck resolves the issue. My guess is that without a sleep, systemd is terminating child processes (ie, exim's queue processing), and something specific to your environment (hardware, load, etc) is unluckily killing exim at a bad point resulting in the errors you've been seeing. I'm planning an upload of logcheck 1.4.5 to unstable later today which will add a 10 second sleep after running logcheck to the systemd service. (This will reach trixie after the 20 day migration delay, but it's also trivial to install from unstable if you'd like the update sooner.) I'll opportunistically close this bug with that upload, but if it isn't fixed after you update to 1.4.5, please feel free to re-open it. Mathias
signature.asc
Description: This is a digitally signed message part
