Re: [pmg-devel] [PATCH package-rebuilds v2 1/1] fetchmail: improve shipped service file

["Max R. Carrara" <m.carrara@proxmox.com>]

On Wed Sep 24, 2025 at 8:05 PM CEST, Stoiko Ivanov wrote:
> fetchmail exits with exit-code 3 if:
> 'The user authentication step failed...' (see fetchmail(1)).
> This also includes the case if there are no accounts configured for
> fetching, e.g. if all accounts are configured with 'skip' instead of
> 'poll'. In PMG you get this when temporary disaling all configured
> accounts in the GUI.

s/disaling/disabling ;)

>
> So we simply should not consider an exit of 3 as failure.
> Additionally adapt the Restart value to 'on-failure' (else systemd
> tries restarting 5 times and gives up)
> see systemd.service(5).

But wouldn't this mean that if I e.g. temporarily disable all configured
accounts, `fetchmail.service` would exit with `3` and then never restart
again? Or am I mistaken here?

That being said, `fetchmail.service` refuses to start for me on my most
recent test VMs, even though I have added a dummy entry in the UI.
I might be holding it wrong, but `/etc/default/fetchmail` is never being
written to; it's empty for me.

# systemctl status fetchmail.service
○ fetchmail.service - fetchmail mail retriever agent
     Loaded: loaded (/usr/lib/systemd/system/fetchmail.service; enabled; preset: enabled)
     Active: inactive (dead) (Result: exec-condition) since Thu 2025-09-25 18:47:38 CEST; 6min ago
 Invocation: e87235808c58412883ec5c2ad55ab248
  Condition: start condition unmet at Thu 2025-09-25 18:47:38 CEST; 6min ago
       Docs: man:fetchmail(1)
    Process: 2239 ExecCondition=/bin/sh -c [ "$START_DAEMON" = "yes" ] (code=exited, status=1/FAILURE)
   Mem peak: 1.7M
        CPU: 5ms

Sep 25 18:47:38 pmg-9-beta-01 systemd[1]: Starting fetchmail.service - fetchmail mail retriever agent...
Sep 25 18:47:38 pmg-9-beta-01 systemd[1]: fetchmail.service: Skipped due to 'exec-condition'.
Sep 25 18:47:38 pmg-9-beta-01 systemd[1]: Condition check resulted in fetchmail.service - fetchmail mail retriever agent being skipped.


All of that aside, I think it would make more sense to keep
`Restart=always` but add `SuccessExitStatus=3`, as well as
`RestartSec=` to control how quickly the service is restarted [0].

Since the daemonized fetchmail sleeps for 300 seconds before running
again, we could maybe set `RestartSec=300`, so that systemd waits for
five minutes before trying to start `fetchmail.service` again. That way
the daemon isn't restarted immediately when it fails; or in other words,
mails are fetched every 5 minutes, even on failure.

Additionally, it might be favorable to explicitly set
`StartLimitIntervalSec=` and `StartLimitBurst=` [1] here, because the
default retry count is 5 (as you already noticed), in case setting
`RestartSec` alone is insufficient for some reason.

FWIW, you can find the defaults for the above two values using
`cat /etc/systemd/system.conf | grep StartLimit`

[0]: `man 5 systemd.service`
[1]: `man 5 systemd.unit`

>
> Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
> ---
>  pkgs/fetchmail/fetchmail-6.4.39/debian/fetchmail.service | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/pkgs/fetchmail/fetchmail-6.4.39/debian/fetchmail.service b/pkgs/fetchmail/fetchmail-6.4.39/debian/fetchmail.service
> index a6e3168..b7260ac 100644
> --- a/pkgs/fetchmail/fetchmail-6.4.39/debian/fetchmail.service
> +++ b/pkgs/fetchmail/fetchmail-6.4.39/debian/fetchmail.service
> @@ -21,7 +21,8 @@ User=fetchmail
>  Type=exec
>  # sort $OPTIONS after "-daemon 300" to allow overwriting the interval using $OPTIONS
>  ExecStart=/usr/bin/fetchmail --daemon 300 $OPTIONS --nodetach -f /etc/fetchmailrc --pidfile /run/fetchmail/fetchmail.pid
> -Restart=always
> +SuccessExitStatus=3
> +Restart=on-failure
>  
>  [Install]
>  WantedBy=multi-user.target



_______________________________________________
pmg-devel mailing list
pmg-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pmg-devel