Open olifre opened 4 years ago
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
If there are no better ideas (as wontfix
suggests?) should we document the hackish workaround?
I can certainly do a PR on this.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
Bumping this with a useless comment and spamming all subscribed watches since the issue persists.
In some cases, postfix is left in a broken state after a fresh machine installation. For that to happen, a mail needs to arrive for a target listed in
/etc/virtual
after the file is filled by augeas but before it ispostmap
ped.I observed the following timeline on CentOS 8 with the master branch of this module:
The system never recovers until either a reboot or restarting postfix.
My interpretation is that
/etc/virtual
is created with wrong initial permissions by Augeas, and this is fixed later whenvirtual.db
is created (and RHEL seems to ship an empty DB in the package by default). However, ifpostfix
has already looked at those files before that happens, we end up in a broken state.My current workaround is rather ugly:
(I use a separate
exec
to prevent a dependency cycle)Any ideas on how to fix this better?