which should all be started/stopped together), previously "restart" would
restart each sub-daemon in turn, but actually it should stop all daemons
and only then start them again. Additionally, as suggested by ajacoutot,
stop the procedure and return an error if stopping one of the rc scripts
failed. ok ajacoutot@ rpe@
(as done for a couple of other ports; bacula was actually the one I wanted
to fix this on but missed the earlier commit ;)
move to MULTI_PACKAGEs (the no_bat PSEUDO_FLAVOR is still available).
This modifies the packages as:
bacula-client - no change
bacula-server - the server part, now unflavored and depends on a proper SQL package support
bacula-bat - no change
bacula-mysql - new, MySQL support for bacula-server
bacula-pgsql - new, PostgreSQL support for bacula-server
bacula-sqlite3 - new, SQLite3 support for bacula-server
All SQL support packages conflict with each others.
For users of mysql and sqlite3, a manual step will be needed to properly
upgrade bacula-server; sthen@ will take care of this in current.html.
Tested by myself and sthen@
ok sthen@
Some info at:
http://blog.bacula.org/p710/http://blog.bacula.org/release-7-0-2/
Now builds with our libtool(1).
While there's a huge amount of changes, existing setups "should" just
work without any modification.
tested by sthen@ and myself
ok sthen@, my new co-maintainer :-)
a new libmysqlclient non-blocking API which utilizes co-routines. The X86
specific GCC ASM co-routine support hid the fact that there was an issue.
The only fallback code so far is POSIX user contexts which OpenBSD does not
support.
Input from and Ok sthen@ jasper@
Adapted from NetBSD pkgsrc and FreeBSD ports.
@sample the script so that people can apply local modifications if
needed.
Tested on an LTO-4, 8 slots.
Why do all chio(1) behave differently!!!???
ok sthen@ merdely@ (maintainer)
works again.
Conflict with older versions of the client.
Make mtx-changer.conf @sample-able so we don't overwrite on update.
ok & help from sthen and aja
packages which depend on each other:
- only create the _bacula user in -main (client parts)
- only create the _bacula-sd user in -server
Change _bacula's home dir from /var/bacula to /nonexistent.
/var/bacula needs to be writable by the _bacula group and since
/usr/libexec/security actually checks that user home dirs are
group-writable, whereas /etc/security just checked if the user
home dir was a writable *file*(!), it whines every day for
every machine with bacula client installed. (This will only
affect new installations so if you're annoyed by these mails
too, just vipw and change it by hand).
ok ajacoutot@ merdely@
- don't use ports libtool, it passes invalid flags to the linker
- since libbaccats-${BACKEND} must have the same .so version as
libbaccats, use a variable to make sure they are kept in-sync
- mark as being SHARED_ONLY and sync/merge plists
