I'm starting to get a handle on pkg's and ports, along with setting up a poudriere server to host my own pkg repo. I think using pkg's makes the most sense for me, along with poudriere to build my own custom packages.
I'm playing around in a home lab for now with maybe 4-6 freeBSD servers. I've been able to do everything I want just fine using the packages, however, I did run into a challenge with a certain package (zabbix3-server) as the default configuration uses MySQL and my preference would be PostgreSQL, as well as a missing library that would allow it to speak with my ESXi server.
Now I could run with the pkg and things would work fine, it just wouldn't be my ideal (which is why we have ports right?). This leaves me with a couple of options I'm aware of:
Option 3 is the most robust and makes the most sense, but it also seems like it may be more work for the limited number of servers I'm running at home. Running a server just to manage packages for 4-6 machines is overkill no?
Any thoughts or wisdom from those of you that have been doing this for longer then me on a good approach that works today and for years to come?
I'm playing around in a home lab for now with maybe 4-6 freeBSD servers. I've been able to do everything I want just fine using the packages, however, I did run into a challenge with a certain package (zabbix3-server) as the default configuration uses MySQL and my preference would be PostgreSQL, as well as a missing library that would allow it to speak with my ESXi server.
Now I could run with the pkg and things would work fine, it just wouldn't be my ideal (which is why we have ports right?). This leaves me with a couple of options I'm aware of:
- I can use pkg's on all my servers except for the server running Zabbix, and on that server I could use ports to customize the configuration
- I can use pkg's on all my servers and accept the default MySQL database as well as the lack of monitoring ESXi.
- I can setup a poudriere server, build all of the ports I need on that machine, and use that as a source of packages for all my remaining servers
- Use ports on all servers
Option 3 is the most robust and makes the most sense, but it also seems like it may be more work for the limited number of servers I'm running at home. Running a server just to manage packages for 4-6 machines is overkill no?
Any thoughts or wisdom from those of you that have been doing this for longer then me on a good approach that works today and for years to come?