Hacker News Reader

Comment by SomeoneFromCA | Hacker News Reader

This all stereotypucal nonsense though, about BSD being great for servers. BSDs are non-LTS, rolling systems, which severely limits their use for servers, where package stability is top priority. They are good for certain kind of servers, but not a replacement for Red Hat.

View on HN

>which severely limits their use for servers

Ahh i found the Oracle user. Redhat has not "more stable packets" but "more known bug's and eventual workarounds and less packets".

>They are good for certain kind of servers

Like Redhat is just good for a certain kind of servers, like that OracleDB for your Delphi-App ;)

BTW: I think 5 years for the base-system is LTS (FreeBSD) and NetBSD even longer, and no BSD's are not rolling-systems hence the release 10. The packets however are rolling or quarterly (for NetBSD and FreeBSD), and ~point for OpenBSD.

This is pointless argument, as you do not understand why industries like stale-like-rock packages of Debian and Red Hat LTS systems. No one cares much that base is LTS, what is important that third party _packages_ stay same for years and only receive security update. Essentially zero-masintenance install and forget for two years systems, not quarterly update-reboot-fix-breakage every 3 month. No go fo many server usage scenarios.

True 25 years in the industry and i still have not learn'd that actively maintaining and continually upgrading and testing systems is the better path then to install rocks in the closet and forget about them...then have to upgrade from PostgreSQL 7.5 to 15 in one week, oh and can you have a look at our nfs v1 server?

Damn our software is not working with Rhel9, can you ask that Software producer if they have it for Rhel9? What? They are out of business since 3 years? Ok let's ask Redhat if we can extend the support for another 5 years, and let the software rebuild..i calculate 10 millions, what about you?

Look Redhat is really good for IoT light-switches and such, with a expected life of 10 years..but not much more...oh and bank and insurance...but that's another story.

BTW: Do you think "Security-updates" are no-risk updates?

I worked in IT for 20+ years too, so you do need to flaunt your experience around.

Yeah, well, this is why you have 2 year update cycle, like Ubuntu, Debian or you have overlapping support periods like Redhat. You can update you system every once in a while, you know beforehand what will be in the next release and when it will happen. Not some insanity of rolling or semi-rolling packages like BSDs have.

Yes, security updates are always a lot less risky than quarterly FreeBSD updates. Quarterly does not receive all sec updates anyway, but do receive some. Worst of both worlds.

In fact I think, FreeBSD is not suitable for majority server purposes exactly due to terrible support cycle, despite clear technological advantages (ZFS, Jails etc.). If you limit yourself to some narrow uses, completely serviced by the base system, then yeah you may use it as server, but that is very niche use.

FreeBSD actually is a good replacement for Arch-like systems, for workstation uses, exactly due to its semi-rolling nature and separate base systems, this will cause less breakage compared to Arch.

You still don't understand what i meant. Breakage means you have not tested anything, it's your fault and your responsibility.

To stay near upstream is incredible healthy, software stacks are a living thing and should be treated like that, remember the Mainframe trap?

FreeBSD supports for example Java8, PostgreSQL12 and GCC4.8.

The port-tree has not just the latest upstream version available.

And why would i want a Red-hat-patched 6 y/o Nginx on my machine? No thanks.

>Yes, security updates are always a lot less risky than quarterly FreeBSD updates

Let's stop here when you start comparing apple with oranges.

https://debugpointnews.com/netbsd-uptime/?amp=1

Reply on news.ycombinator.com