Flavors *COMPLETELY* break the port system (synth and poudriere are useless)

Baho Utot baho-utot at columbus.rr.com
Thu Dec 7 00:34:44 UTC 2017


On 12/6/2017 6:40 PM, Mel Pilgrim wrote:
> On 12/5/2017 2:25 PM, Baho Utot wrote:
>> Thank you for taking a perfectly good system and breaking it as well 
>> as making it unusable, unstable. You just don't know of all the 
>> countless hours spent after running an update and taking a week to 
>> get it working again.
>
> I manage (currently) 104 FreeBSD systems that are a mix of 10.x and 
> 11.x, on-metal and VM.  I use freebsd-update and poudriere. The 
> biggest issue in two years was back before I used poudriere and it was 
> when the default version postgresql change I had to upgrade databases 
> (something for which postgresql doesn't have automated tools).
>
> The last time I had a major stability problem with FreeBSD was when I 
> had a brand new Nocona Xeon system that would get interrupt storms 
> running 5.x and had to run 6-CURRENT on it for a while because 6.x 
> introduced MSI support.
>
> The closest thing to a FreeBSD failure after that was Perl upgrades 
> exploding all over the place because there was a time when the Ports 
> Tree didn't handle Perl upgrades gracefully.
>
> Looking back at two years of stats, all of the unplanned uptime resets 
> are associated with hypervisor or power outages.
>
> Experiences will vary, and it's totally fine if FreeBSD is not your 
> cup of tea, but you're out of line condemning the project as a whole 
> because your individual experience differed.
>
>> It really helps motivate all of us users to continue to have to fix 
>> broken systems due to broken ports system and then be told how great 
>> things are, brings us so much joy and keeps our attitude positive.
>
> I really don't know what advice to offer you other than that, from my 
> observations, there is a very high correlation between people who have 
> chronic stability problems with FreeBSD and people who insist on not 
> using the officially endorsed tools and methods.
>
> (Cue people utterly failing to understand statistics and citing single 
> data points in 3, 2, 1...)

Look at the so called "packaged base" it is nothing but an ungodly mess 
to put it mildly.  FreeBSD should have finished that project before 
starting all these "other" projects.  If you can not have a rock solid 
packaged base you really don't have much to stand upon. Hell when I 
tried "packaged base" and wanted to return to "standard base" No one 
here could tell me how to de-populate  the pkgng database of the 
packaged base entries.  I found a way that was trivaly to do that.  If 
the experts here can not or do not have a way to back out packaged base 
that says a whole lot about the credability of FreeBSD as a whole.

(Cue FreeBSD utterly failing run on newer hardware 3, 2, 1...)





More information about the freebsd-ports mailing list