<div dir='auto'><div><br><div class="gmail_extra"><br><div class="gmail_quote">On Dec 9, 2020 10:51 AM, Scott Dowdle <dowdle@montanalinux.org> wrote:<br type="attribution"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><p dir="ltr">Greetings,
<br>
<br>
----- Original Message -----
<br>
> Any kernel device drivers, for a start. Kind of critical if your
<br>
> SAS/RAID device wont boot, or your network device doesn't come up, or
<br>
> your GUI doesn't start because your display drivers aren't compatible
<br>
> anymore. Just minor things like that maybe?
<br>
<br>
If RHEL 8.4 works, and CentOS Stream 8.4+ makes a change that breaks something... that's what is going to become RHEL... so does it matter THAT much if it breaks a month or two earlier than in the past? Now answering your question with a question, how about another? Did you have much breakage as a result of EL minor updates? I assume the answer is no... and if no is the case, then you shouldn't have much breakage with Stream.
</p></blockquote></div></div></div><div dir="auto"><br></div><div dir="auto">I have hardware on ELrepo drivers; specifically at $dayjob we have a couple of nice, but older, Dell PowerEdge R710s that were donated to us ($dayjob is a 591(c)(3) nonprofit). The kmod-megaraid_sas driver, needed for the PERC H710 used by these R710s, distributed by ELrepo because RH decided they didn't want the hassle of supporting that hardware, DOES and DID change for each point release; I actually needed to enable elrepo-testing to get the version for the kernel in 8.3(2011). I don't know if that kmod is even available for Stream.</div><div dir="auto"><br></div><div dir="auto">The R710s are beefy enough servers for what we need (dual hex core, half a dozen virtual machines each, including a couple of Windows guests each, 96GB RAM). I just simply don't have the capital budget to replace them with "supported" new servers; one new server would be half of my entire operating budget for repairs and parts, and that includes UPS and HVAC maintenance. And I definitely don't have the recurring expense budget for something like RHEV or the virtualization add-on to support Windows guests.</div><div dir="auto"><br></div><div dir="auto">But the idea that it wouldn't be that difficult to wait another $indefinite_amount_of_time for a driver for the boot device is ludicrous, especially when it is consistently true that every roll-up release breaks compatibility (your assumption about no breakage is the exact and polar opposite of reality; even a cursory look at the ELrepo repository contents will show that clearly).</div><div dir="auto"><br></div><div dir="auto">Debian is looking more attractive, but I have a lot of years of invested effort in the RH ecosystem (since RHL 4.1 in 1997); not that different of a system in principle, but transitioning to it is not something I look forward to. I guess I need to take one of the older servers here and try out Debian's kvm support to see how well a Windows guest works.</div><div dir="auto"><br></div><div dir="auto">I also run CentOS on my daily-use laptop, a Dell Precision M6700 (on my 501(c)(3) salary I buy at least a couple of generations older, as there are some really good bargains for really nice hardware to be found). The ELrepo-packaged nVidia drivers are essential boot-time drivers, and yes they do need to change between the kernel changes of one update roll-up and the next, again I needed elrepo-testing enabled.</div><div dir="auto"><br></div><div dir="auto">It sounds to me like the actual use cases for current CentOS Linux simply weren't thought through fully enough by RH. Yes, I know the press release text mentions nonprofit organizations and making it easier to "consume" RHEL, but it is quite nebulous as to what that might look like. </div><div dir="auto"><br></div><div dir="auto">Further, it seems to me that the community wasn't asked about this change by RH because the likely set of answers from members of the user community was both already known and unwanted. For that matter, the same thing would probably be true about hardware support. </div><div dir="auto"><br></div><div dir="auto">And to top it all off, the RH CTO just flat out misinformed the community last September when he was quoted as saying that current CentOS wouldn't change. This especially is a breach of trust, in my opinion. RH can of course invest money into whatever projects it wishes, but when a clear statement of commitment to maintaining the status quo is made and then broken.... that's not ok.</div><div dir="auto"><br></div><div dir="auto">A lot to think about, and decisions to be made for sure. And CentOS 8 was just beginning, in my opinion, to mature and start being stable. A real shame.</div><div dir="auto"><div class="gmail_extra"><br></div></div></div>