Andrew McGlashan wrote: > Yes, that's what I meant, sysvinit is not broken. I rather agree. But the opponents cite corner cases where the previous security model doesn't handle every possible access case. I always hate it when people say such vague statements such as "modern" or "is broken" without actually saying why it is one way or the other. After reading months of arguments these next two postings were the first real postings I had read with any detail in them. Especially the second one. https://lists.debian.org/debian-devel/2014/06/msg00455.html https://lists.debian.org/debian-devel/2014/06/msg00461.html These are things that probably 99.44%[1] of the population hasn't ever needed before. The 99% where everything works for us are all of us crying about the disruption. But for that 0.56% that worried about those corner cases they see the old system as really broken. They are probably right that it is broken for them. But there are better ways to go about improving the system than the unpleasant way that systemd has been rolled out to the community. I would really like to read more reasonable postings of such details such as Simon's postings cited above. > Sure it counts, but if you have 1000s of servers, you likely have many > other considerations and you'll be pooling [at least] those servers in a > cluster type arrangement ... much lessening the need for any machine to > startup so quickly. How many people here work with RHEL / CentOS? Have you booted a CentOS machine lately? I mostly work with Debian and I have come expect a reasonably quick boot time (using sysvinit). Most of my machines reboot in around 30 seconds using sysvinit and about 10 seconds of that is BIOS POST time which is independent of the system. Even if booting were twice as fast it would only reduce my boot times from 30 seconds down to 20 seconds. There is so little time spent there that it can't be a win for me. But when I started rebooting CentOS 5.x and 6.x systems I was shocked at how long it takes to reboot there! It takes several minutes for a fast 8-core workstation to boot. Minutes! Maybe five or six minutes in some cases. Wow. I find it truly shocking how long it takes to boot RHEL / CentOS. Debian with sysvinit boots hugely faster. And so I wonder. Is the entire reason for the push for reboot speed with systemd due to the problems of RHEL having an extremely slow boot? Maybe. Maybe if RHEL / CentOS hadn't been that slow to boot then it wouldn't have caused an itch to scratch it. Maybe if they had been using Debian with the already quite fast boot times they wouldn't have felt compelled to rewrite everything. Maybe. It is one of those possible alternate history stories. Bob [1] Ivory soap advertises itself as 99.44% pure soap. I often use that number when I am just declaring something without actual data. I don't know what the number is but is certainly large.
Attachment:
signature.asc
Description: Digital signature