On Tue, Dec 25, 2012 at 7:14 AM, Michael Mol <mike...@gmail.com> wrote: > > On Dec 25, 2012 3:04 AM, "Canek Peláez Valdés" <can...@gmail.com> wrote: >> >> On Tue, Dec 25, 2012 at 1:38 AM, G.Wolfe Woodbury <redwo...@gmail.com> >> wrote: >> [ snip ] >> > From what has been happening with the systemd stuff, I do not see what >> > advantages it really offers over the SysV scheme and its successors like >> > OpenRC. Someone enlighten me please? >> >> I wrote the following some months ago; I think nothing much has >> changed since then (I added a couple of comments): >> >> Take this with a grain (or a kilo) of salt, since I'm obviously >> biased, but IMHO this are systemd advantages over OpenRC: >> >> * Really fast boot. OpenRC takes at least double the time that systemd >> does when booting, easily verifiable. In my laptop systemd is twice as >> fast as OpenRC; in my desktop is three times faster. (With a solid >> state hard drive, my laptop now boots even faster). >> >> * Really parallel service startup: OpenRC has never been reliable on >> parallel service startup; its documentation says it explicitly. Some >> will tell you that for them "it works", but just like the guys who >> have a separate /usr and refuse to use an initramfs, they just haven't >> been bitten by the inherent problems of it (just ask kernel developer >> Greg Kroah-Hartman). The Gentoo devs recognize that OpenRC is just >> broken with parallel service startup. >> >> * Really simple service unit files: The service unit files are really >> small, really simple, really easy to understand/modify. Compare the 9 >> lines of sshd.service: >> >> $ cat /etc/systemd/system/sshd.service >> [Unit] >> Description=SSH Secure Shell Service >> After=syslog.target >> >> [Service] >> ExecStart=/usr/sbin/sshd -D >> >> [Install] >> WantedBy=multi-user.target >> >> with the 84 of /etc/init.d/sshd (80 without comments). >> >> * Really good documentation: systemd has one of the best >> documentations I have ever seen in *any* project. Everything (except >> really new, experimental features) is documented, with manual pages >> explaining everything. And besides, there are blog posts by Lennart >> explaining in a more informal way how to do neat tricks with systemd. >> >> * Really good in-site customization: The service unit files are >> trivially overrided with custom ones for specific installations, >> without needing to touch the ones installed by systemd or a program. >> With OpenRC, if I modify a /etc/init.d file, chances are I need to >> check out my next installation so I can see how the new file differs >> from the old one, and adapt the changes to my customized version. >> >> * All the goodies from Control Groups: You can use kernel cgroups to >> monitor/control several properties of your daemons, out of the box, >> almost no admin effort involved. >> >> * It tries to unify Linux behaviour among distros (some can argue that >> this is a bad thing): Using systemd, the same >> configurations/techniques work the same in every distribution. No more >> need to learn /etc/conf.d, /etc/sysconfig, /etc/default hacks by >> different distros. >> >> * Finally, and what I think is the most fundamental difference between >> systemd and almost any other init system: The service unit files in >> systemd are *declarative*; you tell the daemon *what* to do, not *how* >> to do it. If the service files are shell scripts (like in >> OpenRC/SysV), everything can spiral out of control really easily. And >> it usually does (again, look at sshd; and that one is actully nicely >> written, there are all kind of monsters out there abusing the power >> that shell gives you). >> >> These are the ones off the top of my head; but what I like the most >> about systemd is that it just works, and that it makes a lot of sense >> (at least to me). >> >> Most of systemd features can be implemented in OpenRC, although the >> speed difference will never be eliminated if OpenRC keeps using shell >> files; however, Luca Barbato said that using reentrant busybox the >> speed difference is greatly reduced (I haven't confirmed this, since I >> haven't even installed OpenRC in months). >> >> Now, this set of (IMO) advantages of systemd over OpenRC pile up over >> the advantages of OpenRC over SysV: the most important one (I believe) >> is that OpenRC has dependencies, so a service starts only when another >> has already started. AFAIK, SysV has lacked this since always. >> >> I don't think I have ever heard anyone saying that we should keep >> using SysV; like a lot of Unix legacies, it should just die. OpenRC is >> much better, but it still uses a Turing-complete language (and a >> really slow one) to simply tell services when to start and when to >> stop, and it doesn't reliably keep track of what services are really >> still running (anyone who has ever used the "zap" command in OpenRC >> knows this). >> >> systemd of course has dependencies, a reliable tracking of service >> status (thanks in part to the use of cgroups), and its service files >> can't enter in an infinite loop. >> >> Hope it helps. >> >> Regards. >> -- >> Canek Peláez Valdés >> Posgrado en Ciencia e Ingeniería de la Computación >> Universidad Nacional Autónoma de México >> > > Thank you. I think that may well be the cleanest set of favorable arguments > I've seen for systemd. > > Now, question: could I not create a "/usr" service and make things dependent > on /usr come after it's been mounted? That seems the single, core missing > piece.
I'm not going into that. People with far more knowledge, experience, and communication hability than me (Greg, William, Michał, Diego, etc.) have explained why the "separated /usr without initramfs" is broken, and that udev/systemd actually *works* in a separated /usr *without* an initramfs (if you use the --with-rootprefix= configure option, which exists exactly for this), but that's not enough for a lot of folks. To be honest, I don't think nothing will be ever be enough, because they are not listening to technical reasons. Which is fine, I guess; they can now use eudev (which the only thing it does so far to is remove features from plain udev), and with luck their systems will "work", in the sense that the breakage doesn't show in their particular setups, like it happened with plain udev for many. If the breakage shows, I'm pretty sure the eudev fork will be of not help at all, because the problem has never been in udev (nor systemd, for that matter). So, no, I'm not trying to answer if you could "create a "/usr" service and make things dependent on /usr come after it's been mounted". I passed almost this entire thread because it's mostly people still hitting the same dead horse; really, if someone believe the eudev fork is the answer, they should go forth and use it. If there are people who don't want to believe developers like Greg Kroah-Hartman or Diego Pettenò when they basically say that eudev is a joke, why they would believe *me*? And, by the way, Diego doesn't like systemd *at all*. I just replied to Wolfe question because it was really concrete, and I thought I could answer it. I'm not going to hit a dead horse. Regards. -- Canek Peláez Valdés Posgrado en Ciencia e Ingeniería de la Computación Universidad Nacional Autónoma de México
