On 02/10/2014 04:19 PM, Armin K. wrote: > On 02/10/2014 01:27 PM, Gregory H. Nietsky wrote: >> >> On 02/10/14 14:04, Fernando de Oliveira wrote: >>> I thought it would be worth sharing what I have just read. Perhaps not >>> everybody knows about it yet. >>> >>> 1. Debian votes for systemd >>> >>> https://lists.debian.org/debian-ctte/2014/02/msg00338.html >>> >> as much as i think systemd is a stinking pile ... for debian [read >> ubuntu if you want by extension ignore they use upstart] it makes sense >> for the non technical consumer. >> >> in my corner of the world ubuntu is rolled out at one of the major banks >> [1 of 4] and they basically use firefox for everything maybe some libre >> office ... the same bank uses asterisk voip server. and for these >> systems its probably the way to go. same applys to daughters school 100% >> ubuntu. ubuntu is a firm favorite here and at one stage i was neighbours >> with M. Shuttleworths brother Grant other distros rarely feature. >> >> for use on a server or on embeded / .... systems its a bad call and i >> worry that it could cause problems that will be blamed on "linux" and >> not pid 0 .... >> > > Actually, people rather prefer systemd for embedded since as you know > embedded hardware isn't that much powerful as the server or desktop > machines (any architecture). Systemd uses pure C code instead of shell > scripts for most of the tasks, so it's a win in preformance and > memory/cpu usage as well for them and it *really matters*. You can > disable lots of things and optimize it for low-end hardware. I believe > it was some car company that used systemd in their embedded software but > I might be wrong. > > As for servers, I personally find it way easier to use and maintain > servers that come with systemd unlike the ones that come with > sysVinit/upstart/whatever. > > If you watched the video I posted few days ago, Lennart did mention that > there's a learning curve and if you got used to sysvinit you *need* to > learn systemd commands and such. Of course, those who spend > years/decades using shell will say that shell is easy, blah blah, etc, > but for beginners (<- note: beginners, newbies, no knowledge or very > limited one about shell) systemd is rather waaay easy to use and to > understand. > > Learning curve is there, but if you are at the beginnings, it's rather > way easier to learn systemd instead of shell. Of course, for some tasks > you'll still need shell, but mostly there's software for common things > that are being done on servers. Do note that not every server > needs/requires some special treatment, but there are actually lot of > them that do. > > If you think systemd is bad choice for servers, think again. Red Hat > Enterprise Linux, who's more focused on servers than on desktops is > shipping with systemd for RHEL 7.0 and they wouldn't do it if it was > *that bad* as people say it is. openSUSE is also using it, so it > shouldn't be surprising if SLES and SLED begin to use it, too. > >> for those who are not sure or would like to try something to see what >> will happen if systemd failed in any way type "kill -9 0" as root. >> >> Greg >> > > Really? Have you actually tried it for *any* init system? The PID 0 > seems to be protected from sigkill and sigterm from userspace and you > can't kill it that way. Any other non-standard way to terminate the > process would cause kernel panic anyways, be it systemd, sysvinit, > openrc, upstart, etc. >
It should read PID 1 instead of PID 0 in the original mail and response. My bad for not spotting it earlier, I just woke up when I wrote this response. The rest still stands. -- Note: My last name is not Krejzi. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
