Slashdot is powered by your submissions, so send in your scoop


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:Bullet point not real feature (Score 1) 508

attacking? So far after dozens of posts back and forth it's only you that so far have cast insults, name calling and attacks. Perhaps you should take a while and contemplate that for a while and wonder why you constantly accuse me of the things that only you do? Questioning your knowledge on what parallel is is not "attacking the person" since you clearly have your own definition of parallel that is shared by nobody else. Had I written it in your style like "oh the usual anti systemd troll fanboi that don't even know basic computing terms like parallel, what are you, like 14?!" then you would have had a point but as it is now you don't.

Parallel have nothing to do with single points of failure. Parallel simply means that you start things at the same time. Now since services started by an init will have dependencies, not all services can be started at the same time and therefore those that have a hard dependency on service X simply have to wait for service X to finish first.

This is not new, this has been the rule for init since back even before sysv and is not some new strange definition of parallel. What systemd brought to the table (and which as I understand it was first implemented in launchd) over sysv was that some services could be started in parallel even when there where hard dependencies due to the socket activation, this however is only enabled on a service by service level since it cannot be just turned on for all services (not all will work with that).

Comment Re:Bullet point not real feature (Score 1) 508

Well then you don't know what parallel is, naturally an init cannot start services with dependencies on services not yet started, parallel or not. This is true regardless of init system used. So if that would prevent an init system to be put into production then no init system would ever be used (as I've said I have had for example sysv do just that for me on several occasions.

So now every one who does not agree with you that systemd is a sign of the end of days is a fanboy, o boy... Anyway you are free to use the old dinosaur system that I don't hate (I just think that systemd is better than sysv in so many ways, but that is not hate), different from you though I will not call you a fanboy or make other insult just because you happen to like some other system than me. Btw thanks for calling me kid, made me feel young again :)

Comment Re:Bullet point not real feature (Score 1) 508

Please point out how I tried to push you to solve your problem to go to RHEL7? Since I don't even run RHEL7 myself (I'm a Debian/Ubuntu user) that would be very interesting to see indeed!

And I'm not trying to send you down a fishing expedition into systemd innards, I'm actually honestly interested in why your problem happens the way it does because due to the design of systemd and my own experience with it this should not happen so fully understanding what happened was something that I found interesting. And if there where a real bug in systemd then that could also have been fixed (I'm a developer, not a systemd developer but fixing code I can do in any project).

Comment Re: Why trust one or two people? (Score 1) 83

Well fuck the people with no patience :-). If we are talking about the US president, the vote takes place on November 8 then the electoral votes will not be case until December 19 and then on January 20 the newly elected President is inaugurated. So why does the no patience people need to know the result two months prior to it actually takes effect?

Comment Re:Bullet point not real feature (Score 1) 508

Where on earth have I implicated that you (or anyone else for that matter) are negligible by not changing to systemd?

Yes I write init scripts, but on systemd machines I #1 don't since I write unit files instead which have no commands what so ever and #2 I have no need to write any systemctl or journalctl command ever in an init script.

WTF is the "systemD's mouse service"? Could you please provide more details on what model of mouse dongle this is, what the service is that depends on it, if any of this is custom things or default things from CentOS7/Fedora. A bug like this must be fixed.

Comment Re: Why trust one or two people? (Score 1) 83

The only benefit of such a system would to get the result sooner than a full paper count and that benefit is quite moot considering that the effect of the vote won't take effect for a very long time anyway. You write that the computer verifies the paper but #1 there is no need to verify the paper considering how paper counts work and #2 wtf do you do if the two results differ anyway, all you know at that time is that there is a difference, not where it is.

Comment Re:Bullet point not real feature (Score 1) 508

No I don't have a "dog in the fight" or trying to push anything. If anything I could say that about you considering that you encountered some problem and then all of the sudden systemd is something that must be erased with fire. I have only ever said that it works for me and others and that if it does not work for you or you don't like it then fine do go and use something else.

The arguments might look verbose but #1 you can tab-complete, and #2 the .service/.target part is not needed, i.e you can type "systemctl start ntp" or "systemctl start ntp.service" they are equivalent. #3 yes the systemd-readahead-collect name is verbose as hell but see due to #1 all you have to do is type "systemctl enable systemd" and then press tab twice to get all the possible options.

Regarding your mouse dongle I don't know why that happens for you, I use such a dongle myself (from logitech) without any issues so it's definitely not something that breaks for all such dongles. That the init is halted is probably due to something having your mouse as a dependency of course then parallelism does not matter since it's a hard dependency for some service. Don't really know why that affects your init either, mine is enabled by the kernel and not by init so perhaps it's not the mouse per say but something that want to use it that halts?

Btw why do you spell systemd with a capital D? Having a small letter d suffix for daemons is a very long tradition within Unix so why mark it specifically with a capital letter? Do you do the same with httpd, named and so on? Just curios because I have seen so many anti-systemd people write it with a capital D like the d had some special meaning.

Comment Re:Bullet point not real feature (Score 1) 508

Well I have had the very same thing happening with sysv, for example it scheduling sshd to start before networking and thus the whole boot hangs forever since the sshd script never timed out and thus the network script could never work, all due to the non parallelism of sysv. I've also had numerous remote servers hang on shutdown forcing me to drive 60 metric miles to pull the plug manually.

So yeah it sucks that your machines experienced these problems but let us not fool ourselves into thinking that there where no problems with sysv either. And no there where no logs to tell me what the problems where with my own issues above either. Since switching to systemd I atleast have had no such problems what so ever since, just to let you know that your experience is not the only one.

And what very verbose commands? is "systemctl" so much more to type than "service" that it's now seen as verbose? And "journalctl" instead of "grep x /var/log/syslog"?

All in all I just don't share your experience, and with that I do not claim that you have your experience, far from it. I just don't share it.

Comment Re:Bullet point not real feature (Score 1) 508

Well sorry to hear that you have such a major problem. Does however sound that your problems is with your vendor mostly and not with systemd to be honest even though of course the change to systemd is what initiated the problem in the first case.

For me it has been the opposite of flaky (upstart and sysv was flaky for me) but then I have no closed vendors software to manage so I guess that I'm lucky that way. For our customers I'm that vendor and we support all the init systems out there (but the systemd one was the nicest to write and the nicest to run)

Comment Re:Errr (Score 1) 56

It's not careful, if I throw the window at the top it gets maximized, if i throw it to either left or the right side it gets sized to take up half the screen but if I move it there and move back a little I get the window to dock to either side instead of getting it to be maximized.

Comment Re:Poor tact but right reaction (Score 1) 294

mesa have been updated like hell during this last year so the version in most repos (I'm on Ubuntu and the latest mesa in their repo is 11.2 while mesa-git is 12.x and with 11.2 there where no support for the RX480 since all the new AMD open source drivers have been put into the 12.x branch. The AMDGPU-PRO which is their closed driver might work ok but I run their fully open stack. My daughter plays Life is Strange and that game requires mesa 12.x (it does not work with AMDGPU-PRO) which is why I went this route.

Slashdot Top Deals

"Don't discount flying pigs before you have good air defense." -- jvh@clinet.FI