Shared publicly  - 
 
Really, the goal was just 200, and all you got was 128? Jeez, I am not sure what disappoints me more, the impudently low goal, or that they didn't even reach that! Did I lose my controversy mojo?
45
8
Michaël PETRE's profile photoThilo Fromm's profile photoTobias Dreyer's profile photoOlivier Le Thanh Duong's profile photo
50 comments
 
Wow, there was an Anti-Lennart-petition? Honestly? Sure, you're doing controversial things. Same goes for Apple: they built a telephone, maybe someone heard of it? ;-) Change is good. Keep doing these 'crazy' things and follow your awesome ideas.
 
Oh boy this post just made my day! Hilarious :)

I think I have to sign the petition so that Lennart gets his mojo back...
 
That looks so sad. I almost want to participate so that it can meet the 200 goal. It's the jolly season after all.
 
Aw I should have signed it then, I was curious to see what he'd do if he reached 200. 
 
You are lame.. that is all. You will also buy me a drink when you next see me immediately. That should raise your controversy ratio.
 
haha awesome. I don't have any opinion on the systems you're involved in, but damn, you must be doing something right to have such an anti-following :) good on ya!
 
A visionary will always get beat up. I'm not completely happy with systemd, but that's because it is just different. Eventually I'll figure it out and it will be the new "norm".
You should have seen this guy when he was battling with getting pulseaudio. hehe.

But what you don't know is what a great photographer he is. It's just too bad that I lost all my pics from 2006 GUADEC.. I got some great ones of Lennart. :D
 
+Dhaval Giani why he's a very humble man.. ;) I'm just hoping he'll buy me beer later.. Kissing his white ass has to account for something.
 
What happened, +Lennart Poettering, I seem to miss something? I don't quite get why these guys are so disturbed. Is there a flame war / mailing list thread or something of some kind to see how this situation started? Google doesn't help much.
+Ingo Molnar span a nice thread yesterday about how integrating new information which conflicts with your inner view of the world makes you feel uncomfortable (https://plus.google.com/u/0/109922199462633401279/posts/cr7Fu8JeHYn). I guess the petitioners are affected.
+Sriram Ramkrishna - he's less of a visionary, more of a good engineer, I guess. All the core ideas have been around before. Instead of coming up with even more cool ideas, however, Lennart seems to focus on integration, making things available to a large user base. But maybe you're just working hard for that one beer :)
 
I fully agree with some of things in the petition. In my 'opinion' - Pulseaudio took fore-ever to get the bugs worked out. I don't "like" binary logging ( from what I have read about Lennart's goals they are far too windowish - best to force devs to "correctly" use the logging that is available rather than re-writing re-conceptualizing logging ).
Can't figure out why Lennart didn't contibute and enhance upstart rather than creating his own, which basically sends different distros based on Linux down different paths. ( I love fragmentation! - sarcasm )

So If you are just a Lennart cheerleader, which from the post above there seems to be many. This is just my point of view and I have neither the inclination nor the time to argue with you.

+Lennart Poettering I suppose the ppl that follow you on G+ must do so for a reason. Posting this will go to ppl that follow you mostly cause they like your projects ( I hope ) so the feedback you get will be tilted and biased. For a independent take on your projects scan the nets... [ Not a hater - just my opinion ]
 
Robert, I too agree that binary logging is a really bad choice, and that pulseaudio took forever to get working, but at least the latter was in many ways not Lennart's fault per se, but rather distros not using it correctly. I have to say though, that I too hope Lennart is at least reading and taking some of the critics the right way. All choices aren't exactly perfect all the time, and taking in feedback from other persons isn't gonna hurt.

But seriously, Lennart killing Linux? Did people forget the issues with, say, oss+alsa+esd? Has this already been forgotten? Wow.

I think crazy south american food killing Lennart is a bigger threat than Lennart killing Linux.
 
I wish, one day, there will be a petition against my work. This will be the landmark that I changed the world a little bit ;-) Next step: having your Wikipedia page vandalized. That's surely one of the coolest thing in the world !
 
I am disappointed too, you should break more stuff in 2012 (after, it will be too late)
 
I love how the "reasons" people are giving as just plain wrong in some cases "systemd breaks separate /usr".... I've just done an install of Mageia with latest systemd + /usr on LVM+ext4... works fine (after fixing some bugs in Dracut). And the whole "binary logging" issues just seems vastly over-played and seems to be people basing opinions on the /. headline rather than any real technical appraisal. Same for the suggestions of "improving upstart"... if any of those commenting had actually read the blogs, they'd realise that this was attempted but it didn't work out.

Well I guess if people want to show themselves as ignorant in public, that's funnier for the rest of us. If they really want to write a technical critique of Lennart's work (and this should always be welcomed - especially by Lennart!) then a silly "petition" is certainly not the way to do it... such approaches are simply childish!
 
I am not convinced about logging as well. I prefer text formats because I can see them w/out specialised programs (early in +GNOME 2.9x I get crash after changing settings and because dconf-editor crashed as well I was unable to alter it). That said I as not sysadmin and I'm using desktop.

+Gustaf Räntilä: As of upstart - the problems with it architecture are discussed on systemd page. It is based IIRC on evenets which can fire and that needs the writer of init script to hardcode the parallelism.On the other hand systemd (by heavy use of Linux-specific features) allows to have to have much less handcoded dependencies relaying on blocking system calls.

The result is that wile Ubuntu-based distros are using upstart the systemd have extended beyond Fedora-based (for example Gentoo and Arch can have systemd as alternative init).
 
+Robert Edouard, +Gustaf Räntilä I do see a lot more advantages than disadvantages in "binary" logging. We're into embedded systems and we pretty much hit the wall with (r)syslog/(ng) when it comes to orchestration of many IP-enabled devices. So we're currently switching to a Couch (http://couchdb.apache.org/) to store syslog lines; now harvesting/synchronizing log lines in a cloud of devices suddenly is not a pain but actually fun. Not to mention searching, filtering, and all the other cool things we can do now without much hassle. A system like Journal with a wider support and a broader user base than our home made approach is definitely welcome here.
By the way, I'm a huge fan of upstart personally, use it on all my machines. +Scott James Remnant is doing a hell of a job there. However I find systemd more suitable for our embedded devices, so we use it there. Having several approaches for the same goal is a good thing, I think.
 
+Thilo Fromm , ahem, dude, Scott is not doing Upstart anymore. I think he's mostly over it. And while Upstart might not be entirely dead yet, it's glacial in its development without him. (and yeah, I am too lazy to respond to the technical nonsense people posted here)
 
+Maciej Piechotka: I never said anything about upstart vs systemd, I'm all for the latter. I just pointed out that all decisions aren't perfect. Binary logging do have benefits for sure, and like +Thilo Fromm said, one can have huge benefits with e.g. a couch based syslog store. It's funny, I have agitated for such a very solution where I'm working now.

I'm just saying for some (perhaps most desktop users), binary logging isn't ideal or preferred. I too have had lots of issues with systems where I need special tools to see what's wrong (debug logs, configurations etc) and where these tools themselves are buggy. Perhaps the log system could have several stores, one for desktop (text), one for embedded (binary), and one for enterprise server situations (e.g. couch).

In any case, I'm happy Lennart is working on systemd. And yes the petition per se is wonderfully pathetic.
 
+Gustaf Räntilä : First of all - I agree that +Lennart Poettering is free to do anything he (alternat alternatively his employer) wants. Sorry about mentioning upstart - I was sure someone mentioned frag fragmentation mentizing of init subsystem when we have upstart (the counterargument is that +systemd can be view as improvement over older attempt).

The mention of logging was rather that I don't like the binary formats when it is possible to avoid them (i.e. on 'normal' systems - I'm not talking about embedded software, messages in HPC etc.). I would prefer the prim primarily text format (possibly with more rigid format and indexing). That said a) +Lennart Poettering can prove me wrong simply by implementing software b) noone forces me to actually use his software and others are free to do it (I would stay with +systemd and +PulseAudio) c) I may misunderstood the concept.
 
Gustaf: Binary isn't ideal for most desktop logging? Really? Why should I care about how it is stored? It needs to be reliable, quick, easy etc. Don't see what the storage format has to do with that. Also, you can still have syslog running along with journal if you care to.

Find the whole stop energy petition funny.
 
+Maciej Piechotka As I understand it, you are missing the possibility to use 'cat' for files created by Journal. All the other benefits of text files, in essence their portability, are also given when using Journals binary log files, if I understand the project goals correctly. So if Journal comes with a tool that dumps log entries to stdout and if this tool isn't depending on a running Journal demon itself, everything will be fine. Or am I missing something?

+Lennart Poettering I recon the key issue for the missed poll is that AFAIK Systemd and Journal are not available on Gentoo yet.
 
Should we start a petition in order to promote a better petition against you? How many people would it take? ;-)
 
I hope the dude who created the petition appreciates that I made it popular... ;-)
 
You should start writing a new online petitioning system. That would piss them off.
 
+Bertrand Rousseau nah, I'd prefer if people would organize one of those pledge drives or so, rather than a petition. I mean, everybody has a price, so let's say mine is five million dollars or so. If they collect that much money, and pass it to me, I promise I'll consider leaving poor sysvinit alone! ;-)

Those naives believe they'd get there wishes for free. They never watched a James Bond, where everybody knows for the villains it's just about the ransom money...
 
+Lennart Poettering ahhh so this is the amazing money making scheme you alluded to in Dublin.... I've been waiting to see what it is for ages. Work on a project and annoy enough people until they pay you to leave it alone! Genius!
 
See, Lennart Poettering is just an assumed name, I was actually born under the name of MAX ZORIN! Muahahahahah! ;-)
 
+Mathias Rüdiger "So if Journal comes with a tool that dumps log entries to stdout and if this tool isn't depending on a running Journal demon itself, everything will be fine."

Ten lines of perl, I'd bet.
 
+Thilo Fromm the default tool to view the journal data by default spits out a pixel-perfect copy of the classic /var/log/messages files (well, except that it is much cooler, since it gets timezones right and stuff), and it is much shorter to type, and can interleave files (including rotated ones and does that implicitly) and knows access controls and a gazillion more things. i.e. "tail -f /var/log/messages" becomes "journalctl -f". Shorter, correcter, better, more aweosome, but processable with the exact same scripts, because i carefully made sure to generate the same pixel-perfect output. Oh, and if you pass "-o json" you get the data in json, and if you pass "-o export" you get the full data in a serialized binary way, that is trivial to parse for other tools who are into it...
 
+Olav Vitters: Why should you care how it's stored? Well, it depends.

What I meant when I wrote that binary might not be ideal for desktop are two things;
1) When shit happens (and shit does happen), you might want to boot up with a livecd, and the only one you have lying around is an old knoppix from 2010. You won't be able to read your logs.
2) Most desktops are fast enough to write text files, and text utilities are fast enough to parse them.

If shit will never happen to you, then you're right, you won't need to bother how logs are stored. For the rest of us, it's not something to rule out too easily. But again, in the long run, when these tools are defacto standard, I see no issue with binary logs. It's the transition period I'm a little afraid about.
 
Gustaf: So your usecase for most desktop users is an outdated live cd? Further, it is important to use text files, because with enough work, you're able to read them?

You said most desktop users. I find these design decision pretty crazy and arbitrary. As I said, you can still log to text files using syslog for a year or so if the distribution (e.g. Debian) cares.

For me (not speaking about desktop), I'd rather hope that Journal makes some things possible that atm for me are quite annoying. E.g. when a program crashes, reliabily add the stderr messages in the bugreport. More easily and quickly filter the logs.
 
That's a good business model for free software. Make a patch in PulseAudio that randomly play a Justin Bieber song and ask money to remove it. It should be compatible with the GPL, isn't it?
 
Most of those guys on slashdot are just a bunch of dead-enders who don't know what they are talking about. They are just giving opinions without actually trying any of it or looking deep to it. If they are systems administrators they aren't worth jack if they can't open their mind to new ideas and new ways of doing things.

That said, I can't wait for systemd to show up at work.. haha, that's going to make people grumble. (fvwm2 is still the major window manager here, although I hear there is a lot of KDE.. but with a home dir of only 128M, it's not going to be that modern) But I'm working on it!
 
Having trustworthy logs trumps the binary issue. Sysog isn't going away. Tools can convert the binary into text. Knoppix from 2007 should still be able to run a compatible binary from a USB key, assuming you wanted to use the trustworthy binary logs and not the potentially-modified text from syslog. Why is this a debate? The "cons" are so easily dealt with, and the pros are enormous.
 
You should invest more in your sales/marketing department. e.g. writing usable documentation with real life "happy family holidays through systemd"-scenarios ... Honestly, as a user familiar with upstart/sysvinit, having to deal with systemd is a big pain. I can't see the beauty of your solution right now.
 
+Yuri Braz Please refrain from speaking on my behalf. I don't care about sysvinit; I certainly don't love it.
 
+Yuri Braz So you know what the majority of Linux users and administrators think? Awesome achievement!! Are you going to FOSDEM and can I touch you? If no, can I stand near to you and take a sneak picture? Stop making me laugh :P
 
Ah, the slashdot crowd found this thread! +Roland Moriz, lemme guess you never ever had any look at systemd or its documentation, have you? Because if you had, you'd know that systemd is much better documented than most projects (probably even one of the best documented projects in free software, especially considering how new it is), and you may criticise it for whatever you want, but if you criticise it regarding its documentation you just make it obvious you never had a look. I mean, it's completely fine if folks make the educated choice not to like systemd, but what you are doing is just uneducated FUDing based on assumptions, guesses, and hearsay. I am not sure why you feel the need to contribute that to the discussion, it is just annoying noise and not useful at all. And +Yuri Braz, listen, I go to a multitude of conferences each year, and talk to people and like to believe to have a good idea what is needed and what people want. I don't think you ever did that (because if you did, I would have met you at one of those occasions, and I didn't and don't know who you are). Now, even though I like to believe that I did my homework I'd still be foolhardy to claim I really knew what the majority of Linux users and administrators believes. But hey, one thing I do know is that you know it even less than I do because I spend considerable time trying to figure it out and you don't, and I hence humbly ask you to stop spreading your nonsense. And anyway, what you and your friends who put together the petition completely misunderstood is how projects get their relevance in Free Software. Because it isn't me who mandates that everybody does what I think is right, it is the distributions who sometimes choose to adopt my ideas, and sometimes don't. I just write code, and propose it. It is technical committees such as Fedora's FESCO who then decide to adopt it -- or not, and I have never been a member of any such committee. If you want to ensure that my ideas don't succeed then you should influence those committees, where the decisions are made, and not me, who just proposes them. Petition the distributions, don't petition me. Finally, if I write something and dump it on the Internet and some random distribution adopts it, then that's entirely and exclusively their decision, and believe me, with the exception of Fedora I personally have not actively proposed systemd or PA or Avahi in any way to any other distribution. Many distros apparently liked what they saw and took it, but you can't blame me for that (I can't control what they do with my code, and that's a core idea of Free Software!). Also note that there are distributions who actively decided not to adopt any of these projects, for example Slackware. I invite you to participate in their community. They are also GNOME-free, which probably makes them particularly interesting to all those folks who detest systemd, GNOME and all the other newer technology coming from the pro-progress camp of the Free Software world. You know, I obviously don't share your refusenik positions, but I am fine if people think like this, but what I don't get is why they believe that they can reverse the development by petitioning me, instead of the ones who actually make the decisions. And now, please leave my Google+ account alone. Please don't post here anymore, because what I write here is not intended for refuseniks. Thank you very much!
 
+Lennart Poettering F*ck those people. I think the idea (of both systemd and journald) is great.

I don't care wether the tool to read my binary logs is tail, less or journalctl. After all there's no such thing as plaintext.

Stuff that needs to die (in it's current form) is SysV, cron and syslog. Looking forward to when the Debian people see that the default should be changed...(and yes I'm too lazy now to argue why I have this opinion)
 
+Lennart Poettering And what to do if your userspace is dead and the only thing u can do is to access file system somehow? How to read a journal in this case?
 
Hee hee - was looking into systemd (it being upcoming in rhel 7 presumably), and people were bitching about "who wrote it". Dug deeper, it was "that pulseaudio guy". 

Well hell - pulseaudio was the first time I actually got WoW and my web browser to not fight over sound and work at the same time. Thank you for that, by the way. Most of the issues I've had with it were mostly people fighting against adopting it (looking at wine here) as opposed to an actual software issue. 

So now I've got to dig into and use it, just to piss off the whiners. Score one for systemd. :-)

PS. Binary logs make me sad, but that's mainly because my old work had some really poorly implemented binary logs. The level of sin here is roughly the same as "impure thoughts". So long as we're moving forward, on the whole I'm for it. 
Add a comment...