You mean a disconnect like the fact that you can pretty trivially 100% a couple of servers running feature extracting daemons processing text based logs at the moment for a small cluster of machines?
Where has this absurd notion that text logs are efficient come from? 16 redundant bytes for a 4 byte IP address? Another 5 for a 2-byte port number? Text based logs generate a huge amount of redundant network traffic. journal-gatewayd might not be the solution people are looking for, but it absolutely makes sense to separate log collection on the local machine and the network interface in some way, since no sane syslog configuration just blasts UDP packets into the ether and hopes they arrive, and network programming is an *entirely* different set of concerns to local programming.
Which on some level, makes HTTP GET just as valid as anything else, since the important parameter is "did the server get the data" not "have I sent it?". rsyslog in TCP mode is how you achieve this currently. But you're left with a lot of text data volume to handle for something like access logs. Of course rsyslog also reads systemd journals natively.
You know I'd been wondering what his position was (was about to look it up while reading this thread) because there seemed to be an awful lot of people jumping to the conclusion that obviously this was systemd related and "look systemd is driving away developers!".
Wow, another long post on systemd spouting ideology and rhetoric without nary a valid technical complaint (nor accurate statement of fact: systemd does not replace 69 services. It co-exists with any number of other utilities providing those services, but the project also aims to provide lightweight implementations of what it considers "core" functionality. None of my systemd systems use systemd cron or DHCP for example, even though those daemons exist.
So I'm going to sabotage the official docs--and thereby lose my job--just so I can score cheap points on Wikipedia? Yeah, that makes lots of sense.
Words have meanings, and their construction has important ramifications for how your comments may be interpreted?
The point is that you're the original source of truth about a thing. That makes you the primary source. Wikipedia and all encyclopedia's don't deal in primary sources. For an obvious example as to why, consider the various reasons people are discouraged from authoring wikipedia entries about themselves. The general rate of defacement that comes from Congressional IPs is a good example.
No we've been going rapidly downhill since the exact moment the current government got elected. That is the sum total of how far this downhill extends: since the current Liberals got power, and proceeded to try and ram through every bit of right-wing nuttery they possibly could.
But I thought system logs never got corrupted! You do realize that system logs aren't flushed to disk line by line in most cases right? That would be a huge amount of incredibly slow disk IO and seeks. That "half line" can be the head of dozens of entries. Which again...when you also have to factor in the speed considerations for the relative size of each the messages...
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Of course, if you and literally every other Anon. Coward in this thread of posts knew what they were talking about, then it wouldn't exist.
The "anti-systemd" brigade seems to consist of a lot of people who have absolutely no idea what they're doing, let alone any idea of how Linux actually works.
To which I asked what version you ran, and you claimed "current production" which is Deb 7, and the installed packages demonstrate that you are warong.
In fact you claimed that I don't know what I'm talking about even though I proved you wrong numerous times. Continuing a lie will NOT make it the truth! A simple "Yup, I meant to claim that in beta/dev Debian it uses systemd, not in a production stable release" would have resolved the issue. Instead of doing this, you keep repeating a false claim that Debian 7 uses systemd as it's default init system.
The existence of a package page does not make it a package installed by default. Looking at the default packages list (or what's installed) determines the default. You lose, good day!
Yeah no I didn't. I said "Debian".
So in addition to having no clue, you also apparently can't read. Which well, I guess one does explain the other.
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Really now? What Debian are you running with systemd and journald? We run thousands of machines on Debian 5-7 and I have yet to use either, see either, or configure either. I have init, and I have a choice of rsyslog or syslog-ng, that's it.
Reading this whole post, I see a whole lot of people fabricating information trying to claim that anyone not for systemd is some type of [ad hominem], I see lots of appeal to authority arguments for systemd, I don't see much honesty when it comes to systemd which is very troublesome.
systemd can launch shell scripts and you can turn off all the other features involved there. So yes, it has been handled - the fallback is you write a shell script, like you already do, that does this manually.
Though again I ask: what reports? Where? I am specifically curious of cases where systemd breaks, because I am running it right now and trying to evaluate it for using it on other systems in the future. Vague concerns do not help that.
Yes because I'm sure the literally microseconds of lag (I mean maybe, there's a lot of assumptions in that statement) is a problem compared to, I don't know, the 100s of milliseconds it takes to flush anything to disk. Presuming it gets there, because if the system goes down, it's unlikely anything gets written out of any memory buffer.
Then of course there's the byte overhead of all those characters rather then concise binary files, so there's a bandwidth cost too.
But it's pretty obvious you don't actually know what you're talking about and are just spreading FUD.
so configure journald to simultaneously spit out text log files to syslog/rsyslog
This works, but it should be the default.
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Of course, if you and literally every other Anon. Coward in this thread of posts knew what they were talking about, then it wouldn't exist.
The "anti-systemd" brigade seems to consist of a lot of people who have absolutely no idea what they're doing, let alone any idea of how Linux actually works.
If it wasn't booting, then there is some sort of error message. Or no error message just a hang maybe! But no, that's never what anyone feels is "worth" mentioning. Just "it broke". I'm sure in debugging init script problems they would've supplied exactly the same information. Or you know, not, because it's extremely unlikely their system was locking up completely.
It also has the great property of letting you install partial configuration overrides per-file by dropping a file in a subdirectory/etc/systemd named after the file you're editing.
Solving the problem once, in the init system, is smarter then having to solve it every single time in every single init script (or you know, it may do anything from not work to stall the entire boot).
And "journalctl" works just fine for systemd, what's your point?
Plus bonus: you can easily tail all the logs on the system with "journalctl -f" and you can of course pipe that output into grep or whatever else you want.
So don't rely on people to accurately report symptoms. Take a sample of their blood, regardless of how well they claim to have been feeling. You can tell from passport records whether they've travelled to or from one of the highly affected countries, so you don't even need them to honestly report where they've been recently, and if they've been there in the past three weeks, require a mandatory blood test. I've heard that a fairly recently developed test for this can be performed in under 15 minutes, requiring only a couple of drops of blood that can be obtained simply by pricking the person's finger.
Blood tests don't detect Ebola reliably before symptoms manifest.
Also because there's little to study, it's hard to test a vaccine, and its also downright dangerous to study. Mundane pathogens that nonetheless kill a lot of people have the delightful property of being almost completely survivable if you contract them accidentally in a research setting.
It almost certainly does only survive for minutes on hard surfaces. The surfaces it survives on longer are those which are literally covered in blood and bodily fluids for it to reside in.
I don't know, he's the president, it's kind of his job to decide what to do as commander of the military. So far he's just been doing stuff, hasn't really accomplished much.
This might be a crazy idea, but have you considered that bombing things, in fact, violence of any kind! might not be the solution to all of America's concerns in the Middle East. I mean has it really worked out for you lately?
Slashdot Mirror
Yeah, we shouldn't have released X11 until it was 100% bug free either...
You mean a disconnect like the fact that you can pretty trivially 100% a couple of servers running feature extracting daemons processing text based logs at the moment for a small cluster of machines?
Where has this absurd notion that text logs are efficient come from? 16 redundant bytes for a 4 byte IP address? Another 5 for a 2-byte port number? Text based logs generate a huge amount of redundant network traffic. journal-gatewayd might not be the solution people are looking for, but it absolutely makes sense to separate log collection on the local machine and the network interface in some way, since no sane syslog configuration just blasts UDP packets into the ether and hopes they arrive, and network programming is an *entirely* different set of concerns to local programming.
Which on some level, makes HTTP GET just as valid as anything else, since the important parameter is "did the server get the data" not "have I sent it?". rsyslog in TCP mode is how you achieve this currently. But you're left with a lot of text data volume to handle for something like access logs. Of course rsyslog also reads systemd journals natively.
You know I'd been wondering what his position was (was about to look it up while reading this thread) because there seemed to be an awful lot of people jumping to the conclusion that obviously this was systemd related and "look systemd is driving away developers!".
Wow, another long post on systemd spouting ideology and rhetoric without nary a valid technical complaint (nor accurate statement of fact: systemd does not replace 69 services. It co-exists with any number of other utilities providing those services, but the project also aims to provide lightweight implementations of what it considers "core" functionality. None of my systemd systems use systemd cron or DHCP for example, even though those daemons exist.
So I'm going to sabotage the official docs--and thereby lose my job--just so I can score cheap points on Wikipedia? Yeah, that makes lots of sense.
Words have meanings, and their construction has important ramifications for how your comments may be interpreted?
The point is that you're the original source of truth about a thing. That makes you the primary source. Wikipedia and all encyclopedia's don't deal in primary sources. For an obvious example as to why, consider the various reasons people are discouraged from authoring wikipedia entries about themselves. The general rate of defacement that comes from Congressional IPs is a good example.
Or you know, just don't insist on being the contributor to an article about things you can't be a contributor to?
It's pretty obvious why an encyclopedia wants to work that way.
No we've been going rapidly downhill since the exact moment the current government got elected. That is the sum total of how far this downhill extends: since the current Liberals got power, and proceeded to try and ram through every bit of right-wing nuttery they possibly could.
Which is also still a lot less then all the other completely preventable deaths in every other industry.
Keep on backpedalling.
But I thought system logs never got corrupted! You do realize that system logs aren't flushed to disk line by line in most cases right? That would be a huge amount of incredibly slow disk IO and seeks. That "half line" can be the head of dozens of entries. Which again...when you also have to factor in the speed considerations for the relative size of each the messages...
YOU DID!
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Of course, if you and literally every other Anon. Coward in this thread of posts knew what they were talking about, then it wouldn't exist.
The "anti-systemd" brigade seems to consist of a lot of people who have absolutely no idea what they're doing, let alone any idea of how Linux actually works.
To which I asked what version you ran, and you claimed "current production" which is Deb 7, and the installed packages demonstrate that you are warong.
In fact you claimed that I don't know what I'm talking about even though I proved you wrong numerous times. Continuing a lie will NOT make it the truth! A simple "Yup, I meant to claim that in beta/dev Debian it uses systemd, not in a production stable release" would have resolved the issue. Instead of doing this, you keep repeating a false claim that Debian 7 uses systemd as it's default init system.
The existence of a package page does not make it a package installed by default. Looking at the default packages list (or what's installed) determines the default. You lose, good day!
Yeah no I didn't. I said "Debian".
So in addition to having no clue, you also apparently can't read. Which well, I guess one does explain the other.
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Really now? What Debian are you running with systemd and journald? We run thousands of machines on Debian 5-7 and I have yet to use either, see either, or configure either. I have init, and I have a choice of rsyslog or syslog-ng, that's it.
Reading this whole post, I see a whole lot of people fabricating information trying to claim that anyone not for systemd is some type of [ad hominem], I see lots of appeal to authority arguments for systemd, I don't see much honesty when it comes to systemd which is very troublesome.
Fabricating info? Then I guess this doesn't exist?
There are Debian packages for systemd-init. There are Ubuntu packages.
So we return to my original opinion...
systemd can launch shell scripts and you can turn off all the other features involved there. So yes, it has been handled - the fallback is you write a shell script, like you already do, that does this manually.
Though again I ask: what reports? Where? I am specifically curious of cases where systemd breaks, because I am running it right now and trying to evaluate it for using it on other systems in the future. Vague concerns do not help that.
Yes because I'm sure the literally microseconds of lag (I mean maybe, there's a lot of assumptions in that statement) is a problem compared to, I don't know, the 100s of milliseconds it takes to flush anything to disk. Presuming it gets there, because if the system goes down, it's unlikely anything gets written out of any memory buffer.
Then of course there's the byte overhead of all those characters rather then concise binary files, so there's a bandwidth cost too.
But it's pretty obvious you don't actually know what you're talking about and are just spreading FUD.
so configure journald to simultaneously spit out text log files to syslog/rsyslog
This works, but it should be the default.
Well good news, this is the default, at least on Debian. In fact Debian doesn't even store journalctl logs, it fowards then straight through to rsyslog.
Of course, if you and literally every other Anon. Coward in this thread of posts knew what they were talking about, then it wouldn't exist.
The "anti-systemd" brigade seems to consist of a lot of people who have absolutely no idea what they're doing, let alone any idea of how Linux actually works.
If it wasn't booting, then there is some sort of error message. Or no error message just a hang maybe! But no, that's never what anyone feels is "worth" mentioning. Just "it broke". I'm sure in debugging init script problems they would've supplied exactly the same information. Or you know, not, because it's extremely unlikely their system was locking up completely.
It also has the great property of letting you install partial configuration overrides per-file by dropping a file in a subdirectory /etc/systemd named after the file you're editing.
Solving the problem once, in the init system, is smarter then having to solve it every single time in every single init script (or you know, it may do anything from not work to stall the entire boot).
And "journalctl" works just fine for systemd, what's your point?
Plus bonus: you can easily tail all the logs on the system with "journalctl -f" and you can of course pipe that output into grep or whatever else you want.
So again, what's your point?
Indeed. People keep saying this and never what they were.
So don't rely on people to accurately report symptoms. Take a sample of their blood, regardless of how well they claim to have been feeling. You can tell from passport records whether they've travelled to or from one of the highly affected countries, so you don't even need them to honestly report where they've been recently, and if they've been there in the past three weeks, require a mandatory blood test. I've heard that a fairly recently developed test for this can be performed in under 15 minutes, requiring only a couple of drops of blood that can be obtained simply by pricking the person's finger.
Blood tests don't detect Ebola reliably before symptoms manifest.
Also because there's little to study, it's hard to test a vaccine, and its also downright dangerous to study. Mundane pathogens that nonetheless kill a lot of people have the delightful property of being almost completely survivable if you contract them accidentally in a research setting.
It almost certainly does only survive for minutes on hard surfaces. The surfaces it survives on longer are those which are literally covered in blood and bodily fluids for it to reside in.
What plan then -
I don't know, he's the president, it's kind of his job to decide what to do as commander of the military. So far he's just been doing stuff, hasn't really accomplished much.
This might be a crazy idea, but have you considered that bombing things, in fact, violence of any kind! might not be the solution to all of America's concerns in the Middle East. I mean has it really worked out for you lately?