« first day (1349 days earlier)      last day (3606 days later) » 
04:00 - 21:0021:00 - 00:00

4:52 AM
Helllo is any one on?
Hello?
Is there any one there?
Can anyone help me?
 
 
4 hours later…
8:52 AM
@Andrew two things:
 
-2
A: Why is Wayland better?

armornickIn addition to the above, the X Windowing Server was also not designed to be used like Ubuntu uses it. Like the name implies, it is mostly developed for server use, like remote desktop use or to link several thin clients to a single operating system. Essentially, Wayland will drop the server sup...

 
one, if you had actually said your question, then someone could have come back later and answered you. as it is, this is useless.
 
souldn't be answering questions if you don't know what you are talking about...
 
@Andrew ?
 
Example?
 
8:54 AM
@Andrew I have no idea what you're talking about anymore
maybe start over?
 
Wayland doesn't drop any server support. Wayland is the server.
 
Well that nebie doen't know what he is talking about
X is not a sever in the context that he is using it as
give it a downvote it only takes 10 to have it remove
 
That's true. In X terminology the client is the server and vice versa.
 
Yea
Hi B
 
@mikeserv this really confused me for a minute...
 
9:01 AM
Geez. You're better than I. It really confused me for years.
 
@mikeserv lol I meant your wording
 
But it's not my wording.
 
because actually what you meant was the X11 client runs on the server and vice versa, not the X11 client is the server
@mikeserv ?
 
I hope so. Sometimes it still confuses me..
Well, no, I think I meant closer to it is the server. But I'm thinking in multiple machine configurations.
 
I can't remember when I learned the X11 architecture
I don't think it took me that long but I'm not really sure
 
9:06 AM
I once had this pretty cool setup with containers.
Each had its own vt.
 
@mikeserv no let's say you have a hub-and-spoke network. the hub is the server and the spokes are thin clients.
 
All on the same desktop kind of.
Ok. I had a horse and buggy network.
 
in that case X11 clients run on the server, the X11 server runs on the spokes
 
Just kidding - you're right.
 
ah. merci.
 
9:09 AM
I used XDM to distribute it over sockets between the containers.
So they all got hardware accelerated.
It was pretty cool till I broke it and couldn't be bothered to find out how.
But switching x desks with CTRL+ALT+Fn was neat.
 
@mikeserv ooo, you fancy
 
Nah. I just couldn't afford more than one monitor.
Anyway, that's how I learned about the spokes.
 
haha nice
bloody hell Debian Stable is so freaking OLD
 
HAHAHA.
 
it's just so goddamn stable agh it drives me nuts
that sounded sarcastic but it actually wasn't it's bloody annoying
 
9:28 AM
@strugee It's good at what it does. Maybe your requirements are different and you should consider a distro that is optimized for something closer to your expectations.
 
and vim is better.
But emacs has a keyboard shortcut for that.
Did anybody see that question the other day - something like How does Ubuntu make Debian Unstable stable or something?
I laughed at the title.
It was funny on lots of levels.
Sorry guys. Just trying to stir the pot. Nothing like a good stew.
 
9:46 AM
@mikeserv Ya I don't think the OP had a clue how software development in general works, much less distro release processes.
@mikeserv Vim isn't better. It's best. And yes @strugee should use it as his distro.
 
yay. i love stew.
@Caleb - while I agree, it was also funny because it was a very good question.
Not that Debian Unstable isn't a stable enough platform, but then why isn't Ubuntu?
 
With the answer being somewhere between "they don't" and "it's really freaking complicated" ;
)
 
Somewhere.
A stable enough platform for a desktop os, I mean.
Could be inconvenient for a 24-7 lab or some other distributed deployment setting - which is where I think Debian shines.
Silly Shuttleworth. Anything happen with Mir yet?
 
10:13 AM
any vmware experts out here?
 
@BTRNaidu don't ask to ask
just do it
@Caleb I use Arch daily
 
ok.. here i go then
 
I'm on Debian Stable because I'm writing Puppet modules targeted for it
:P
 
I have a windows vm with 60 GB disk running on VMWare Workstation. I see the total size of the folder containing is close to 115 GB. I had some 7 snapshots and though snapshots are the reason. Delete few of them, performed defrag and compact but no use. The enter folder size is still same.
 
@Caleb I spit in the face of Vim users.
 
10:15 AM
I see a huge list of

Windows-s0xx.vmdk Windows-000007-s0xx.vmdk Windows-000005-s0xx.vmdk Windows-000004-s0xx.vmdk Windows-000003-s0xx.vmdk Windows-000002-s0xx.vmdk and Windows-000001-s0xx.vmdk

files.
 
@Caleb something broke here
 
@BTRNaidu This shounds like something you should work up as a full question for the main site, it's not so much a discussion thing.
 
Thanks
 
@mikeserv unfortunately it has not burned in the fiery depths of hell like we all hoped it would. yet.
 
@strugee :g/spit residue/ d
 
10:19 AM
@Caleb sorry, I don't speak stupid
no but actually I don't know what that does besides the fact that it's global (g)
 
@strugee That would search the entire buffer (usually the open file), find any lines containing the string and summarily delete them.
 
@Caleb ah
 
10:51 AM
Linux you kill me sometimes. Why can I drag open URL's to new tabs both directions between Firefox and Chromium and I can't drag a tab from one Chromium window to another?
 
slm
11:18 AM
@Caleb sounds like you're doing it wrong, I do that all the time in chrome
 
@slm No it's just broken, and it's been widely known to be broken for some time. But the bug only shows up with some multi-monitor + tiling window manager setups.
 
slm
@Caleb - hmm. Do you happen to have a link? Also you're using chrome or chromium. I'm able to drag tabs b/w windows w/o issue
 
slm
@Caleb I'll try it when I get to the office where I have 2 monitors. But I'm using chrome not chromium, and I'm pretty sure I've dragged b/w 2 windows across monitors, so it must be a very specific bug to a WM + chromium
 
@slm I'm using AwesomeWM but at one point I checked at it was broken in Gnome 3 too while at the same time I found something else where it worked so it is WM related but it's not a fluke.
Looks like we need a new bug report though, they don't seem to be taking that one seriously.
 
slm
11:34 AM
@Caleb - what's the distro you're using?
 
@slm I'm on Arch today but it is not distro specific. I've confirmed the same behavior on several others (note this has been out there for a couple years!)
Looking through the issue reports there are hundreds of tab dragging related issues, many of them with lots of sightings and no resolution. All a little different though.
 
slm
11:51 AM
@Caleb Yeah I was looking through the issues and saw a lot of that, I was interested if it was chromium only or chrome too
 
@slm It appears to be a mish-mash of both and ever more complicated with the Aura integration code being in both.
 
 
1 hour later…
1:09 PM
Does anyone here have any ideas for this?
0
Q: git fetch - missing commits

GraemeI am running etckeeper on two different machines, boxA and boxB. The OS, etckeeper repository and all, on boxB was copied from boxA back in January. In order to apply some setup changes made on boxA to boxB, I added the repository on boxA as a remote repository on boxB with the intention of doing...

 
1:50 PM
@Caleb I can't reproduce. Chromium 35 and I can drag tabs between chromium windows with no problems
Version 35.0.1916.114 (270117) to be exact
and It seems I misread. I'll try reproducing the actual issue you mentioned, not the one I made up :)
This works for me also. I can drag URLs to tabs in same window and other windows (chromium to chromium)
and the windows were on different monitors if that matters (I see slm mentioned it)
but I'm multi-monitor KDE and not a tiling WM
 
 
1 hour later…
3:21 PM
I'm very much not a unix expert. Can someone please look over this answer of mine that I just made substantial edits to and make sure I didn't accidentally tell people to do something very bad?
 
@Graeme You might have better luck with #git on freenode for example. I use etckeeper but not git.
So, don't have anything useful to say. I do find #mercurial on freenode quite helpful, though there often not a lot of people paying attention. otoh, git has a larger user (and presumably dev) community.
 
3:48 PM
@undergroundmonorail It looks fine to me, +1. Oh, and an Arch user using a tiling window manager and no file manager is hardly a newbie :)
Thanks for taking the time to post your solution by the way.
 
4:06 PM
How safe is it to convert from ext3 to ext4, and can it be done will the block device is mounted?
 
@terdon Normally I'd agree with you, except that my whole experience with Linux before Arch was a couple weeks of Ubuntu. I realized that if I was going to just use it like Windows (the way I was) there was no point, so I decided to jump into Arch. I'm not using expert tools because I'm an expert, I'm using them because it helps me learn. :P
Also I'm using a tiling WM exclusively because I like this gif i.imgur.com/pQT0l.jpg
 
4:23 PM
@undergroundmonorail Ah, fell in love with the geeky coolness huh? I know how you feel :)
 
I love "movie hacking" so much. It's why I like Uplink.
 
4:38 PM
@undergroundmonorail What is movie hacking?
@terdon geeky coolness of what?
 
@FaheemMitha The "hacking"characters do in movies. Anyone who knows things about computers can see that it's nonsense, but it's visually impressive to everyone else.
Example: Every sci-fi movie except Tron: Legacy. That was actually fairly realistic in its hacking scene.
 
@undergroundmonorail Specific example?
 
@FaheemMitha I will try to find one for you
 
@undergroundmonorail ok
 
@FaheemMitha This is not the one I was looking for but it's pretty good youtube.com/watch?v=aApTVqeGJMw
 
4:55 PM
@undergroundmonorail Usual Hollywood silliness.
Thanks for the sample though. I think I've seen funnier and sillier.
 
I'd appreciate some guidance on how to award my bounty here. Could somebody who understands memory management better than I do give an upvote to one of the answers?
5
Q: What used the linux memory? Low cache, low buffer, not a VM

JasonFirst of all, yes, I have read LinuxAteMyRAM, which doesn't explain my situation. # free -tm total used free shared buffers cached Mem: 48149 43948 4200 0 4 75 -/+ buffers/cache: 43868 4280 Swap: 38287...

@Gilles @derobert ?
 
will look in a second, actually finally editing my mod questions answer
 
That's fine, whenever and if you get the chance. It's just that the bounty will be auto awarded in 24 hours and I can't really judge whether this answer is actually relevant.
 
@derobert For a virtual machine, do you typically default to chroot unless there is a compelling reason otherwise? Eg. wiki.debian.org/chroot ?
 
@terdon None of those answers deserve an upvote. In fact all of them deserve a downvote
actually, the memory leak one, if a somewhat crappy answer, may at least be the most promising
 
5:21 PM
@FaheemMitha I'm not following. VM != chroot. They do different things...
 
@Patrick Yeah, that's what I was afraid of. Well, knock yourself out but bear in mind that anything you upvote will probably end up with my bounty :)
 
Or, I should say, a chroot is not a VM.
 
@terdon 1 minute remaining! The pressure!!!
Going to abstain. I dont think any of them are definitively better.
 
@derobert I thought a chroot counted as a vm.
@Patrick Remaining for what?
 
@FaheemMitha The time left on terdon's bounty
 
5:24 PM
@Patrick Ok. Never mind.
 
@FaheemMitha I wouldn't count it as one. It provides hardly any isolation, and further isn't actually a virtual machine. E.g., you can't boot Windows on it.
... or even a different Linux kernel version. It's maybe a very, very poorly isolated container.
 
@derobert At least you can run a different Linux distribution in one. I'm not aware of a formal defn of a virtual machine. Anyway...
Suppose you just want to build stuff for testing? Would you use a chroot?
 
@FaheemMitha Well, you can run some different distros in one (as long as its udev is compatible with your kernel, etc.). And only sort of. You still get, for example, one port 80 to listen on.
@FaheemMitha Yes. I'd just use a chroot for that. Even better, pbuilder/cowbuilder (which, AFAIK, use chroots)
 
@derobert Ok. Have you used pbuilder/cowbuilder? which do you prefer if any?
 
@FaheemMitha Its been a while. Last time, I just used pbuilder.
Note that if you additionally need to test your package (no idea if you're just setting up a build host or not), you'll probably want a full VM (KVM, etc.) to actually boot up testing.
 
5:39 PM
X just crashed again. This is not good. Suggestions for debugging?
@derobert Hmm, I just want to do some basic testing. Probably doesn't require a full environment.
 
@terdon I'm going to have to agree with @Patrick on that one...
 
Just did it like for the first time a day or two ago. Forget exactly when, but I commented on it in this channel.
Maybe yesterday?
23 hours ago, by Faheem Mitha
My X server crashed for no reason. A bit worrying. Why would this happen? Maybe my graphics card is dying? It is quite old. The logs just said server terminated unexpectedly.
Yes, almost exactly a day ago.
 
@FaheemMitha Well, the standard stuff: (1) did you change anything? (2) anything useful in the X log (probably not)
 
23 hours ago, by casey
@FaheemMitha could be that, or memory, or a misbehaving X client...
@derobert (1) not that I am aware of. Hardware, definitely not.
As far for (2), you mean the usual X log which show the server starting?
[1081568.802] [mi] EQ overflowing. Additional events will be discarded until existing events are processed.
[1081568.802]
[1081568.802] Backtrace:
[1081568.802] 0: /usr/bin/X (xorg_backtrace+0x36) [0x7f17a0d56d16]
[1081568.802] 1: /usr/bin/X (mieqEnqueue+0x26b) [0x7f17a0d37fbb]
[1081568.802] 2: /usr/bin/X (0x7f17a0bd8000+0x66902) [0x7f17a0c3e902]
[1081568.802] 3: /usr/lib/xorg/modules/input/evdev_drv.so (0x7f17979a8000+0x55d8) [0x7f17979ad5d8]
[1081568.803] 4: /usr/bin/X (0x7f17a0bd8000+0x8d947) [0x7f17a0c65947]
Doesn't give the time clearly. May be unrelated.
Could Adobe Flash crash X? I think I was playing a movie via putlocker both times.
Possibly a coincidence.
 
I blame adobe
are your nvidia drivers up to date?
 
5:53 PM
@FaheemMitha That's the time since boot in [ ]
@FaheemMitha I was thinking software. You have to suspect updates from between the last time you restarted X and the first time it crashed.
 
@terdon ssl's answer doesn't make much sense to me, and has at least completely wrong point (but it might not invalidate everything). I commented.
 
@derobert In what units?
 
Avery Payne's hypothesis of a kernel bug isn't unlikely — but more details would be appreciated
 
@FaheemMitha seconds
 
@derobert That's a long time. Do you suggest a reboot?
 
5:54 PM
@FaheemMitha seconds
 
If it was me I'd ask for more information — he says he did “delve a little deeper”, so what are the specific problems that he found?
@FaheemMitha did you do a RAM test recently?
 
Current uptime is 12 days.
 
@FaheemMitha It is indeed a long time, but the libraries & X aren't reloaded when they're running, so updates don't take effect. /var/log/apt, /var/log/aptitude, and /var/log/dpkg may help
 
@Gilles No.
 
12 days isn't long. I typically go about 60-180 days between reboots, depending on when I feel like upgrading my kernel
 
5:55 PM
What do you suggest, memtest86?
 
and X runs for most of that time, though I do ocassionally logout and back in when the nvidia drivers get updated
 
@casey True. I often have this computer up for months. I think I had it cleaned, that is why it was taken down.
 
@FaheemMitha wait, you had it cleaned?
 
mine also goes down occasionally when the power goes out for more than 3 minutes. I love my power company :/
 
@derobert Right. And inspected.
Gets very dirty here.
 
5:57 PM
just open the case and spray some compressed air in there :)
 
@casey No UPS?
 
@FaheemMitha the UPS is why It goes down 3 minutes after the power goes out. Technically I think I can stretch it to 7 minutes, but I don't push it.
 
@casey Well, it requires a little more than that, after it has been sitting here a few months.
 
@FaheemMitha pretty safe. It isn't exactly a conversion, but rather activating some extra features. Though it does require a run of fsck if you enable the uninit_bg option (I have no idea whether it is beneficial in practice)
 
The server downstairs can go 30 minutes on battery
 
5:58 PM
@FaheemMitha So someone did touch the hardware. You could try a memtest, but I'd first check to make sure you didn't just suffer an nvidia upgrade, or similar
 
@derobert Unlikely, but Ok. I've certainly upgraded some stuff, but I nothing I think would cause a problem.
 
Hah, my UPSes are good for at least 2h...
 
@casey 3 min is a pretty short time. Do you hava big/heavily loaded system?
 
@derobert mine will run that long if I don't load it up :) These are just cheapo residential units to mostly massage the incoming power before it gets to my PSU and keep hard power downs from occurring every time the wind blows and knocks a tree down. (3-4 times a year, sadly)
 
@Gilles Ok, thanks. Can I run it on a mounted filesystem which is in use? One of these is /home.
 
6:00 PM
@casey I hear ya. I spent over $300/UPS for mine... And I have three of 'em.
 
@FaheemMitha right now the USP says I'm pulling 240 W. I'm not really doing much (0.3 load avg) but I do run 6 spinning disks and a GPU driving 3 displays.
 
@casey UPS with display? Nice. I couldn't get one here, at least not APC.
 
@FaheemMitha BTW, speaking of interesting bugs filed, bugs.debian.org/cgi-bin/bugreport.cgi?bug=183486 ... Interesting, because Stephane hasn't found out about it yet, and he knows all. (see his comment at unix.stackexchange.com/questions/56093/…)
 
@derobert they are worth it. I'd have gone bigger if it made sense. Here if the power goes out its either back up in a minute, or down for 12 hours or longer. No real middle ground.
 
However, to get maximum benefit from ext4, you need to recreate the fs from scratch. Ext3 conversion won't convert your existing directories to btrees, for instance.
 
6:03 PM
@casey Yeah. We get some 30-60m outages.
 
@derobert checked apt logs. nothing noteworthy, except chromium seems to have had quite a lot of upgrades recently. considering this is supposed to be debian stable. i'd rather they just left it alone, but i guess these are security bugs.
 
@FaheemMitha memtest86+, packaged in all good distros
 
@Gilles Bummer.
 
@FaheemMitha I think you can run tune2fs on a mounted filesystem and it takes effect on the next mount, but I'm not sure
 
@Gilles I wonder, does e2fsck -D do that?
 
6:06 PM
@derobert my favorite "UPS" was the one in the calling center my dad worked in in the 90's. Almost 50 12V lead-acid car batteries hooked up to a controller of some kind powering a Nortel Meridian 1 switch and a few racks worth of autodialers.
didn't last long either, just long enough to go from mains to generator power
 
@derobert hmmm, it looks like it does
 
@casey That's how data center UPSs are built, rooms full of batteries
 
Did it always do this? I remember researching this back when I was planning to convert my home machine to ext4, and concluding that I'd have to rewrite (so rewrite I did)
 
@Gilles Not sure if it always did.
 
Maybe I got it wrong then, maybe I misremember, maybe the possibilities have improved
 
6:10 PM
@derobert I once helped a friend move a server from an apartment to a house he'd just bought. He didn't want to shut it down, so it was a very delicate move with 3 of us each carrying the computer, the UPS and the large heavy batter that was rewired in place of the normal UPS battery. I was impressed the disks survived the drive...
 
@casey Wow, was that just to preserve uptime bragging rights? I hope so!
 
@derobert iirc, that was exactly the reason
At that point it was at 400 days or so
that friend was a bit crazy though. First thing he did to that house was have the electric company run another line (200A iirc) that just ran to two of his rooms and then had 4-socket electrical plugs installed every 16" or so along the walls of those rooms.
A bit overkill, but I think there was ~50 computers in there at its peak
 
@Gilles Ok, thanks. It's not really urgent/important - ext3 still seems fine.
@casey Sounds risky.
@casey That's a lot of computers. What was your friend using them for?
 
@FaheemMitha we wrote a client to a then-popular MMORPG and were massivly exploiting it in parallel.
 
So, would people care to estimate the probabilty that Flash running on putlocker is crashing X? I'm in the middle of watching the Hobbit, and would like to continue, but not if it crashes X. Thorin and Co. have just reached Rivendell. I must say, they've taken a ton of liberties, most of them not improvements.
 
6:20 PM
@FaheemMitha not too bad. He had the power properly installed and had an additional AC unit put in just for those rooms. I'd say expensive more than risky :)
 
Jackson really isn't much of a filmmaker though he seems to have a lot of money.
@casey Do you make money off that? Unclear what that is.
 
It was quite profitable for a very short span of time.
 
@casey Ok.
 
Has mikeserv been at other people's throats lately, or is he just targeting me?
 
@Gilles thats normal for him
 
6:24 PM
For the record, I'm trying to get out of any further interaction with him. He's made me more frustrated with U&L than anyone else had ever come close to.
 
@Gilles Huh?
 
@mikeserv I am now asking you to stop commenting on my posts. Let us live separately. Downvote my posts if you don't like them; I'd prefer not to (unless they are actually incorrect of course), but if playing silly buggers with the serial downvoting detector makes you happy, enjoy. As for plagiarism being “common practice”, I'll ask you not to slander me. This is, hopefully, my last interaction with you. Take any further concern to the moderators. — Gilles 2 mins ago
 
@Gilles he seems pretty nice in general. I think this is one of this case where peoplecan get bananas when you push the right button.
We all have our corners of insanity.
 
@casey not understanding stuff and being stubborn (while pretending he's trying to learn) is normal for him. But it's gotten to stalking levels lately, plus the accusations of plagiarism
 
Good morning.
 
6:26 PM
@Gilles yea, seeing what you posted there, I agree he's a bit more on edge than normal
 
@FaheemMitha mikeserv, pretty nice???? I'm trying to steer clear of his buttons
 
@Gilles Ok, I take that back. The comments in those threads seem pretty strange. Not sure what is going on there.
@Gilles Well, when I've chatted with him on chat. Hardly much of a basis, granted, but that is all any of us have.
@casey Maybe he is having a really bad day. Though taking it out on strangers on the net isn't cool.
 
@FaheemMitha Gilles isn't really a stranger to U&L or people who chat here regularly.
 
@FaheemMitha The 1st Hobbit movie was pretty bad. 2nd one is better. Of course, what someone needs to do is (once all hundred or so are out) cut together the scenes that Jackson didn't just pull out of his rear and then you'd have a good 2-3h version of the Hobbit!
 
@casey Well, I mean stranger in terms of knowing the person in real life / in the flesh.
I mean, we could all be aliens or robots talking to each other, for all any of us know.
@derobert That might be a useful project. Would be a bit disjointed though.
 
6:31 PM
@FaheemMitha managed to work out what was wrong by myself in the end. Git is confusing when things don't work! Good way to learn though. Thanks for the suggestion anyway.
0
A: git fetch - missing commits

GraemeSo the problem here turned out to be that there was a detached head in the repository on boxA after an incomplete fix of a botched rebase. When this happens the output of git branch is as follows: # git branch * (no branch) master The fix was simply to create a branch on boxA and then merge ...

 
I remember the mines of moria scene being pretty good, but it has been a while since I watched it. I don't think Jackson really gets Tolkien otherwise, though.
 
@Gilles I have him on ignore here in chat, and I won't touch anything of his, not even to downvote it.
 
@Graeme Pretty trivial suggestions, but you're welcome.
I find git super-confusing myself.
Though maybe there is a reason to it all.
 
Can the action of change user in a terminal be interpreted as a context switch?
 
That scene of Radagast running away from the Wargs being pulled by a pack of rabbits was pretty bad. Tolkien would have died.
 
6:33 PM
@Lucio The term normally means when the processor goes from user space to kernel space, or vice versa....
 
Could you give me an example?
 
@Lucio please provide some more context. (No pun intended)
 
@derobert I'm not really a Tolkien expert or even a fan, but I do think people's work deserves to be treated with respect. Something Hollywood never understands.
 
@Lucio etymologically, yes, it's a change (a switch) of something that could be called a context
 
@derobert yes, sorry, I do understand the concept but it is very abstract without identifying the action in an example
 
6:35 PM
@Lucio If a program makes a syscall, then the processor must stop executing user (unprivileged) code, and instead start executing kernel (very privileged) code. That's a context switch
 
but as an OS professional, when I hear “context switch”, I immediately think “switch between tasks with their own virtual memory space”. Users and terminals don't enter into it.
 
Oh, I see.
 
@derobert Interesting patch, but shouldn't this go upstream?
 
@FaheemMitha It didn't? Figured the maintainer would have forwarded it on...
 
@derobert not just kernel/user but also any task switch can be called a context switch (even if a user/user context switch involves a sequence of two context switches, user/kernel and kernel/user)
 
6:36 PM
@derobert Ok,maybe it did. I should have said, why not submit it directly to upstream?
 
@Gilles Sure, I'll agree with that.
@Lucio both what Gilles and I gave are how context switch is used in Unix, in a strictly technical sense.
@Lucio if you're trying to use it in a different sense, well, it's hard to say if its OK without knowing more background.
 
Yeah, I just wanted to have the basic technical concept.
Both points are helpful for me
 
@Lucio a processor can only do one thing at a time. To multi-task it just switches what it is doing many times per second. Each of these switches is a context switch.
 
Exactly, but that action is performed by OS
 
you are basically saving the state of the processor in your task, and then loading the saved state from a different task.
 
6:40 PM
@Lucio Correct. It's one of the core responsibilities of a multi-tasking OS.
 
May I (an user) do a context switch from terminal? This question may be found as pointless but is something that I really don't know..
What I think is that only the OS and a father process may do it. Am I wrong?
 
@Lucio run a program. You just caused a context switch. Or just sit there and stare at the terminal, lots of context switches are happening anyway
 
@Lucio Sure. Run ls from a shell. There's at least a context switch when ls starts and one when it finishes
 
Ahhhh
 
@Lucio Well, if you switch from writing code to reading an instant message, you could say you've done a context switch.
 
6:44 PM
but that won't be true if I'm running a multi-core processor where one processor is free at that time, right?
 
@Lucio You'll still take a few from the syscall context switches I was talking about earlier.
 
@Lucio technically, yes, but it's highly unlikely
 
right
 
and that's only talking about process-to-process context switches
if you count all of them, then every syscall involves two context switches
 
And, for example, Linux may decide to leave the 2nd core unused to save electricity. So maybe it'll run your ls on the 1st core.
 
6:46 PM
also, when you're typing ls, many processes are involved: the X server for input and again later to display the characters, the terminal emulator (again both in and out), the shell
 
Yep. Each character you type actually causes many context switches.
 
@Gilles that is true. Identifying PIDs I could see that any new command would have at least bring one more process executed.
@Gilles store and restore, you mean
 
@Lucio A switch from your process's context to the kernel's context, and then a switch from the kernel's context to your process's context
Even on a fairly idle system, there are hundreds of context switches per second.
 
oh, right, you wrote that above
 
Watching vmstat 1 here, for example, there are about 1000/sec on my system right now. And it's not really doing much...
 
6:51 PM
wow
 
By wiggling the mouse around a lot, I can add another 500/second
 
That gives a nice idea, he
 
Actually, found a nice spot to wiggle it between more different windows, and added 2,000/second.
As you can guess from the number of them, trying to understand each one of them individually would keep you busy for a very long time.
 
The system is begging for you stop doing it! Haha
Yeah, now I have a good basic idea of it.
Thank you guys
All of you are great!
 
and if you have a cheap NIC, every time a packet arrives on the interface its going to interrupt and cause a context switch
 
6:55 PM
Are there NICs that handle themselves the packages, do you mean?
 
@Lucio Some NICs (most, actually) can be set to only interrupt only n packets when the packet rate gets high.
 
oh, that should give a big queue
 
Well, depending on n, yes, it could be a big queue. But its much more efficient to process a few packets at once than one at a time. It does of course increase latency, which is why its only used when the packet rate gets high.
@Lucio there is an explanation over at Server Fault: serverfault.com/q/241421/1006
 
Now I am learning about network administration, heh
 
Performance tuning, really :-P
There's a lot you can learn from reading answers around Stack Exchange.
 
7:11 PM
That is only because of the people :)
I have to go now.
I appreciate the help and time explaining everything, thanks a lot! ciao
 
7:37 PM
The instructions at the beginning of the configuration section in wiki.debian.org/chroot don't work. I get
root@orwell:~# cat > ./usr/sbin/policy-rc.d << EOF
> #!/bin/sh
> exit 101
> EOF
bash: ./usr/sbin/policy-rc.d: No such file or directory
What am I doing wrong?
 
@FaheemMitha running this from the wrong directory? You should be at the root of your chroot
 
There needs to be a checkbox when editing a question to say that "OP was so incompetent further upvotes should be credited to nobody".
 
@Gilles That's right. Thanks.
I switched directories without realising it mattered.
 
8:30 PM
What are applications of
1
A: Is there apt-get for windows?

derobertI think you're looking for apt-offline. On your Linux machine with no Internet connection, let's say you want to install emacs. You'd do: # apt-offline set --install-packages=emacs /media/usbstick/offline.sig Then, on your Windows machine with an Internet connection, you'd do: E:\> apt-offli...

Burning debs to CD in windows to transfer to a Debian machine?
 
@FaheemMitha yes
 
@Gilles ok
 
@FaheemMitha Yep. Though a USB stick is probably more normal.
 
or nowadays, usually a USB stick
 
Say you have a fast Internet connection at school, but dialup at home. Download your packages at school, bring them home on USB.
 
8:32 PM
been there done that too
 
people have it so easy today. I remember my first cd-burner. 2x and I shared it with a friend since we pooled our money to buy it.
 
@derobert Right. Though I'd hate to have dialup at home. Been there, done that, hated it.
 
except for me it was before USB sticks, I used zip drives and floppies
 
teamviewer kind of software will be more helpful right?
 
I got a cable internet connection in 2002, I wonder now why I waited to make the switch.
 
8:33 PM
@FaheemMitha I had dialup at home, 15 years ago. Maybe 16.
But I didn't have anything faster elsewhere.
 
@derobert So late 90s?
 
Yeah.
 
i started with a 2400 bps hayes AT modem dialing into local BBS. I didn't get cable internet until I moved out of my parents place in 1999
 
Afterwards I wished I'd switched earlier. It was $30/40 at the time, but dialup caused me much more than that amount of trouble.
 
I don't remember what year we got internet in, but it was probably 96
 
8:36 PM
@casey Ah, the good old days.
I wonder how many people in the US are still using dialup.
 
Probably way too many. I know various places around here still use dialup credit card terminals.
 
@casey In the summer of 1998 I remember wandering Chapel Hill trying to find a CD burner to use. I and a friend were trying to burn Debian to CDs. We failed.
Apparently burners were a rare commodity once upon a time.
@derobert That sucks. Money or access?
 
Yes, they were, I think my first CD burner was well over $200.
 
"dad, can I leave the computer on overnight and tie up the phone lines to download this 0.5 MB file?"
 
@FaheemMitha I'd guess its cheaper for them.
 
8:39 PM
@derobert Is cable still around $40 for a reasonable quality connection?
 
Hah, I tied up the phone line for something like 6 nights straight...
 
@casey did you live in a place with unmetered local calls and unmetered Internet? Because in France this just didn't exist. You paid by the minute.
 
@FaheemMitha I pay ~50/month for 30 mbit/s down and 6 up
 
@FaheemMitha Somewhere around there. Probably less with a TV bundle, maybe a little more without.
 
@casey That's common in the US now?
 
8:40 PM
@Gilles Most of the US at that point had unlimited local calls
 
@Gilles thankfully no per-minute charges here. Some internet providers were time limited, but I went with ones that were not
 
@derobert I thought it was close to 100%.
 
you just needed an ISP with a local number
 
@casey Right.
 
@FaheemMitha Yeah, probably was. Not sure what percent had a local ISP number.
 
8:41 PM
I remember that.
 
At one point had a network inside my parents house, with a dialup modem on a Linux server. It was set to auto-dial on Internet traffic... So much nicer once we switched to cable.
 
choclatey looks interesting. Assuming one is cursed with using Windows.
 
ninite.com is useful too on Windows
 
04:00 - 21:0021:00 - 00:00

« first day (1349 days earlier)      last day (3606 days later) »