/dev/chat

Braiam

00:53

3

^ this is useful sometimes...

terdon

@Braiam Ha! Very true :)

Tim

01:19

Is apt only used for installation, uninstallation, upgrade of applications, not for getting info about installed packages?

dpkg seems to me can do all: installation, uninstallation, upgrade of applications, but also getting info about installed packages

so the functionality of dpkg is a superset of those of apt?

hi slm. cong for becoming a mod.

strugee

01:37

@Tim no, APT is the high-level package manager, DPKG is the low-level package manager

APT is a system wrapped around DPKG

DPKG tracks what packages are installed. APT tracks where packages come from and does intelligent things with that information.

e.g. APT keeps track of whether a package was installed automatically or not, and what repository on the internet it was installed from.

so really, if anything, APT functionality is a superset of DPKG functionality

@FaheemMitha I have

hey @FaheemMitha if I was to start learning TeX, what should I look into? TeX? LaTeX?

something else?

@Braiam that's only for section 1 and... what, 6?

Tim

01:55

@strugee thanks. DPKG tracks what packages are installed. Does APT also track what packages are installed.

strugee

@Tim APT, IIRC, tracks what packages are installed through APT.

it does this so that e.g. it can track whether or not a package was installed as a dependency.

mikeserv

@terdon - it's better like this:

Tim

can we call APT to get info about installed packages?

mikeserv

tr \\0 \\n </dev/zero |
MAILPATH= z0=0 ENV= \
PS1='$((n=${2-1}*$((l=l${l++1}))))\
${z:=${z0#$((n+${2-1}>${3-100}))}}${1-
}${z:+$(echo $n >&2;kill -PIPE $$)}' \
dash -si 2>&1 -- , 5 10009 ||:

strugee

@Tim yes, I believe so

maybe. I'm not sure.

mikeserv

02:06

I rearranged it a little so I don't check the line number twice. Also, before I noticed that if the upper limit didn't align with the interval it would print the separator once more following the last number printed. It doesn't now. I think I need to change it to avoid multipliers as well and just use addition since I'm pretty much already doing so.

Ramesh

What exactly is the difference between 32 bit and 64 bit OS?

mikeserv

I put the MAILPATH= thing in because strace showed me that for every $PS1 printed the shell was stat() my mail dir?

strugee

@Ramesh it has to do with the hardware

Ramesh

@strugee, I installed 32 bit OS in a 64 bit machine.

strugee

specifically whether the hardware is able to handle 64-bit numbers or not

mikeserv

02:08

@Ramesh - digitaltrends.com/computing/32-bit-64-bit-operating-systems/…

Ramesh

However, I am trying to convince that there won't be much of performance difference.

strugee

@Ramesh that will work just fine, but bear in mind that it will work just like a 32-bit computer. it will have all the limitations associated with 32-bit hardware.

e.g. not being able to access >4 GB of core

mikeserv

Also, the upper memory allocation limit on a 32-bit processor is typically 4GBs - that's its bandwidth you might say.

Ramesh

@strugee, no.

strugee

@Ramesh ?

Ramesh

02:09

We can access more than 4 GB. In fact we can access upto 64 GB.

mikeserv

But with processor extensions that is largely expanded.

Yeah - there you go.

Ramesh

If we enable PAE it is possible.

mikeserv

But MS never built that in.

strugee

okay, fine. but still.

Ramesh

And by default, PAE option is enabled in RHEL systems.

strugee

02:10

I don't understand why you wouldn't just install a 64-bit system.

Ramesh

So, I mean if we have 64 GB RAM, I am not sure if there would be a performance difference between 32 bit and 64 bit.

@strugee I already installed thinking it was a 32 bit.

mikeserv

Actually - in some cases the 32-bit os will outperform the 64-bit os.

Ramesh

It was my bad. But still I am trying to gather enough info to let people know that there won't be much performance difference.

mikeserv

Its binary lower limit is smaller.

strugee

@Ramesh 64-bit also allows things like better ASLR. not sure if PAE will fix that.

Ramesh

02:11

@mikeserv Any example?

mikeserv

Also - and this is the real problem with PAE - it's not very well supported. I'll get some examples. But I can say that kernel maintainers have made it clear that PAE and 32-bit processors in general are not large concerns of theirs.

Ramesh

Ok, how about this? Since the hardware supports 64 bit, can I install 64 bit JVM and other things?

strugee

@Ramesh no.

since your OS is not 64-bit.

you'd need 64-bit libraries and, most importantly, a 64-bit kernel.

mikeserv

^^this

Ramesh

@strugee ok. But exactly in what way I will have a performance difference? I mean still I can use 64 GB RAM. So the address space I can use is still enough, right?

@mikeserv ??

mikeserv

02:16

27

Q: Are 64 bit programs bigger and faster than 32 bit versions?

I suppose I am focussing on x86, but I am generally interested in the move from 32 to 64 bit. Logically, I can see that constants and pointers, in some cases, will be larger so programs are likely to be larger. And the desire to allocate memory on word boundaries for efficiency would mean more w...

Ramesh

@mikeserv This one gives an opposite view.

3

Q: What a 64 bit Linux can do that 32 bit linux can't?

I ordered a server from fdcserver. Then I realized that, unless changed, fdcserver actually provides 32 bit operating as default. I logged in using SSH, typed uname -m, and to my horror I saw that I was actually running a 32 bit Linux. I will definitely change this near the end of the month. Me...

linux memory 64bit x86 32bit

mikeserv

I guess I was wrong - MS did enable the extended mem addressing for PAE in XP, but then removed it...

Physical Address Extension

In computing, Physical Address Extension (PAE) is a feature to allow 32-bit IA-32 central processing units (CPUs) to access a physical address space (including random access memory and memory mapped devices) larger than 4 gigabytes. History PAE was first implemented in the Intel Pentium Pro in 1995, although the accompanying chipsets usually lacked support for the required extra address bits. It was extended by AMD to add a level to the page table hierarchy, to allow it to handle up to 52-bit physical addresses, add NX bit functionality, and make it the mandatory memory paging mo...

Ramesh

What does this statement mean?

Each application can only see 4 GiB at a time (and some of that 4GiB must be used for other things, the exact amount depending on the "Memory split" kernel setting)

mikeserv

Took me a few minutes to find this, but I knew I'd read it somewhere. Here it is from the horse's mouth:

phoronix.com/scan.php?page=news_item&px=MTUwMzI

Honestly, I don't know @Ramesh. I guess that the kernel has to shift memory around as requested so probably it can't mmap() a single chunk of 4gbs? I don't know.

Ramesh

so with 64 bit, this restriction won't be there?

mikeserv

02:29

I don't know. But if it is as I say, then the application trying to do an mmap() like that probably needs to refactor.

casey

02:59

@Ramesh PAE and the 4GiB stuff only apply to 32 bit

Ramesh

@casey, It's actually interesting. RHEL used to provide hugemem for 4 and 5 which they have stopped.

with hugemem, 32 bit can support more than 4 GB.

redhat.com/resourcelibrary/articles/…

casey

its not really needed these days except on older machines

Ramesh

I unfortunately installed 32 bit on a 64 bit without realizing the consequences.

With 32 bit PAE enabled, upto 4 GB of memory space can be used at a time.

However, with 64 bit OS this limitation is not there.

casey

yep, my research code routinely allocates ~20 GB of RAM, something I couldn't do on a 32 bit OS (even with PAE)

Ramesh

Another interesting observation. Mysql maximum query size can be set from my.cnf file.

Assuming, I set the maximum query size in the my.cnf file to 8 GB will it fail in my 32 bit system?

See, 32 bit system supports 8 GB but only in 4 GB chunks each.

And mysqld needs to be run as a single process.

strugee

03:38

@Ramesh performance-wise, not a huge amount.

but like I said, 64-bit affects other things too, like ASLR

ASLR can be broken with medium difficulty on 32-bit machines. on 64-bit, it is much much harder.

Ramesh

03:54

@strugee, thanks.

From wiki page of PAE,

The Linux kernel includes full PAE mode support starting with version 2.3.23, with Linus Torvalds mentioning PAE's 4 MB page support in 1.3.15,[18] enabling access of up to 64 GB of memory on 32-bit machines.

Now, I understand with 32 bit OS, I can use 64 GB physical memory.

I am confused with the virtual memory concept here.

Do you have an idea on how virtual memory will work on 32 bit systems?

strugee

04:28

@Ramesh do you understand virtual memory in general?

Ramesh

@strugee, just now was reading about it.

strugee

@Ramesh ok, let me know if you have any questions

I see you posted a question. nice job remaining true to the Stack Exchange Way(tm)

Ramesh

@strugee, I got a wonderful answer too. :)

0

Q: Maximum memory usable by a 32 bit RHEL 6 system

I have installed a 32 bit RHEL OS on a 64 bit system. From the wiki page of PAE, The Linux kernel includes full PAE mode support starting with version 2.3.23, with Linus Torvalds mentioning PAE's 4 MB page support in 1.3.15,[18] enabling access of up to 64 GB of memory on 32-bit machines....

rhel process pae

strugee

@Ramesh ooooo, nice!

Ramesh

@strugee, this explains it all.

If you have a four-car garage at your house, you can still own fifty cars. You just can't keep them all in your garage.

strugee

04:32

@Ramesh that is one of the best metaphors I've seen in a long while. A++ upvoted

Ramesh

yeah. It just makes the concept clear as water.

mikeserv

How many monster trucks?

Ramesh

2

mikeserv

OK!

Ramesh - that's a super QA.

Ramesh

@mikeserv, the dude's metaphor is amazing.

mikeserv

04:38

You talk about yourself in the third person now?

Ramesh

That is not my answer :) copy pasted from superuser

mikeserv

That point about swapping it out is what I meant by the application should refactor - allocating that much memory in a single mmap() is crazy. Doing cat /proc/$pid/maps will show you how a process addresses memory.

The kernel builds out virtual file descriptors for every mapped address.

Ramesh

@mikeserv, is there any application that you could think of which could use more than 4 GB at any point of time?

mikeserv

Well, sure. lots - your answer has a good point about video processing. Also hypervisors usually do memory ballooning - which are large memory overcommits managed by the hv.

But in a single map? I doubt it.

Ramesh

yeah exactly that was my question.

mikeserv

04:45

I dunno though. Maybe there's a use case for it. It just seems like it would be a lot of work for the programmer to write any kind of code that could do anything useful with a chunk like that.

Ramesh

Or as per dude's metaphor, we can phrase it as, having 20 people in a single car...

mikeserv

Yeah - and he points out shmem too.

You can privately map memory or map it as shared. shared memory sticks around until all pointers to it are closed, I think.

And the kernel handles all of it. The big deal with virtual memory is that it isn't necessarily where the kernel tells an application it is - that's what is meant by virtual. The addresses exist in kernel - they're not actual memory adresses. The kernel proxies all calls to memory.

Ramesh

So in what ways people say 64 bit is advantageous over 32 bit? Is it the virtual memory allowed more in 64 bit case?

mikeserv

Well, I don't know specifically. I think at present most applications are pretty much fully backwards compatible.

But I think the real advantage lies in future-proofing what you do - people don't really like doing stuff twice, so I guess sooner than later that backwards compatibility will fade.

Ramesh

Why is the choice of 32 bit or 64 bit OS not made automatically by the OS? I mean, the kernel can probe the hardware and determine it in runtime and based on that install the necessary drivers and so on, right?

@mikeserv Yeah, RHEL7 doesn't provide support for 32 bit I guess.

mikeserv

04:53

Well, yeah, but this isn't about drivers - this is about where those drivers are kept.

The kernel has to communicate with the processor which speaks a certain language.

It says - yo, cpu, put this driver here and be ready when I ask for it.

Ramesh

I may sound naive but still that capability should not be left to user choice right?

mikeserv

Those are the memory addresses - they're all proxied by kernel space. It's the kernel's domain completely - userspace never gets a glimpse at real memory addresses. Vm only and totally.

Ramesh

►

cartmenasa, you rack disciprine. :D

mikeserv

You've got a lot of those south park videos.

Ramesh

Karate Instructor: Cartman-san! What are you doing?
Cartman: I'm doing some sweet banzai moves. I'm a little better than everyone else here.
Karate Instructor: Eric-san you must forrow direction! You rack disciprine!
Cartman: Nuh-uh, I don't rack disciprine!
Karate Instructor: Risten! You all need more disciprine! True disciprine come from within.

@mikeserv It is the best timepass

I could say southpark would be apt for any situation.

mikeserv

05:06

@terdon: I edited my $PS1 thing again:

    tr \\0 \\n </dev/zero |
    MAILPATH= z0=0 ENV= \
    PS1='${n+$((o=n-${2-1}))${1-
    }}${z:+$(echo $o >&2;kill -PIPE $$)}\
    ${z0#["${z:=${z0#$(($((n=n+${2-1}))>=${3-100}))}}"0]}' \
    dash -si 2>&1 -- ',' 5 100 ||:

Ramesh

what is ps1?

mikeserv

The regular shell prompt.

^^that thing turns the shell into a kind of nl.

Or maybe seq.

The , 5 100 arguments tell it to separate each number with a , and count upwards to 100 in increments of 5.

It uses the tr thing as a source - the prompt is written out for every \newline read in until the upper limit argument is reached - or 100 if none is provided - at which point the shell kill -PIPEs itself.

But the entire thing is done without a single subshell until the very last instance when I need to use kill - all tests and everything else occur only as side effects of arithmetic ops and or variable expansions.

Crap. The second to last line needs a > rather than a >=.

Ramesh

@mikeserv, quick question. How is the process size determined?

mikeserv

What do you mean?

Ramesh

I mean, for example, if my vim editor has to open a 10 GB file, I suppose it will be a process.

So here, the vim editor can use page tables to get the data that is needed and so on.

I mean, the entire 10 GB file need not be in virtual memory or address space, right?

mikeserv

05:21

Oh - well if you open a 10 GB file the kernel provides a file descriptor - it points to a file and retains your file offset.

I think that reading backwards in a file means either reopening the file and getting a new descriptor, or reading backwards in a buffer that the process has stored.

So if I open a 10 gb file I should buffer it by read region - which allows me some give and take from my offset - but mapping the whole thing is probably an awful idea.

Ramesh

Ok. Who determines whether or not it should be mapped?

Kernel? Hardware?

mikeserv

Well I do - I read it in from my descriptor and request the memory maps for the regions I wish to retain in memory.

But attempting to map 10gbs is almost certain to fail - which means I'll have to do the whole thing over again anyway.

Ramesh

So that design is based on the coding, I suppose. Am I right?

mikeserv

It's - I think it's called the ABI?

Ramesh

Seems so. APplication binary interface.

mikeserv

05:29

Maybe this will help?

kernelnewbies.org/ABI

And that's the same reason the 4gb map won't happen either - the kernel's bound to want that memory back if it gives it out at all, and as soon as it takes it you have to read in the whole 4gbs all over again.

Ramesh

Nice. I have bookmarked the page. Will give a read tomorrow pbbly.

cartman-san time!!

mikeserv

In the source .../Documentation there's a huge ABI readme, too.

Faheem Mitha

07:57

@strugee Most people use LaTeX. Some people use Context. A tiny fraction of people use plain TeX.

Faheem Mitha

08:19

I'd generally recommend LaTeX. It has far the largest userbase. Though I heard Context has some things to recommend it.

Its design might be better. It is more modern.

I think it has native PDF support.

Braiam

11:47

@strugee apt tracks whatever dpkg says it's installed, just with less detail, apt doesn't know if a package is half-configured or not

got debian bronze badge, yay!

Faheem Mitha

@strugee No, apt knows whatever dpkg tells it. The package doesn't have to be installed through apt, it can be installed directly via dpkg or using some other front end.

@Braiam Congrats.

@Braiam Apt does know about the package status, because it gets that information from dpkg.

Braiam

@FaheemMitha what is the difference?

btw, apt doesn't know the detailed status, only differentiate between installed, uninstalled, purged

Faheem Mitha

12:03

@Braiam The difference between what?

Braiam

@FaheemMitha what I said and what you said

Faheem Mitha

@Braiam It knows whatever dpkg tells it, afaik.

@Braiam Hmm, I guess maybe a matter of semantics. Depends how you define "know", I suppose.

I don't actually know what API dpkg presents to apt.

Braiam

@FaheemMitha dpkg doesn't present api to apt, apt just query the same database where dpkg stores stuff

Faheem Mitha

@Braiam Oh, it that so? Yes, I guess that makes sense. That wacky text only database? One would expect that to be a bit of a speed bump (no hashing?) but I don't know about performance issues.

@Braiam do you answer more questions on AU?

Faheem Mitha

15:13

LWN says Freshmeat has just shut down. Apparently was renamed to freecode before it was.

Namrata

@terdon Hi Terdon. Ok so what I want to do is, I want to compare Contig1 23 of File1 with all the columns of File2 first. Since the last value of File2 is Contig1 24. That is the reason it is not printed since it falls in the range (23+10). Next I compare Contig1 42 of File1 to all the values of File2. Since the first value of File2 is Contig1 40, as 40 belongs to the range (42-10). Next I compare Contig2 68 of File1 to all the values of File2. Since there is Contig2 90, which DOES NOT fall in the range of (68+10) or (68-10). Therefore, I print it.

terdon

@Namrata Yes, I am beginning to understand.

Faheem Mitha

@Namrata you are a bioinformatics PhD student?

Namrata

@terdon Yes I am PhD student in Bioinformatics

terdon

@FaheemMitha the last one was for you.

@Namrata just out of curiosity, what are the numbers associated with the contigs?

Faheem Mitha

15:27

@terdon got it. :-)

terdon

(I am also a bioinformatician)

Namrata

@terdon The numbers associated with the Contigs are the Score of the Contigs.

Faheem Mitha

And I've done some bioinformatics work.

terdon

From assembly?

Faheem Mitha

@Namrata you might do better to handle this kind of thing in R. A database is also a possibility, but R would probably cope easier.

terdon

15:28

I'm just wondering why you would have multiple scores for the same contig in the same run (assuming each file is a separate run).

Namrata

@terdon Yes from a genome assembly.. Sorry the numbers are position of a SNP in a contig.

terdon

Ah! That makes more sense:)

One more thing, are your values space separated or tab separated? Do you have a preference?

Namrata

:) They are space separated. I used awk to print these two columns from a vcf output file.

terdon

OK, I'm working on a Perl solution since I find that easier. Someone else might give you an awk one too. Is Perl OK?

Namrata

Yes fine for me.

Faheem Mitha

15:47

I wonder why trad unix tools are so popular for bioinformatics people. it's the last thing i'd think of using.

Anthon

Someone just suggested an SQL engine over CSV files in a comment

Just found a SQL engine over CSV files: github.com/harelba/q — LatinSuD 32 mins ago

Why deal with only one problem when you can have two !!

Faheem Mitha

@Anthon Sheesh, stop picking on databases.

God, I never thought I'd say that. I'm defending databases now. What happened to me?

Huh, not a database after all.

Braiam

16:02

@FaheemMitha mm?

Faheem Mitha

@Braiam More than here, I mean.

Braiam

@FaheemMitha well, here I have 200 answers vs. AU 1139

Faheem Mitha

@Braiam So, that's probably a yes?

Braiam

@FaheemMitha yeah, but rep wise I have higher rep/answer ration here then on AU

Faheem Mitha

@Braiam You calculated it?

what are the respective numbers?

Chap

16:21

@terdon Thanks for response, and - no problem :-)

Namrata

17:52

@terdon Do you think it is complex programming code?

strugee

18:07

Frew

11

emacs elisp apache gnome keyboard-shortcuts linux monitoring

@terdon @slm @all keep an eye on this guy.

seems like he's potentially someone trying to get reputation so he can spam us later.

you'll probably notice him if you've been in the Low Quality queue

Tim

18:28

hi, any one with Ubuntu or Debian?

know about fpm?

strugee

@Tim just ask your question.

don't wait for someone to come by and say it's OK.

Tim

it is here unix.stackexchange.com/questions/143830/…

strugee

I could be answering your question right now if I knew what it was.

@Tim please don't post questions in chat. there are a core group of regulars here who regularly check the front page.

we will see your question eventually. patience, grasshopper.

Tim

Do you know what fpm is about?

I am still not very clear

Are you running Debian or Ubuntu, by any chance? I have my deb here for 32-bit dl.dropboxusercontent.com/u/13029929/nodejs-0.6.0_i386.deb

strugee

@Tim I run Debian on a server

@Tim why don't you just use the repository packages?

Tim

18:35

Not sure why it can't install here. Either my OS is the problem, or the deb?

because I learn to use fpm

I create the deb by fpm, following the instructions perfectly.

Seth

apt-cache search fpm | grep ^fpm
fpm2 - password manager with GTK+ 2.x GUI

What fpm are we talking abut here?

Tim

github.com/jordansissel/fpm/wiki

Seth

Uhhhh, that got starred? ok then.

oh help. I smell ruby.

Tim

@Seth what is teh matter?

Is ruby bad? Should I learn it

Seth

no, I'm just not used to it.

Tim

18:40

from the fpm wiki, I am not so clear about what fpm can do

Seth

So.. why don't you just install libssl0.9.8?

Tim

I already has libssl1.0.0 1.0.1-4ubuntu5.12, and libstdc++6 4.6.3-1ubuntu5. They seem to satisfy -d "libssl0.9.8 (> 0)" and -d "libstdc++6 (>= 4.4.3)".

Seth

I read that.

strugee

@Seth I just thought it was funny

with no context

Tim

@Seth you mean to uninstall a higher version, and install a lower version of libssl?

Seth

18:42

@strugee Dunno why ¯\_(ツ)_/¯

@Tim I mean, it wants libssl0.9.8. Install that. Does it fix it? ok then. Now we need to find out why.

strugee

try creating a local repository and using APT to install your package

Seth

It's part of troubleshooting.

strugee

so you get dependency resolution

Tim

Hasn't higher version satisfied the requirement of lower version?

From my experience, whenever ./configure, if my installed version is higher than what is asked for by configure, then that is ok?

Seth

I would think usually, yes. Apparently not in this case ;)

(>> 0) is leaving me slightly confused though. Why the double >?

Tim

18:47

have you look at the example that i follow github.com/jordansissel/fpm/wiki/PackageMakeInstall

I don't know eeither

Seth

tell you what, let me fire up a VM and test it myself.

argh VMware updates

Tim

@Seth I am always wondering how people draw that cutie pie ¯_(ツ)_/¯

Seth

I just copy it ;)

Tim

how is it done by typing

Seth

ツ is a Japanese character so I can't really type it easily.

Tim

18:49

do you have a repository where you can copy whenever you like?

Seth

@Tim Yeah:

39

Q: I saw something in chat I don't understand

What are some of the common chat items? I want to understand what people who are more familiar with other chat clients are trying to say to me.

discussion chat

oh brilliant. VMware is updating and VB crashed.

Tim

is VB visual basic?

Seth

No, VirtualBox.

Braiam

@Seth 1st world problem... here would be "meh, there is a blackout right now so I can't check"

Seth

@Braiam heh ;P

That's true though.

Tim

18:51

Is VMware running on a VB, or the othe rway around

Seth

okay, so apparently that vm is borked. I need to learn to use snapshots.

@Tim Neither, I have both on my main machine.

Tim

So I created a deb file with fpm on my Ubuntu 12.04 (32-bit). Will it be able to install on Debian?

strugee

@Tim not always

Tim

on other releases of Ubuntu, e.g. 14.04

Seth

@Tim probably, not always.

strugee

18:54

@Tim no

Seth

For example, I installed the 12.04 .deb of Skype on 14.04. Works fine.

strugee

at least not consistently

Tim

what is my deb supposed to install on?

strugee

it depends on a lot of things

Seth

^^

strugee

18:55

in no particular order, things that can break:

* ABI incompatibility (aka library upgrades)

* kernel ABI mismatches, it it uses kernel modules

* interpreter mismatches, if it uses e.g. Python

Seth

Is this a 32bit or a 64 binary?

ah, 32.

strugee

* 32-bit vs 64-bit

Tim

what is the difference between "ABI" in the first and "kernel ABI" in the second?

strugee

@Tim do you know what an ABI is?

Tim

no

strugee

18:58

also:

* DBus interface mismatches, if the desktop environment has changed significantly

* init system differences, if your stuff deals with lower-level things like the session and init

Application binary interface

In computer software, an application binary interface (ABI) is the interface between two program modules, one of which is often a library or operating system, at the level of machine code. An ABI determines such details as how functions are called and in which binary format information should be passed from one program component to the next, or to the operating system in the case of a system call. Adhering to ABIs (which may or may not be officially standardized) is usually the job of the compiler, OS or library writer, but application programmers may have to deal with ABIs directly when...

Braiam

@Tim I recommend you to do a structured learning, it will be better in the long run

Tim

yes, I am doing my best to organize my learning. what is in a structured learning?

strugee

@Tim he means e.g. reading a book or something

Tim

Yes, I am now reading fpm wiki

I will some books too. :) soon

I have read some books before. :)

doing my best to structurize my knowledge and learning.

Do you think fpm will be a good tool to you as an administrator?

what do you use if you are not using fpm?

Faheem Mitha

19:14

hi @strugee

Tim

hi Faheem

terdon

@Chap Not quite sure what you're referring to but you're welcome and I'm glad there's no problem.

@Namrata Not very, just more complex than I thought at first. I'm sorry but I don't have time to do it right now, I will try as soon as I can.

Ramesh

One problem with this command is, it doesn't give the output in tab delimited format.

0

A: Merging multiple files by row

You can accomplish it fairly with the join command. join -j 1 <(sort file3) <(sort file1) > file4 | join -j 1 <(sort file4) <(sort file2)

can someone help me fix this? I tried the -t option of join.

It's not working.

slm

19:39

@Ramesh is everything tab delim. for files1-3?

Ramesh

@slm, yeah.

terdon

Just pipe through sed 's/ /\t/g'

strugee

@FaheemMitha hi!

Faheem Mitha

hi @Tim

Busy evening here?

slm

@terdon I usually use column to add tabs

Namrata

19:45

@terdon Yes I agree. Looking forward to it. :)

Ramesh

@terdon thanks. see modified answer now.

0

A: Merging multiple files by row

You can accomplish it fairly with the join command. join -j 1 <(sort file3) <(sort file1) > tmp | join -j 1 <(sort tmp) <(sort file2) | sed 's/ /\t/g' > file4 I first use join on file3 and file1 since file3 has the keys. Now, I write the output of the above command to tmp and now again do th...

@slm I tried. But unfortunately it merged the lines together.

casey

19:57

@slm @terdon how's the diamond working out now that you've had some time to get used to it?

strugee

20:12

DAMMIT

guess who just locked himself out of a remote server

Seth

I can't guess.

Ramesh

@strugee chuck Norris?

strugee

@Ramesh I wish

Seth

@Ramesh Chuck Norris doesn't access remote servers, they come to him.

Ramesh

ha ha..

We need some facts for our site like this.

366

Q: Jon Skeet Facts

I'm looking for Chuck Norris Facts style answers. In case anyone is curious, this question was inspired by Jon's own comment to this question. EDIT: If you're into cryptography, you may enjoy these facts. Now with official sanction from the powers that be!

discussion community fun

Seth

20:15

@strugee I did that to myself once, except I was lucky because it was a local server.

Patrick

20:33

Are you saying the numbers in the %idle category is just for CPU? If so, why is the overall CPU 87% idle? — Rene 3 hours ago

"If so, why is the overall CPU 87% idle?" <- I feel like this is a trick question. I want to answer "because it's 13% used"...

Braiam

20:47

Are you really asking why there is a ~1% difference between all the counters in both scenario? — Braiam 18 secs ago

A side note to this... I knew someone once who offered $500 to someone who could solve a rather complicated domain-specific question, in the comment space for when a bounty is added. Nobody seriously took them up on it. I was once stuck on a problem and my wife jokingly said, "tell them I'll make them cupcakes if they help you with the problem", so I added that to my bounty post. Within seconds, I had 5 answers and one guy happened to live across town so I did deliver a couple dozen cupcakes for the effort. — Brad 27 mins ago

come on!

slm

@casey still adjusting. The hardest part is not being able to simply be a 1 of 5 vote on VtC and other things, so I'm trying to back off, but it's hard 8-)

@strugee I can't login to the server to check 8-)

@Seth Chuck Norris is who all the servers are serving 8-)

strugee

@slm the worst part is that this is something I use to feel smug when other admins do it

slm

21:03

@strugee I've had to drive to a datacenter twice in my life b/c of this 8-)

Ramesh

Is setting java heap to 16G good?

strugee

@slm ouch

it was just cause of bloody capitalization too

Defaults Insults instead of Defaults insults

Faheem Mitha

21:33

@Braiam I want cupcakes too.

We should change the system of reward here from Imaginary Internet Points to cupcakes.

9

mikeserv

22:00

@FaheemMitha - the idea holds promise. How many cupcakes?

Faheem Mitha

@mikeserv Dunno. I'd hadn't worked out the details.

But cupcakes are tastier than IIP.

strugee

@Seth mine's local-ish

I just have to wait until I'm home

Seth

ah.

That's good.

@FaheemMitha You aren't the first to suggest that, irrc. I'm cool with cupcakes, but if I ate them would my privileges go down? Maybe if I kept the wrappers that would be proof enough that I had them once upon a time...

Faheem Mitha

22:26

@Seth I don't think you can eat your cake and have it too. Or so I have heard. But I have not tested this via the method of direct experimentation aka the scientific method.

Seth

@FaheemMitha If you ever do decide to test it ~~call~~ ping me up! I'll gladly handle the quality control on those cupcakes for you. Can't have bad cupcakes in a scientific experiment!

Faheem Mitha

@Seth Ok :-)

Seth

(good one with the can't eat & have thing ;)

Faheem Mitha

@Seth Thanks. I aim to please.

Seth

lol

Gilles

23:09

2

A: tail -f produces no output in Ubuntu live CD

If it's a case of tail not working at all, then it could be because your liveCD is using the overlayfs filesystem, which has a bug regarding notifications of modified files. You could try to move the log to another filesystem, such as /tmp if the application creating the log has an option to do ...

that's good to know

Transcript for