« first day (1288 days earlier)      last day (3658 days later) » 

slm
12:02 AM
@Ramesh Please by all means clog the site. We're fighting over like 20-40 questions a day!
 
@derobert ah, I understand
basically the idea is that there are a bunch of computers called notaries. you choose what notaries you use.
every notary will, by default, retrieve the certificate of the site you're trying to connect to. then it'll pass it back to you so that you can do a compare.
that way you know you got a legitimate certificate.
there's also a bunch of stuff done with caching and routing to ensure privacy, but that's the basic idea
 
neat, so its basically certificate pinning, but pinned across different views of the Internet, not time
 
yeah
 
so it'd defeat attacks close to you
 
yes.
the interesting part is it's designed to be extensible, such that notaries can validate with anything they want. e.g. you could have a notary that does the traditional validation, but then also have one that does DNSSEC validation.
the EFF ran a notary for a while that tied into the SSL Observatory. you can even run a notary that verifies using the legacy CA system
 
slm
12:12 AM
We should probably downvote this A on this Q
8
Q: If I connect a physical device, how can I ever know which device file belongs to it?

rightfoldSay I plug in several USB drives which don't get automatically mounted. How can I find out which device file belongs to which physical device, so I can mount it for example? I'm running Mac OS X but I rather like an answer that works on all (or at least the most popular) Unix systems. I had this...

8
A: If I connect a physical device, how can I ever know which device file belongs to it?

rozcietrzewiaczUsing udev: You can get useful information querying udev (on systems that use it - almost all desktop-type Linuxes for sure). For instance, if you want to know which attached drive is associated with /dev/sdb, you can use: udevadm info --query=property --name=sdb It will show you a list of pr...

The Q is about OSX and the highest A is showing how to do this in Linux via UDEV. If I'm not mistaken OSX doesn't do UDEV.
 
"I'm running Mac OS X but I rather like an answer that works on all (or at least the most popular) Unix systems. I had this problem with Linux in the past." ... from the question.
 
is the talk. I've skipped to the part where he actually talks about Convergence, it's not actually all that long
@slm not in the slightest. not even close
@derobert impossible
no way in hell you can make something like that even vaguely portable
 
slm
Shouldn't we either delete that A or DV it then?
 
12:28 AM
are we talking about the top-voted A?
 
slm
@derobert - so should we leave it then?
yeah
isn't that incorrect? I'm not sure what we should do here. It's clearly wrong for the Q
 
well, the question is from 2011
 
slm
I left a comment to the A'er suggesting a new Q to provide a home for that A, since it's good.
I know but it's come up several times for me when I'm searching for a linux way to do this.
That seems convoluted to me since it's an OSX Q
 
certaintly if the answerer moves it then we should get rid of the answer
otherwise I'm not in favor of removing information
 
or just edit the question and steal it...
 
12:30 AM
worst comes to worst, we make it community wiki
 
@slm yeah, but given that none of the answers say anything useful about OSX, it would be more productive to edit the question to be about Linux
 
brb
@Gilles exactly, just change the question
 
anyway, brb
 
is the only thing we would have to worry about but that answer is crap
so I say go for it
 
12:31 AM
@strugee it has 0 votes, very little content, and is it even true?
 
@Gilles it's the OS X equivalent of lsblk
 
slm
@Gilles I'm fine with that as well. Shall I do that then?
 
@slm go for it!
 
@Gilles but like I said, IMO it's crap
 
slm
@Gilles Do we remove the OSX from it or lace in OSX and Linux together?
 
12:33 AM
@slm no, make it just about linux. This is completely different from unix variant to unix variant, there's no sense in combining them all in one question.
 
slm
@Gilles - OK, I just didn't want to invalidate all the OSX answers either, but I think there's really only one now.
 
12:53 AM
@slm, this user is asking the same question 3rd time.
-2
Q: How to compile openSSL

thisisjndI'm using Cygwin on Window, trying to use THC Hydra. When I try to initiate a command, I get an error: [ERROR] Compiled without SSL support, module not available I cant find any advice on how to compile/append the openSSL dependency

 
Its an old, not very good answer from someone who showed up on the site once, and the answer has received no votes... It can just be deleted.
Flag it (reason: other) and explain to the mod what happened, you can link to the chat transcript
 
@derobert, thanks. I have flagged it with appropriate reason
 
Someone needs to edit the question too...
 
@derobert, the same user asked the exact question yesterday too.
It was closed yesterday and he has asked the exact question with slight change in words today.
 
@Ramesh I was talking about unix.stackexchange.com/questions/18064/… which is from 2011... Not sure which one you're talking about
 
12:59 AM
@derobert isn't @slm editing?
 
Possibly. I know I'm not.
 
slm
I'm going to
 
Is Cygwin even on topic?
 
slm
was trying to help the yoga laptop guy first
 
yes
 
12:59 AM
@derobert, sorry about that. I was talking about some other question.
 
slm
but will edit it tonight unless anyone else wants to
8-)
 
@Ramesh and he's cross-posted
I think we need to suspend him
I'm trying to vote to close but the bloody dialog won't load
 
@strugee yeah, he has not made any effort from yesterday.
 
and he's an idiot
 
I mean he has posted the same question with the same content, just the heading different.
 
slm
1:02 AM
@Ramesh - which Q are you talking about? The SSL one, right?
 
@slm, yeah.
 
slm
OK, voted to close it and flagged it as low quality, just for spite
 
thanks
 
slm
I'd already DV'd it
 
stupid dialog
 
slm
1:03 AM
Gilles was trying to help him too.
 
@slm, as strugee suggested you probably want to look into suspending this user.
 
@Ramesh we have to ask a mod
which @slm is not (shocker!)
@MichaelMrozek can you suspend this guy?
 
@strugee, that is a shocker for me. Really?
 
slm
We'd have to ask a mod to do that. I see 2 Q's that he's asked about this.
Seems extreme
 
@Ramesh yep. neither is Gilles
 
1:04 AM
Two questions, one self-deleted
 
slm
yes gilles and I are just here all the time, not mods
 
@slm he's also cross-posted and shows no indication of wanting to learn the rules
 
But yeah, this is a case of cluelessness, not malice... doesn't seem to warrant a suspension. At least if he stops
 
he's posted a duplicate question three times
 
@slm Is it because, you guys opted out of the moderators post?
 
slm
1:05 AM
yeah I was about to say the cross posting is more a reason to close IMO esp. if he's not following the rules
 
you can tell because two are in the list of questions he's posted and he has the peer pressure badge
 
slm
yeah 3 times would seem to be enough to justify a timeout
 
@Ramesh no
@slm's been here for less than a year
 
Not sure when the last U&L mod election was...
 
Gilles presumably just wasn't interested when they were having moderator elections
 
1:06 AM
2011, apparently
 
I asked a Q about elections once
 
slm
@Ramesh Nope, it comes up from time to time. We already have enough mods, it's better for us to A the Q's.
 
@derobert site graduation
 
slm
we have plenty of regulars and power users to manage the day to day
 
UL is easy of moderating
 
1:08 AM
oh ok. I like this site more than stackoverflow for the reason it is easy going on its users :)
 
there are enough 20kers and 3kers to deal with most of the problems
 
slm
I've had an account for 3 years and been here for ~15 months
 
@slm I stand corrected
 
slm
I actually started actually trying in Nov. of 2012
 
@slm, 75k points in 15 months?
 
1:09 AM
unix.stackexchange.com/questions/125289/… <--- there is one of the deleted ones, need 10K rep to see of course
 
That's awesome :)
 
slm
yes It started on a dare
 
slm
my co-worker dared me to get 3k on the original 3 sites and I said pff, and I'll do it on 3 + this site
 
@slm and it just spiraled?
 
1:10 AM
I know the approach is not the best, but I think I saw something about sending the same commands for several ssh questions
 
slm
couldn't stop on this site. They offered to buy me lunches everytime I hit plateaus of rep.
 
@Braiam not with GNOME Terminal, I don't think
 
@slm, great going :) So now are you trying to reach Giles points?
 
slm
I'm trying to get 100k
 
honestly it sounds like his problem is that he has a server farm and is in desparate need of Puppet
 
1:11 AM
@strugee I didn't want to be both a mod and the user who posted all the answers
 
@Gilles ah. why?
 
slm
I'm following Gilles lead on this as well 8-)
 
because too often I'd end up with a conflict of interest, e.g. deleting an answer to a question where I'd also posted all the answers, or closing a question as a duplicate of one that I'd answered
 
slm
we'd have to change the url to gilles.stackexchange.com
3
 
Thanks, this helped me a lot. — Space Ghost 1 min ago
 
1:12 AM
I would vote up for giles.se.com :)
 
but then I ended up also being the highest-rep user and a mod on French Language
not to the same extent I used to be on Unix & Linux, however
 
@Gilles ah, that makes sense
 
slm
Gilles is at no risk of me ever catching him either. I've calculated the rate I'd have to run at to gain him. It would take ~3-4 years
 
assuming there's no variance
 
slm
right
 
1:14 AM
@slm You just have to get @Gilles to ask a bunch of questions, and offer bounties.
 
slm
one thing I've noticed is we've grown our user base of late
which is good for the site, but it's been more crowded at the top now, which is good for the site, but makes it harder to catch him
 
really?
 
@slm, one thing I liked here in this site is everyone is friendly :)
 
I'm spoiled on this site
 
slm
yes that's why I never finished my dare on SF.
 
1:15 AM
this was my first SE site that I really participated in and now I can't stand other communities
 
slm
me neither, everyone is about trying to get the best solutions in place and everyone is very encouraging of others here
I've noticed several new faces of late that have been writing up good answers. Grame's one and Patrick has been much more active
there are many others
 
yeah
and mikeserv, despite his coarse personality, is good at what he does
 
glenjackman is also very good. But I have never seen him in chat window.
 
I'm good at crying when people run mdadm --create :-)
 
I haven't noticed any of his stuff
 
@derobert mkfs.btrfs. problem solved.
 
Its actually been a while since I've seen --create as a raid "recovery" method on the site.
 
runs mdadm --create in a VM
meh, virtualbox is being picky with me... maybe a reboot will help
 
@Braiam hah, so much for your mdadm attempts!
(If you're not familiar with mdraid: --create is perfectly fine when used as intended, i.e., to create a new, blank array. It should almost never be used to recover data from an existing array.)
Anyway, time to get some food. Then back to working on this webapp...
 
@derobert I've been "working" on github.com/strugee/prism-break for about an hour and a half now...
actually just been on here :)
 
1:28 AM
Hah, you can see I did a little! I committed something.
Anyway, BBL
 
\o/
yeah, so far I have one commit that changes a string.
("truecrypt-audit" branch)
sent a pull request but haven't gotten much else done
anyway, I gotta go eat. then I'm off to a 5-year-anniversary party for a local hackerspace.
see you all!
 
slm
2:16 AM
@strugee yes I think he's warmed up to us, he's been a nice addition, has pretty good knowledge of boot and file redirection
@Ramesh yeah jackman has excellent awk/sed/shell skills. He pops in/out of here, AU and SO.
@strugee I think often times ppl come in to SE sites expecting to be attacked, so when they get to this SE site they're on the defensive. We like to keep things running smoothly here w/o a lot of drama
 
 
2 hours later…
slm
4:46 AM
@strugee - he's back
0
Q: How do you compile dependencies manually? Cygwin

thisisjndI installed cygwin on windows and the setup.exe doesnt include all the dependencies. in the folder however, there are many dependencies that arent included in the setup.exe. when i type in a dependency in the search bar at the top left corner, it doesnt appear, but it IS in the cygwin folder. ...

 
5:09 AM
@slm he's trying to convince you cygwin is off-topic
 
slm
@derobert - who's he in that sentence?
 
slm
I'd normally just ignore this type of Q.
 
5:33 AM
@slm that seems like it would have been a good plan
@derobert, do you have an email or somewhere we could chat? please? ive been having trouble getting cygwin to work for days now and im going nuts — thisisjnd 42 secs ago
Guess I'll create a chat.SE room. I think I can then add him to the access list, to let him in despite the lack of rep.
UGH. Apparently only a mod can do that. I get an error saying he needs at least 20 rep :-(
 
 
2 hours later…
7:15 AM
@derobert i can upvote whoever if you want. how much rep does he currently have?
 
 
1 hour later…
8:21 AM
@Braiam Well, that is polite.
 
8:34 AM
@FaheemMitha negative
not actually but if it was possible he would
@slm I think that's true. the Big Three have a reputation that precedes them
 
8:53 AM
@strugee the big three?
 
 
4 hours later…
12:44 PM
Can we reopen this?
2
Q: Strange error (?) when I run `man` command from a folder that no longer exists

Radu RădeanuLet's have a look at the followings: radu@Radu:~$ mkdir test radu@Radu:~$ cd test radu@Radu:~/test$ rmdir ~/test radu@Radu:~/test$ man ls man: can't change directory to '': No such file or directory Normally, I would say that the last line from the previous output from my terminal is an error. ...

It's not a dupe, it looks like a bug in man 2.6.7.1, the OP is asking why the man command specifically fails when run from a deleted directory. Not why other commands might.
 
slm
@terdon - voted. Just tried it and man does work in a deleted dir for me.
man 2.6.3
 
@slm me too. The OP is using man 2.6.7.1 so it looks like it's specific to that version. He's not a newbie, I'm confident that what he describes is actually real.
 
slm
@terdon - yes I recognize him from AU
 
12:59 PM
This sounds like some kind of compositing problem but I don't know what to suggest:
1
Q: Mouse is flickering in new debian install

ArmanSchwarzI initially thought this might be a display driver issue but after installing the proprietary ATI drivers and now the free ones I am still getting it. My mouse disappears when I scroll or write sometimes (but not always). It also flickers when I move, especially when I am in the system settings ...

 
@terdon I don't have the bug in Debian
 
@Braiam With the same man version?
 
1:17 PM
 
@Braiam Thanks, so yes, it's a bug in his particular version of man.
 
@terdon no, is a bug in that particular version of man in Ubuntu
I presume they messed up a patch
 
slm
1:42 PM
@Braiam @terdon - any ideas on what's up with this one?
@slm I reinstalled the proprietary drivers (wiki.debian.org/ATIProprietary) and restarted. Running gksudo amdcccle as per that answer produces this output: (gksudo:4955): GConf-WARNING **: Client failed to connect to the D-BUS daemon: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. GConf Error: No D-BUS daemon running and nothing opens.. — ArmanSchwarz 6 mins ago
 
@slm dbus is not running
 
slm
$ service dbus start
 
IMO, is the DM that starts dbus
 
@slm lol :)
 
slm
@Braiam agreed
 
1:47 PM
@terdon I finally wrote a response to your letter.
I mean, sent a response to your letter.
 
slm
don't know what's up with this Q but the OP just said adding to xorg.conf fixes it, I'm inclined to let it lie
 
@terdon as I said in the letter, I see your point with the HMM, but I'm not sure about what to change, so perhaps you can suggest something?
 
@FaheemMitha I just replied. I think the changes you made are fine. My point about HMMS is that they can take the surrounding nucleotides into account but only in a limited way and for a limited number. So saying that they can't at all is problematic, while saying that they're not very good at it is fine.
 
@terdon ok. If you don't think other changes are necessary, that is Ok. Do you agree the Eddy quote is useful here?
 
@FaheemMitha Sure, it highlights the limitations of HMM based approaches. And he should know.
 
1:54 PM
@terdon got your email. what does N mean there?
Ok, you wrote N-dimensional HMMs. but I still don't know what that means
is that the state space?
 
@FaheemMitha No, I should have written order not dimension
 
@terdon btw, I've read about HMMs but never used them, so no, I wouldn't say I understand them.
@terdon still not sure what that means. can you elaborate?
 
OK, for example a 4th order HMM will look at 3 nt before the one it's checking
So, an N-order HMM will look at N-1 nt before the current one. This just doesn't scale well to large values of N
 
@terdon ah, right. sorry, I didn't remember that usage of the terminology. thanks.
what is nt?
next to, perhaps
no, nucleotides
ok, off to dinner. take care, guys
 
@FaheemMitha Sorry, biological jargon, nt == nucleotide
Also, here's an example of how a 5th order HMM can be sued:
> All exon and intron sequences were extracted from the
MR multiexon data set. A Markov model of order 5 was
estimated to model both exon and intron sequences,
that is, we estimated the probability distribution of
each nucleotide given the pentanucleotide preceding it
in exon and intron sequences.
 
2:03 PM
 
2:27 PM
Finally got a handle on null-bytes. Kinda proud of it...unix.stackexchange.com/a/125549/52934
 
Does anyone here deal with data center?
 
2:44 PM
@terdon Yes, I see. I hadn't previously considered this possibility. As I said earlier,I have not used those things. I think of HMM as using just the adjacent nucleotide. I'll think about it, but I don't think it alters my basic point.
Which is really an issue of locality.
@terdon if you are ok with it, i'd like to add you to the acknowledgements. If that is ok, i'll send you an email, since, as you know, written permission is required.
@Braiam It seems the poster is not aware of multiarch.
 
3:11 PM
is there anywhere on meta where people have listed possible blog ideas?
 
 
3 hours later…
5:58 PM
@FaheemMitha are you out there?
 
@MERM I'm here.
 
I tried installing apt-show-versions by manually copying it over from a new machine, but it didn't work. I also tried dl the package and installing it via perl Makefile.PL and that didn't work either
 
@MERM ok. i think my last recommendation was to fix apt. do you remember what it was? one sec, let me pull up the question
 
What I'd like to ultimately do is get rid of all of the unstable stuff and replace it by stable versions. Can I do that package by package, or wholeslae?
 
actually, why don't you just post the link?
 
6:01 PM
aptitude is better at resolving dependencies — Kiwy yesterday
 
@MERM You'll probably have to do it package by package, but it depends on the situation. you said you had a bunch of machines - are they all buggered in the same way, or don't you know.
 
the whole problem is that apt is messed up
 
@MERM that's a matter of opinion.
 
yes all of the machines are similarly messed up
 
@MERM can you give me the link?
 
6:02 PM
3
Q: apt python error prevents me from resolving dependency problems for apt-get install

MERMOn Debian 6 servers dependency issues are preventing me from installing nis (or any package). When I try to resolve that issue I get a python error. How do I get past this issue and re-gain the ability to install packages? I have googled and found a few people reporting the same error messag...

 
@MERM but with variations, i assume. i mean, they aren't exactly the same install, are they?
 
no that would be too easy
 
ok, my last recommendation was
yesterday, by Faheem Mitha
run apt-cache policy apt libapt-pkg-perl libglib2.0-dev libglib2.0-0
let's do that then.
 
I don't know why they chose to mix unstable an testing instead of just using stable sources, but I don't think it was a consious decision
ok
 
@MERM i think someone screwed up. it is actually a very common error.
 
6:04 PM
# apt-cache policy apt libapt-pkg-perl libglib2.0-dev libglib2.0-0
libglib2.0-0:
Installed: 2.40.0-2
Candidate: 2.40.0-2
Version table:
*** 2.40.0-2 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
100 /var/lib/dpkg/status
libglib2.0-dev:
Installed: (none)
Candidate: 2.40.0-2
Version table:
2.40.0-2 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
apt:
Installed: 0.8.10.3+squeeze1
Candidate: 0.8.10.3+squeeze1
Version table:
1.0.1 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
 
@MERM you've got a bit of a mess here. tell me the apt-cache policy libapt-pkg4.12
 
# apt-cache policy libapt-pkg4.12
libapt-pkg4.12:
Installed: (none)
Candidate: 1.0.1
Version table:
1.0.1 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
 
@MERM ok, before you do anything else, you need to set your preferences to something sensible. add the following to /etc/apt/preferences.
Package: *
Pin: release a=testing
Pin-Priority: 50

Package: *
Pin: release a=unstable
Pin-Priority: 50
Unless you have it there already.
the testing one is not currently necessary, but doesn't hurt to have it
 
Yes, we put it there the other day. Later I can push this ot to all of the mahines via ssh
I am working with one machine currently, to find a procedure I can use
 
@MERM ok, good
tell me what apt-get -f install is currently showing
 
6:10 PM
apt-get -f install
Reading package lists... Done
Building dependency tree
Reading state information... Done
0 upgraded, 0 newly installed, 0 to remove and 73 not upgraded.
 
Ok. You seem to somehow have the apt from squeeze, which is very weird, and a bunch of stuff from unstable which shouldn't be there.
just to confirm, what does dpkg -l apt give?
 
# cat sources.list
#deb http://ftp.us.debian.org/debian/ stable main
#deb-src http://ftp.us.debian.org/debian/ stable main

deb http://ftp.us.debian.org/debian/ unstable main
deb-src http://ftp.us.debian.org/debian/ unstable main

# Line commented out by installer because it failed to verify:
deb http://security.debian.org/ stable/updates main
# Line commented out by installer because it failed to verify:
deb-src http://security.debian.org/ stable/updates main
# dpkg -l apt
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Description
+++-==================-==================-====================================================
ii apt 0.8.10.3+squeeze1 Advanced front-end for dpkg
 
@MERM yes, apt from squeeze. ok, try apt-get install apt.
 
# apt-get install apt
Reading package lists... Done
Building dependency tree
Reading state information... Done
apt is already the newest version.
0 upgraded, 0 newly installed, 0 to remove and 73 not upgraded.
 
@MERM ok, weird. apt-cache policy apt
 
6:14 PM
@FaheemMitha # apt-cache policy apt
apt:
Installed: 0.8.10.3+squeeze1
Candidate: 0.8.10.3+squeeze1
Version table:
1.0.1 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
*** 0.8.10.3+squeeze1 0
100 /var/lib/dpkg/status
 
@MERM ok, that's strange. have you run apt-get update recently? in any case, do it now.
Ok, never mind, why do you have stable sources commented out?
 
Like I said, I didn't set this up.
 
actually, can i take a 5 min break?
 
sure see you in a few
 
@MERM go ahead and uncomment those, then run update.
 
6:16 PM
running update now and it is still going, 172 lines
 
also, change your security lines to
deb http://security.debian.org/ wheezy/updates main contrib non-free
deb-src http://security.debian.org/ wheezy/updates main contrib non-free
 
ok
 
@MERM interrupt the update. and change those first, then run update. back in 5
 
thanks
 
@MERM ok, back now
 
6:24 PM
ok
 
What does apt-cache policy apt give now?
 
I was trying some things on another machine, put the preferences and changes the sources, did apt-get update and apt-get -f install
# apt-cache policy apt
apt:
Installed: 0.8.10.3+squeeze1
Candidate: 0.9.7.9+deb7u1
Version table:
1.0.1 0
50 http://ftp.us.debian.org/debian/ unstable/main amd64 Packages
0.9.7.9+deb7u1 0
500 http://ftp.us.debian.org/debian/ stable/main amd64 Packages
*** 0.8.10.3+squeeze1 0
100 /var/lib/dpkg/status
this is on the original machine we were working on
 
@MERM Ok, try apt-get -f install and if that is ok, then apt-get install apt.
 
it is the same on the other machine too
 
@MERM what is?
 
6:27 PM
apt-cache policy apt
# apt-get install apt
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following extra packages will be installed:
apt-utils aptitude aptitude-common libapt-inst1.5 libapt-pkg4.12 libboost-iostreams1.49.0
libept1.4.12 zlib1g zlib1g-dev
Suggested packages:
apt-doc python-apt debtags
Recommended packages:
aptitude-doc-en aptitude-doc apt-xapian-index libparse-debianchangelog-perl
The following packages will be REMOVED:
libept1
The following NEW packages will be installed:
 
@MERM ok. maybe you should take notes along the way. or use version control
@MERM ok, that looks good
 
I have history and notres
 
does it complete normally?
 
How do I tell if everything is good?
apt-get install apt did complete successfully
 
@MERM it just exits, no errors given
ok, run dpkg -l | grep libapt
 
6:29 PM
see full text
# dpkg -l | grep libapt
ii libapt-inst1.5 0.9.7.9+deb7u1 deb package format runtime library
ii libapt-pkg4.12 0.9.7.9+deb7u1 package managment runtime library
 
@MERM that's everything?
 
yes
 
if so, try installing apt-show-versions
 
apt-show-versions installes now
# apt-show-versions
acpi:amd64/stable 1.5-2 upgradeable to 1.6-1
acpi-support-base:all/stable 0.137-5 upgradeable to 0.140-5
acpid:amd64/stable 1:2.0.7-1squeeze4 upgradeable to 1:2.0.16-1+deb7u1
adduser:all/stable 3.112+nmu2 upgradeable to 3.113+nmu3
apache2-mpm-prefork:amd64/stable 2.2.16-6+squeeze10 upgradeable to 2.2.22-13+deb7u1
apache2-utils:amd64/stable 2.2.16-6+squeeze10 upgradeable to 2.2.22-13+deb7u1
apache2.2-bin:amd64/stable 2.2.16-6+squeeze10 upgradeable to 2.2.22-13+deb7u1
apache2.2-common:amd64/stable 2.2.16-6+squeeze10 upgradeable to 2.2.22-13+deb7u1
 
apt-show-versions | grep unstable
 
6:31 PM
# apt-show-versions | grep unstable
console-setup:all/unstable *manually* upgradeable from 1.92 to 1.107
console-setup-linux:all/unstable *manually* upgradeable from 1.92 to 1.107
cpp-4.4:amd64/unstable *manually* upgradeable from 4.4.7-3 to 4.4.7-8
fontconfig-config:all/unstable 2.11.0-5 uptodate
fonts-dejavu-core:all/unstable 2.34-1 uptodate
g++-4.4:amd64/unstable *manually* upgradeable from 4.4.7-3 to 4.4.7-8
gcc-4.4:amd64/unstable *manually* upgradeable from 4.4.7-3 to 4.4.7-8
gcc-4.4-base:amd64/unstable *manually* upgradeable from 4.4.7-3 to 4.4.7-8
 
@MERM hmm, your libc6 is the unstable one. that is not good.
 
how do I upgrade it? can I just tell aptitude to upgrade everything?
 
@MERM ok, you want to bring your server to stable, right?
 
yes
 
@MERM ok, then you need to get rid of as much of the unstable stuff as you can. you can't get rid of libc6, because that will break your machine, though.
 
6:34 PM
what is the procedure?
 
So, try doing apt-get purge pkg1 pkg2 for all the not-essential packages listed as on unstable
I think everything except libc6 can be removed, but you may lose X. are you on a console at the moment?
 
I am in via ssh
 
are you running gnome, kde, or some other window manager, or are you on a remote ssh connection?
@MERM ok
 
The server is a few thousand of miles away
I am just ssh in from my local machine
I have a local window manager
 
@MERM yikes. so if you break the machine, what happens?
 
6:36 PM
but not using X remotely
client gets mad
there are non-skilled local hands
 
@MERM ah. well, you don't have a locally messed up machine you can practice on?
 
did I mention that these machines are live, providing service?
well, there are a few in NY which is only 50 miles away, as opposed to the ones in mexico and columbia which are thousands
 
@MERM no, you didn't.
unfortunately, you're going to have to take down some services, at least temporarily.
anyway, most of the packages that are at unstable can go without problems, but at least libc6 may be problematic.
 
I can take some short down time, as long as it is restorable
I can work on one machine and see if I can fix it, there are some spares I can move the live customers to in case of a problem
 
@MERM try taking out the python and gcc packages listed. look carefully at the other packages that are listed, to make sure it it not taking out anything you don't want.
 
6:40 PM
if I can figure out a procedure, I'll do them one by one
so I want to purge them, and then re-install them?
 
apt-get purge python-minimal libpython-stdlib libpython2.7-minimal
to start with
@MERM yes, at stable versions
just check and see what else it is taking out
 
# apt-get purge python-minimal libpython-stdlib libpython2.7-minimal
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following packages were automatically installed and are no longer required:
python2.6 python2.6-minimal
Use 'apt-get autoremove' to remove them.
The following packages will be REMOVED:
libpython-stdlib* libpython2.7-minimal* libpython2.7-stdlib* mercurial* mercurial-common*
python* python-minimal* python2.7* python2.7-minimal*
0 upgraded, 0 newly installed, 9 to remove and 332 not upgraded.
 
as long as it doesn't take out ssh you are probably ok. if it is going to take out really important packages, it will warn you
 
but then the install failed
# apt-get install python-minimal libpython-stdlib libpython2.7-minimal
Reading package lists... Done
Building dependency tree
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 
@MERM no, first get rid of all the unstable that you can
 
6:42 PM
ok
 
first, see if there are any other python packages from that unstable list
 
ENC01_root 113 # apt-show-versions | grep unstable | grep python
ENC01_root 114 #
 
@MERM ok
@MERM ok, then do it for the other stuff in that list, leave out libc6 for now
 
ok
 
if it tries to take out ssh or anything else that looks important, stop
i'm not sure what else it might try taking out
 
6:44 PM
roger that!
I can remove the gcc stuff? If I do that I can't compile as part of installs
 
@MERM you can reinstall it later
 
ok
 
and yes, you can remove all the stuff that doesn't break ssh, i think, or that the system doesn't actually warn you against removing
one thought - open a couple more ssh connections to that machine in case something happens to the one you have
 
already have two, I'll add one
 
afk for another 5
 
6:52 PM
ok
 

« first day (1288 days earlier)      last day (3658 days later) »