1140048195 Q * prae Quit: Pwet
1140048698 M * michal` Bertl: do you know what's the difference between bug and feature ?
1140048711 M * michal` i would say it is purely cosmetic ;p
1140048713 M * michal` http://pics.pkilian.com/bug.jpg
1140048823 J * grant_ mep@p5091AC0E.dip0.t-ipconnect.de
1140049240 Q * grant Ping timeout: 480 seconds
1140049574 M * derjohn micah, lol
1140049589 M * derjohn michal`, lol
1140049593 M * derjohn micah, sry :)
1140049651 M * cehteh haha
1140050125 Q * michal` Ping timeout: 480 seconds
1140050605 Q * SNy Ping timeout: 480 seconds
1140050760 A * SamVilain will begin his tut in 15 minutes...
1140050799 A * SamVilain fiddles in frustration with his laptop and projector, cursing every now and then
1140050800 M * Bertl excellent!
1140051056 M * Bertl http://vserver.13thfloor.at/Stuff/MANDRAKE/linus-git-1.2.0-0.1.20060mdk.src.rpm
1140051059 M * Bertl http://vserver.13thfloor.at/Stuff/MANDRAKE/stgit-0.8.1-0.1.20060mdk.src.rpm
1140051073 M * Bertl (for the folks who didn't prepare that yet :)
1140051099 Q * Doener Quit: Leaving
1140051147 M * derjohn SamVilain, what the tut?
1140051166 J * SNy 5c27df7953@bmx-chemnitz.de
1140051168 M * SamVilain "Managing the VServer patch with Git and stgit"
1140051223 M * derjohn SamVilain, some kind of podcast?
1140051250 M * SamVilain nono, just IRC.  Don't worry, I haven't tried this before either
1140051313 A * SamVilain just notices qgit has StGIT support
1140051324 M * derjohn well then I tune in this channel ... SamVilain we be back after this short break ;)
1140051562 M * SamVilain ok then
1140051566 M * SamVilain Bertl, the topic, please
1140051578 M * Bertl hmm?
1140051621 M * SamVilain oh, just thinking it would be good in case people join :)
1140051629 M * Bertl well, change it :)
1140051663 T * SamVilain http://linux-vserver.org/ | Tutorial "Managing the VServer patch with Git and stgit" now in session
1140051676 M * SamVilain rightio, then
1140051681 M * SamVilain    
1140051684 A * SamVilain turns on his mic
1140051689 M * SamVilain ah, there we go
1140051700 M * Bertl *test* *test* *1* *2* *3*
1140051712 M * mnemoc o_O
1140051723 M * SamVilain ok.  We're here to talk about git.  Firstly, I'll talk a little about the git filesystem, then we'll get onto the worked demos
1140051727 M * derjohn *monitor up left speaker*
1140051755 M * SamVilain ok, as we all know, there have been a great many new VCS systems out there in the last few years
1140051767 M * SamVilain But what does GIT do differently?
1140051794 M * SamVilain Well, the key difference between git and virtually everything else (except monotone, which inspired git) is that Git is a _content based_, not _revision based_, VCS
1140051834 M * SamVilain So, for instance, with GIT you do not have file revision numbers like CVS, or repository revisions like Subversion, but long SHA1 content revisions
1140051882 M * SamVilain each file or object's identity is the SHA1 hash of its contents,
1140051909 M * SamVilain and the directory's is the SHA1 hash of the directory listing, including the SHA1 hashes of each file
1140051919 J * michal` ~michal@www.rsbac.org
1140051936 M * SamVilain so, if you have two commits with the same "tree ID", their content can be determined to be the same
1140051955 J * nem ~nem@leibniz.catalyst.net.nz
1140051957 M * SamVilain so, the history doesn't matter.
1140052051 M * Bertl sidenote: hashes are supposed to identify files in an unique manner
1140052052 M * SamVilain Let's see this happening in a real repository, I'm going to point at a couple of gitweb URLs
1140052073 M * SamVilain ok, the first is http://vserver.utsl.gen.nz/gitweb/?p=vserver.git;a=shortlog;h=2.6.14.y
1140052103 M * SamVilain If you select the "commit" link of the topmost commit, "Official chrisw 2.6.14.4 patch", then you can see a page with the commit information
1140052155 M * SamVilain There are three IDs there; a commit ID that uniquely identifies this commit (and its history), a tree ID that uniquely identifies the content at that point, and a parent ID that identifies the last commit
1140052178 M * SamVilain If you now look at chrisw's tree;
1140052181 M * SamVilain http://www.kernel.org/git/?p=linux/kernel/git/chrisw/linux-2.6.14.y.git;a=shortlog
1140052205 J * Smutje_ ~Smutje@xdsl-87-78-42-227.netcologne.de
1140052228 M * SamVilain then scroll down to the commit labeled "Linux 2.6.14.4", by Greg Kroah-Hartman, and select the "commit" link again,
1140052259 M * SamVilain You'll notice that the "tree" IDs of that commit and the head commit on the 2.6.14.y branch in my (utsl) repository are identical
1140052315 Q * Smutje Ping timeout: 480 seconds
1140052315 N * Smutje_ Smutje
1140052322 M * SamVilain however, the "parent" link is different, because it has a different history; I simply applied the official patch and committed that to get the version in my repository
1140052340 M * SamVilain There are some other important differences between git and the rest.
1140052348 M * SamVilain For a start, git is blindingly fast
1140052376 M * SamVilain Linus happens to know a thing or two about writing a good FS ;)
1140052423 M * SamVilain Secondly, it is designed to be extremely modular
1140052445 M * SamVilain The core tools, commands starting with git-, are called the "plumbing"
1140052470 M * SamVilain Programs which wrap the git- commands to provide a nicer user interface, are called "porcelain"
1140052486 M * SamVilain You may also hear reference to "spillage",
1140052510 M * SamVilain This is when you lose something.  But personally I haven't encountered that yet ;)
1140052541 M * SamVilain So, StGIT is an example of a git porcelain
1140052600 M * SamVilain Another important porcelain is Cogito, which tries to be more like CVS or Subversion in its interface.  But I won't be talking about that, just git-core and stgit
1140052622 M * SamVilain ok, so anyway, along with the examples.
1140052666 M * SamVilain You should have brought with you a clone of Linus' git tree,
1140052684 M * SamVilain as produced by:  git-clone rsync://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
1140052716 M * SamVilain This will have made a directory with "linux-2.6", and inside you will find a copy of the head of Linus' 2.6 tree
1140052727 M * SamVilain s/with/called/
1140052754 M * SamVilain you can see the last commit with "git-whatchanged" from inside the tree
1140052785 M * SamVilain ok, so the first thing we want to do is update this to the latest, freshest code from Linus
1140052818 M * SamVilain we will use a command called "git-pull" to do this, but first, we have to tell the git client about a better protocol than rsync for syncing git repositories
1140052844 M * SamVilain In your .git/remotes/, there is a file called "origin"
1140052854 M * Bertl Q: why is rsync bad for git?
1140052908 M * SamVilain when you "pack" a git repository, it will make a "pack file" out of the revisions present.  the filename of this, as well as the contents, change markedly every time it is packed.
1140052930 M * SamVilain There is a little bit of explaining to do there,
1140052956 M * SamVilain if you look in .git/objects/??/, you will see a whole load of files with just an SHA1 sum as their name
1140053002 M * SamVilain These files are actually revisions of the files (the complete content, gzip'ed, minus the 20 byte gzip header)
1140053029 M * SamVilain Early versions of git just kept masses of these files, without regard to worrying about disk space efficiency
1140053075 M * SamVilain However after a short while they made a "delta format", a pack file that can hold revisions in a space efficient (but slower to access) way
1140053116 M * SamVilain So, the .git/remotes/origin has two lines.
1140053135 M * SamVilain The first is the URL, obviously.  Edit the file, and change the "rsync" part to "git"
1140053155 M * SamVilain The second line controls what to sync.
1140053181 M * SamVilain What "master:origin" says, is that we want to pull revisions from the upstream branch called "master", and put them in a local branch called "origin"
1140053200 M * SamVilain when we are pulling from this remote source called "origin"
1140053233 M * SamVilain so, with the git:// change there, you should be able to use "git-pull", from your linux-2.6 directory.
1140053271 M * SamVilain You should get something like "Already up-to-date", or a progress report of the update, things about fast forwarding, etc
1140053319 M * SamVilain So, that's probably not very interesting if you've already got the most recent version.  So, instead, let's get someone else's tree, like chrisw
1140053358 M * SamVilain so, create a file in your .git/remotes/, called "chrisw-2.6.15.y"
1140053365 M * SamVilain add these two lines to it:
1140053370 M * SamVilain URL: git://kernel.org/pub/scm/linux/kernel/git/chrisw/linux-2.6.15.y.git
1140053370 M * SamVilain Pull: master:2.6.15.y
1140053411 M * SamVilain now, all we have to do is make the local branch "2.6.15.y", for the changes from that tree to go.
1140053429 M * SamVilain "git-branch" with no arguments shows you what branches you have
1140053438 M * SamVilain or, with stgit, the command is "stg branch -l"
1140053455 M * SamVilain You should see two branches initially, "master" and "origin"
1140053516 M * Bertl Q: what does the asterisk mean?
1140053522 M * SamVilain we are going to branch on the tag "v2.6.15"
1140053530 M * SamVilain which asterisk?
1140053546 M * SamVilain oh, in git-branch output
1140053555 M * SamVilain it means the branch that you have currently checked out
1140053576 M * SamVilain you can switch to another branch with "git-checkout somebranch" or "stg branch somebranch"
1140053600 M * SamVilain you can create a new branch with git-branch, git-checkout or stg branch -c
1140053620 M * SamVilain so, we can use:
1140053644 M * SamVilain git-checkout -b 2.6.15.y v2.6.15
1140053666 M * Skram dYo.
1140053690 M * SamVilain Actually, this is one thing to note.  With other VCS systems you can usually check out an old revision without creating a branch.  However git enforces this so when you commit, there is somewhere for it to go.
1140053698 M * Bertl Q: what was the v2.6.15 for?
1140053721 M * SamVilain if you look in .git/refs/tags, there is a file called "v2.6.15"
1140053738 M * SamVilain contents dab47a31f42a23d2b374e1cd7d0b797e8e08b23d
1140053751 M * Bertl @: so a tag (handle) for that checkout?
1140053778 M * SamVilain this labels that revision with a nice user-visible tag
1140053812 M * SamVilain if you type "git-rev-parse v2.6.15", you can do just the version number to ID conversion; it will print the ID I listed before
1140053853 M * SamVilain then, you can type "git-cat-file -t dab47a3" (you don't need to type the whole SHA1 ID), to see what type that object has
1140053860 M * SamVilain it should print "tag"
1140053900 M * SamVilain "git-cat-file tag dab47a3" will show you the object
1140054038 M * SamVilain ok, normally, git-cat-file, etc, is pretty much an advanced tool, but it is a useful one to know about
1140054066 M * SamVilain so, on our new branch, with the new remote specified, we can start pulling revisions from the other tree
1140054099 M * SamVilain first check you are on the right branch with git-branch (git-pull will try to merge what it's pulled into the current checkout)
1140054192 M * SamVilain ok, sorry, instead of using "git-pull", we'll use stg for this
1140054205 M * SamVilain it's just a little less fiddly.
1140054223 M * SamVilain use "stg init" to set up stgit on this branch.  Otherwise, it will refuse to do anything.
1140054266 M * SamVilain stgit will set up information in .git/branches/2.6.15.y
1140054319 M * SamVilain then, you can pull with "stg pull 2.6.15.y"
1140054342 M * SamVilain correction, "stg pull chrisw-2.6.15.y"
1140054388 M * Bertl Pulling from "chrisw-2.6.15.y"...
1140054388 M * Bertl fatal: 'chrisw-2.6.15.y': unable to chdir or not a git archive
1140054388 M * Bertl fatal: unexpected EOF
1140054416 M * SamVilain ok, this helpful message means you didn't call your remote spec file .git/remotes/chrisw-2.6.15.y
1140054429 M * SamVilain so, it thought you were trying to pull revisions from a local directory
1140054463 M * Bertl Q: which means?
1140054507 M * SamVilain you are confused about the second statement?
1140054510 M * SamVilain or the first?
1140054523 M * Bertl what did I do wrong?
1140054572 M * SamVilain do you have a file .git/remotes/chrisw-2.6.15.y, and does it have the two lines I echoed to the channel?
1140054583 M * SamVilain those lines were
1140054584 M * SamVilain URL: git://kernel.org/pub/scm/linux/kernel/git/chrisw/linux-2.6.15.y.git
1140054584 M * SamVilain Pull: master:2.6.15.y
1140054632 M * Bertl hmm, I added that to the origin, as I think you said?!
1140054642 M * SamVilain aha, I see
1140054658 M * SamVilain right.  no, you should have two files, one called origin, the other chrisw-2.6.15.y
1140054664 M * SamVilain each two lines
1140054679 M * SamVilain there is probably a utility that configures these from the command line, of course :)
1140054685 A * Bertl is now fixing that ...
1140054687 M * SamVilain I just haven't found it/cared :)
1140054732 M * Bertl k, works now
1140054763 M * SamVilain ok.  So, we'll go and import a second tree, just because we're having so much fun.
1140054784 M * SamVilain When branching to pull in another tree, it is important that you branch on a revision that both the tree you have and the tree you are importing share
1140054802 M * SamVilain say we want to import ebiederm's tree
1140054852 M * SamVilain http://www.kernel.org/git/?p=linux/kernel/git/ebiederm/linux-2.6-ns.git;a=summary
1140054865 J * Aiken_ ~james@tooax7-229.dialup.optusnet.com.au
1140054905 M * SamVilain We can see if we look at the shortlog that the last revision that looks like it was upstream was by Benjamin Herrenschmidt
1140054911 M * SamVilain ie, http://www.kernel.org/git/?p=linux/kernel/git/ebiederm/linux-2.6-ns.git;a=commit;h=ac171c46667c1cb2ee9e22312291df6ed78e1b6e
1140054996 M * SamVilain so, we first branch on this common revision
1140055000 M * SamVilain git-checkout -b ebiederm-pspace ac171c46667c1cb2ee9e22312291df6ed78e1b6e
1140055085 M * SamVilain this time we won't use a remotes file, just use git-pull
1140055096 M * SamVilain git-pull git://kernel.org/pub/scm/linux/kernel/git/ebiederm/linux-2.6-ns.git
1140055134 M * SamVilain using "git-whatchanged", you should be able to see the last few commits are by ebiederm
1140055148 M * SamVilain I got  78 files changed, 2079 insertions(+), 911 deletions(-)
1140055185 Q * Aiken Ping timeout: 480 seconds
1140055217 M * SamVilain Right, so there we have pulling revisions from somebody else's repository
1140055259 M * SamVilain After a short recess while I grab a coffee, we'll import a split set of patches from Herbert
1140055274 M * Bertl hmm, I got a merge failed
1140055290 M * Bertl Automatic merge failed; fix up by hand
1140055292 M * SamVilain ok, are you on the right branch?
1140055325 M * SamVilain the "git-checkout -b ebiederm-pspace ac171c4" should have made a new branch called "ebiederm-pspace"
1140055346 M * Bertl * ebiederm-pspace
1140055374 M * SamVilain ok, so what's your last few revs with "git-whatchanged" ?
1140055426 A * SamVilain goes to the kitchen &
1140055436 M * Bertl hmm, seems I missed the ac171c4 part
1140055471 M * Skram Bertl: ever thought about merging this with the channel on freenode.net?
1140055520 M * Bertl Skram: don't know about 'meges' but we had a 'bridge-bot' for some time
1140055550 M * cehteh would be nice, then i can quit here ;)
1140055550 M * Skram meges?
1140055553 M * Skram Same
1140055555 M * Bertl *merges
1140055561 M * Skram Wel, I have a bridge bot..
1140055603 M * cehteh -ChanServ- [##unavailable] You've tried to join a channel which is unavailable.
1140055611 M * cehteh mhm .. no #vserver on freenode?
1140055627 M * Skram cehteh: #gentoo-vserver to precise
1140055633 M * Skram Bertl: mind if i started the bot up?
1140055636 M * Skram and we can test it?
1140055647 M * Bertl not right now ...
1140055650 M * Skram I take it you are the "maintainer" of this irc room
1140055652 M * Skram Oh, Okay.
1140055664 M * Skram Id be willing to host it on our companies dedicated server..
1140055666 M * Skram Alright
1140055666 M * Bertl when Sam is finished we can try 'again'
1140055677 M * Skram Sam? Alright..
1140055680 M * Skram well, hit me up with a pm
1140055705 M * cehteh Skram: well i am not interested in gentoo .. and vserver should be distribution agnostic
1140055712 M * Skram Yeah
1140055719 M * Skram fine #vserver, no biggy
1140055739 M * Bertl we closed that one after the last user moved here :)
1140055746 M * Skram :(
1140055751 A * SamVilain returns with a cup of coffee
1140055768 A * Skram cyphins SamVilain's coffee out into his own cup.
1140055790 A * SamVilain takes a phone call
1140055805 M * cehteh Bertl: lol i was sometimes there and like freenode more than OFTC (not because of the network policies, freenode staff seems to smoke wrong pot sometimes) .. but there are the right people :)
1140055828 M * Skram I agree, somewhat
1140055893 M * Bertl well, depends on the perspective, I guess :)
1140055939 M * Skram Indeed, Indeed.
1140055983 M * mnemoc most people is on both servers i guess
1140055993 M * mnemoc networks*
1140055996 M * SamVilain heh, that was the rental car compant
1140055999 M * Bertl at least those with better clients :)
1140056006 M * SamVilain after some excess
1140056008 M * SamVilain http://www.utsl.gen.nz/media/roadtrip06/tn/sany0185.med.jpg
1140056130 M * SamVilain ok, back to the program
1140056146 M * SamVilain We'll focus this half of the talk on stgit
1140056172 M * SamVilain Bertl: did you get your ebiederm branch to pull in the end?
1140056220 M * Bertl no, how to remove the old one?
1140056232 M * SamVilain to delete a branch,
1140056238 M * SamVilain first checkout a different branch
1140056248 M * SamVilain then use "git-branch -d badbranch"
1140056280 M * Bertl fatal: you need to resolve your current index first
1140056331 M * SamVilain ok, "current index" refers to which files are changed in the currently checked out branch
1140056336 M * SamVilain git sure has some funny terms
1140056346 M * SamVilain so, you can run "git-status" or "stg status"
1140056366 M * SamVilain do they say "nothing to commit"?
1140056376 M * SamVilain (or zero output, in the stg status case) ?
1140056386 M * Bertl okay, gives me an endless list
1140056405 M * SamVilain aha.  something went wrong.  you can revert your checkout with git-checkout -f
1140056409 M * Bertl many, many modified
1140056412 Q * FireEgl Ping timeout: 480 seconds
1140056435 M * SamVilain this is probably the result of this bad merge operation
1140056439 M * Bertl okay, nothing to commit
1140056496 M * Bertl how do I 'checkout' a different branch
1140056501 M * SamVilain git-checkout branchname
1140056511 M * Bertl git-branch 2.6.15.y
1140056511 M * Bertl 2.6.15.y already exists.
1140056523 M * SamVilain git-branch creates a branch
1140056535 M * SamVilain git-checkout checks out a branch
1140056537 M * Bertl okay, do I have to do that just to delete the bad one?
1140056549 M * SamVilain git-checkout master; git-branch -d ebiederm-pspace
1140056556 M * Bertl ah, got it
1140056595 M * SamVilain er, -D :)
1140056604 M * Bertl too late
1140056611 M * SamVilain well, -d would have complained
1140056617 M * Bertl git-branch -d ebiederm-pspace
1140056617 M * Bertl Deleted branch ebiederm-pspace.
1140056651 M * SamVilain right, so the branch you switched to was a parent of the branch you made.  that gives a hint as to what went wrong
1140056655 M * SamVilain git-checkout -b ebiederm-pspace ac171c
1140056661 M * SamVilain was the correct command to create the branch
1140056670 M * Bertl @: now what is the 'important' difference I have now with the SHA, compared to before
1140056693 M * SamVilain right, instead of the default, which would have been where you were, which was 2.6.15.y
1140056711 M * SamVilain so now
1140056712 M * SamVilain git-pull git://kernel.org/pub/scm/linux/kernel/git/ebiederm/linux-2.6-ns.git
1140056714 M * SamVilain should work
1140056744 M * SamVilain in my case it was a lot faster, because it already found the commit objects in my tree
1140056778 M * Bertl k, did something
1140056790 M * SamVilain this was because "git-branch -d" does not actually delete the objects on disk; only "git-prune" deletes unused objects
1140056829 M * SamVilain in short, it just removed .git/refs/heads/ebiederm-pspace, which is just a file holding a pointer to the head revision for that branch
1140056895 M * SamVilain right, so let's import the del-2.6.16-rc1-vs2.1.0.9/ series
1140056915 M * SamVilain first, we will need to switch to the correct base and make a branch
1140056941 M * SamVilain git-checkout -b 2.6.16-rc1-vs2.1.0.9 v2.6.16-rc1
1140056989 M * SamVilain right, now that we will be importing patches, we should set up some environment variables
1140057002 M * SamVilain wilber:~/vserver/test-linux-2.6$ env | grep GIT
1140057002 M * SamVilain GIT_AUTHOR_NAME=Herbert Pötzl
1140057002 M * SamVilain GIT_COMMITTER_NAME=Sam Vilain
1140057002 M * SamVilain GIT_COMMITTER_EMAIL=sam.vilain@catalyst.net.nz
1140057002 M * SamVilain GIT_AUTHOR_EMAIL=herbert@13thfloor.at
1140057036 M * SamVilain This is what I have been using; I have been noticing encoding problems, probably because I use a UTF-8 locale
1140057048 M * Bertl just use Poetzl instead
1140057075 M * SamVilain okay
1140057089 M * SamVilain right, so we first init this branch with "stg init"
1140057093 M * mnemoc there is any good comparison between tla-family and git?
1140057123 M * SamVilain you mean arch/bazaar/etc?  hmm
1140057146 M * mnemoc yes
1140057177 M * SamVilain I haven't really looked, although of course arch pretty much defines the "changeset oriented" paradigm I mentioned in the introduction to this tut
1140057208 M * SamVilain So, git is "Content oriented".  For more, read the backlog :)
1140057215 M * mnemoc i read it :)
1140057230 M * mnemoc but ok, i'll keep searching
1140057243 M * Skram !seen SuperLag
1140057287 M * SamVilain Other than that, I haven't really used git very much, but Martin Langhoff, who gave a tut at Linux.conf.au about git (which Linus attended), is an arch veteran who can talk at great lengths about the failings of arch ;)
1140057305 M * SamVilain er, I haven't really used *arch* very much :)))
1140057314 M * mnemoc :)
1140057331 M * SamVilain anyhoo
1140057359 M * SamVilain We'll import the first patch in the patchset.  Assuming you've already downloaded all the files in http://vserver.13thfloor.at/Experimental/del-2.6.16-rc1-vs2.1.0.9/,
1140057393 M * SamVilain you can either apply the patch, then name it, or name it then apply the patch
1140057419 M * SamVilain however with current stgit you need to use "-f" when naming it (though the next version doesn't care)
1140057428 M * SamVilain if there are already changes
1140057450 M * SamVilain so, we'll use this to name/describe the patch:
1140057495 M * SamVilain stg new -m "From del-2.6.16-rc1-vs2.1.0.9/01_include.diff" --authdate=2006-01-31 include
1140057528 M * SamVilain without -m, stgit will pop up your $EDITOR to edit the comment
1140057537 M * SamVilain the first line is the patch summary visible in all the shortlogs etc
1140057557 M * SamVilain or, the subject line when exporting patchsets to e-mails
1140057566 M * SamVilain then, apply the patch as normal;
1140057610 M * SamVilain patch -p1 < ../path/del-2.6.16-rc1-vs2.1.0.9/01_include.diff
1140057629 M * SamVilain now if you run "stg status" or "git-status", you will see which files have changed
1140057656 M * SamVilain if there are any new files, you have to add them with "git-add filename"
1140057671 M * SamVilain however unlike CVS/SVN/etc, there is no need to mark files as removed
1140057718 M * SamVilain so, the first patch only changes existing files, so we can go ahead and type "stg refresh" to refresh the patch with local changes
1140057751 M * SamVilain typing "stg series" will show you what patches you have created
1140057797 M * SamVilain then we carry this on for the rest of the patches in the series
1140057802 M * SamVilain stg new -m "From del-2.6.16-rc1-vs2.1.0.9/02_inline.diff" --authdate=2006-01-31 include
1140057807 M * SamVilain bah
1140057814 M * SamVilain that last paste should have been
1140057829 M * SamVilain stg new -m "From del-2.6.16-rc1-vs2.1.0.9/02_inline.diff" --authdate=2006-01-31 inline
1140057849 M * SamVilain patch -p1 < ../patches/13thfloor/2.6.16-rc1-vs2.1.0.9/02_inline.diff 
1140057869 M * SamVilain if you run "git-status" this time, you'll notice that new files have been added,
1140057891 M * SamVilain so you have to "git-add include/linux/vs*"
1140057901 M * SamVilain before "stg refresh" will pick up those files
1140057918 M * Bertl Q: can I somehow auto-add all new files?
1140057974 M * SamVilain heh, I normally use a command from patchy git to do that.
1140057983 M * SamVilain I'll just put my "import-quilt" script online, sec
1140058003 M * SamVilain http://vserver.utsl.gen.nz/scripts/import-quilt
1140058022 M * SamVilain \argh
1140058047 M * SamVilain missing dependency
1140058185 M * SamVilain ok, there we go, it has the other script it called inline now
1140058268 M * SamVilain Right, so anyway, we now should have two patches on our stgit stack
1140058274 M * SamVilain you can see the stack with "stg series"
1140058289 M * SamVilain we can remove patches with "stg pop"
1140058303 M * SamVilain or, perhaps "stg pop -t patchname"
1140058346 M * SamVilain so, what we will do is pop all the patches off, delete them, then import them en masse with import-quilt
1140058350 M * SamVilain use:
1140058352 M * SamVilain stg pop -a
1140058357 M * SamVilain stg delete include
1140058359 M * SamVilain stg delete inline
1140058414 M * SamVilain ok, I have updated the import-quilt script again, removing the options it was passing to stg new
1140058430 M * SamVilain such as specifying the author and committer
1140058495 M * SamVilain ok, how about I test it first
1140058508 M * SamVilain what's that golden rule of tutorials, never do a live demo?  ;)
1140058524 M * Bertl :)
1140058553 M * SamVilain ok, anyway, it should work now, if you have the split set in a directory, and everything is popped off and deleted, then "import-quilt ../path-to/del-2.6.16-rc1-vs2.1.0.9" should pull them all in one by one
1140058640 M * SamVilain ok, I have 3 minutes left to cover re-basing, moving changes between patches in our patchset and exporting the patches :)
1140058660 M * SamVilain re-basing is simple.  You just use:
1140058662 M * SamVilain stg pop -a
1140058680 M * SamVilain apply the changes that bring you to the new base
1140058684 M * SamVilain then stg push -a
1140058716 M * SamVilain or, another way is to make a new branch,
1140058728 M * SamVilain then use "stg pick" to pull in patches one by one to the new branch
1140058752 M * SamVilain however, we should first tell stgit how to merge
1140058767 M * SamVilain http://vserver.utsl.gen.nz/scripts/smartmerge  # what I use to merge
1140058784 M * SamVilain stgit comes with an example rc file
1140058791 M * SamVilain  /usr/local/share/stgit/examples/stgitrc
1140058812 M * SamVilain "diff3" is the default merger, which is good, but it can't handle conflicts
1140058846 M * SamVilain using "emacs" for a merger is not good either, because if the merging doesn't need intervention, I don't want to have to close the popped up window
1140058875 M * SamVilain so, I use the smartmerge script I posted, and my ~/.stgitrc looks like:
1140058885 M * SamVilain [gitmergeonefile]
1140058886 M * SamVilain merger: smartmerge "%(branch1)s" "%(branch2)s" "%(ancestor)s" "%(output)s"
1140058935 M * SamVilain This tries diff3 first, then if that fails, falls back to ediff-merge-files-with-ancestor, which I've found to be da bomb.
1140058961 M * SamVilain customise to taste :)
1140059012 M * SamVilain so, for instance, to re-base that old set against the git HEAD
1140059043 M * SamVilain I'd use : git-checkout -b HEAD-vs2.1.0.9 master
1140059049 M * SamVilain stg init
1140059080 M * SamVilain patches=`stg series -b 2.6.16-rc1-vs2.1.0.9 | awk '{print $2}'`
1140059085 M * SamVilain then
1140059097 M * SamVilain for x in $patches; do stg pick $x@2.6.16-rc1-vs2.1.0.9; done
1140059158 M * SamVilain however, I didn't really need to do all that.  I could have just used "stg pull"; and that's what I would use to keep this new branch up to date
1140059171 M * SamVilain But I wanted to keep the 2.6.16-rc1-vs2.1.0.9 branch for later reference
1140059263 M * SamVilain Moving changes between patches is also fairly easy, especially if you are moving them UPwards
1140059302 M * SamVilain you just "stg pop -t targetpatch", apply the change, then "stg push -t sourcepatch".  If you're lucky, diff3 will DTRT, otherwise you'll have to manually revert the change from the later patch
1140059325 M * SamVilain when you want to export to a series of files, you can use "stg export -nd ../path/to/export"
1140059353 M * SamVilain before doing that, you should remove ../path/to/export, and make sure all your patches are pushed
1140059364 M * SamVilain as stg export will only export all the ones you have pushed so far
1140059434 M * SamVilain ok, well, I've pretty much covered the main course of this meal.  My apologies if the later parts seemed a little ... rushed :)
1140059481 M * Bertl guess we will revisit some points at a later time ...
1140059506 M * SamVilain sure.  If you'd like to pull in my vserver-inclusion working tree, you can use;
1140059539 M * SamVilain git-branch -b 2.6.16-rc2-vsi v2.6.16-rc2
1140059567 M * SamVilain er, git-checkout -b 2.6.16-rc2-vsi v2.6.16-rc2
1140059567 M * SamVilain :)
1140059649 A * SamVilain tests his command
1140059724 A * SamVilain taps his feet
1140059747 M * SamVilain ok
1140059748 M * SamVilain git-pull git://vserver.utsl.gen.nz/git/vserver.git 2.6.16-rc2-vsi
1140059750 M * SamVilain  :)
1140059826 M * SamVilain however, perhaps it is better to get http://vserver.utsl.gen.nz/patches/utsl/2.6.16-rc2-vsi.tar.gz, and use the import-quilt script
1140059851 Q * shuri Remote host closed the connection
1140059871 M * SamVilain except, that will not work as it is - as that tarball wasn't made with stg export -n so the files are not in order
1140059878 A * SamVilain sighs
1140059885 M * SamVilain well, that's where I'm at, anyway :)
1140059891 A * SamVilain turns off the mic
1140059897 N * SamVilain mugwump
1140059934 T * mugwump http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.1-rc6/3, 2.0.2-rc5/1 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 
1140059942 M * mugwump thank you all for your time
1140059955 M * mugwump I mean, listening to me :)
1140059985 M * mnemoc :)
1140060005 M * Bertl mugwump: thank you for your informative introduction!
1140060020 M * mnemoc yep, thanks a lot
1140060043 M * mugwump hmm, huge audience ;)
1140060056 M * mugwump a good thing the tape recorder's running
1140060063 M * Bertl :)
1140060090 M * mugwump note that there is a minor ... issue with the way that stgit handles repository revisions
1140060102 M * mugwump when you "stg pop", you're effectively uncommitting a change
1140060124 M * mugwump then you "stg push" with some changes, and the new commit id is different
1140060160 M * mugwump so, using the core git tools, somebody who pulled the top of your stgit patch can't pull any more
1140060198 M * mugwump There's a current thread on the GIT list about this.  Patchy Git (pg) was designed to fix this, but it's early days yet.
1140060273 M * mugwump note that it is easy to "wind back" a branch of a repository; just change .git/refs/heads/branchname
1140060287 M * mnemoc o_O
1140060317 M * mugwump yes, if you have published those commits then people pulling your revisions must know how to deal with it :)
1140060331 M * mugwump but hey, this is all extremely impressive for a VCS less than a year old
1140060339 M * mugwump oh, we didn't even pull out gitk
1140060386 M * mugwump use something like: gitk --since="3 weeks ago" 2.6.16-rc2-vsi ebiederm-pspace for an interesting look at the revision history
1140060459 M * mugwump each line on the left is a branch, and you can see branching and merging as these lines joining or forking
1140061616 M * Bertl okay, off to bed now .. quite tired ... back tomorrow!
1140061626 M * Bertl have a good one, everyone!
1140061638 N * Bertl Bertl_zZ
1140061664 M * mugwump seeya Bertl_zZ ... a shame our $TZ are so different
1140062050 P * nem wishes he was off to hack lisp
1140062764 M * cehteh Bertl_zZ: has strange timezones anyways :) its 5am here
1140067211 M * matti cehteh: Hehehe.
1140067228 M * matti cehteh: Not so different.
1140067239 M * matti cehteh: But Bertl have his own timezone.
1140067252 M * matti cehteh: We called it BUT - Bertl Universal Time.
1140067253 M * matti ;]
1140067260 M * mnemoc *G*
1140067288 A * cehteh should normalize his TZ too
1140067312 M * Skram haha
1140067330 M * matti Anyone wants some coffee?
1140067330 M * matti ;]
1140067334 M * Skram YES.
1140067335 M * Skram I need some
1140067342 M * eyck nope, thanks, 
1140067348 M * mnemoc no no... i want to sleep tonight
1140067350 M * matti I am in GMT +1, to coffee is more than appropriate right now ;]
1140067360 M * matti s/to/so/
1140067376 A * matti gives Skram cup of hot coffee.
1140067382 M * matti Skram: Milk? Sugar?
1140067393 M * Skram Black, baby.
1140067411 M * matti Aye aye sir!
1140067413 M * matti So, enjoy.
1140068016 M * cehteh i better go to bed for some hours
1140068059 M * matti So, sweet dreams then.
1140068059 M * matti ;]
1140068078 M * matti cehteh: http://graffias.estrefa.pl/~matti/tmp/go2sleep.swf
1140068079 M * matti ;-p
1140068158 A * cehteh waits how long it takes until the horror pic flashed up
1140068166 M * matti Hehehe.
1140068178 M * matti Right ;-p
1140068210 M * matti cehteh: Well, after horror event, ya can relax with http://graffias.estrefa.pl/~matti/tmp/bubblewrap.swf ;]
1140068259 M * cehteh lol
1140068266 M * matti ;]
1140068381 M * cehteh http://www.pipapo.org/people/ct/Karriereleiter.swf ... for the ones who understand german
1140071031 J * FireEgl Atlantica@Atlantica.DollarDNS.Net
1140071310 M * ebiederm_oO mugwump: I don't know if you are around but a quick tutorial update.
1140071342 M * ebiederm_oO you can just do git-fetch <remote-repository> remote-branch:local-branch
1140071352 M * ebiederm_oO Then git-checkout local-branch
1140071366 M * ebiederm_oO git-fetch is a lot less likely to cause problems.....
1140072165 Q * shedi Ping timeout: 480 seconds
1140072693 J * shedi ~siggi@inferno.lhi.is
1140073123 J * Aiken__ ~james@tooax6-174.dialup.optusnet.com.au
1140073445 Q * Aiken_ Ping timeout: 480 seconds
1140074626 N * ebiederm_oO ebiederm_zZ
1140078653 Q * shedi Quit: Leaving
1140078943 Q * michal` Quit: chilling out
1140079947 J * prae ~prae@ezoffice.mandriva.com
1140083790 J * meandtheshell ~markus@85-125-230-44.dynamic.xdsl-line.inode.at
1140086212 J * shedi ~siggi@tolvudeild-204.lhi.is
1140086239 J * lilalinux ~plasma@80.69.35.186
1140087656 J * mnemoc_ ~amery@200.75.27.38
1140087762 Q * mnemoc Ping timeout: 480 seconds
1140087762 N * mnemoc_ mnemoc
1140088002 Q * mcp Read error: Connection reset by peer
1140088089 Q * Hollow Read error: Connection reset by peer
1140088092 J * mcp ~hightower@wolk-project.de
1140088092 J * Hollow_ ~hollow@home.xnull.de
1140090375 Q * Aiken__ Ping timeout: 480 seconds
1140090591 Q * SNy Remote host closed the connection
1140090614 J * SNy 55011e3077@bmx-chemnitz.de
1140090702 J * Doener doener@i5387EFE7.versanet.de
1140091148 J * pflanze ~chris@84-73-63-10.dclient.hispeed.ch
1140091156 M * pflanze Hello
1140091538 Q * Hollow_ Remote host closed the connection
1140091544 J * Hollow ~hollow@home.xnull.de
1140091556 M * Hollow morning
1140091576 M * Doener  morning Hollow, pflanze 
1140091600 Q * mikeb Remote host closed the connection
1140091618 J * mikeb ~baptiste@tigger.msbnetworks.net
1140091950 M * Hollow hey Doener! everything fine?
1140091976 M * Doener sure, except for the usual lack of time ;)
1140091998 M * Doener just wondered where my memory went until I noticed that my slab cache grew to 1.3G
1140092036 M * Hollow heh, i just finished a trenary search tree implementation for vserver-utils
1140092053 M * Hollow took me two days.. lmao
1140092072 M * Hollow s/lmao/lmaa/ *g*
1140092087 M * Doener heh
1140092538 J * k4mper ~maio@mail.progamers.cz
1140092541 M * k4mper hi
1140092590 M * k4mper i have sucesfully (nearly :) installed my first vserver but when i install ssh and try to ssh to it from another host i'm connecting to my real machine and not vserver
1140092598 M * k4mper any idea what might be wrong?
1140092638 M * SiD3WiNDR your host's sshd is bound to 0.0.0.0
1140092646 M * SiD3WiNDR you have to restrict it to the main ip
1140092651 M * SiD3WiNDR listenaddress in the sshd_config file
1140092662 M * k4mper thanks. gonna try that
1140092693 M * SiD3WiNDR then, on the vserver, restart the sshd
1140092698 M * SiD3WiNDR cause it will not have started because the port was in use
1140093030 M * k4mper thanks SiD3WiNDR 
1140093218 Q * wibble_ Quit: Lost terminal
1140093458 M * pflanze Has anyone run UML inside vserver?
1140093513 M * pflanze I've got problems with it's networking, and additionally the uml booting process hangs while starting atd or crond
1140093529 M * pflanze (Might be the latter is related to the former.)
1140093575 M * pflanze I'm trying to run uml-switch on the host, and bind mount it's var directory (socket + pid file) to the guest.
1140093616 M * pflanze the users and groups match, but  cat > /var/run/uml-utilities/uml_switch.ctl  says permission error.
1140093676 M * pflanze while if I run uml-switch inside the guest, the above command says No such device or address
1140093694 M * pflanze but then it cannot use the tap0 device.
1140093748 M * pflanze So is there any solution to either the permission problem (how can this be? is there a protection of unix sockets bindmounted to other vservers?), or the tap usage inside a guest?
1140094482 Q * Doener Ping timeout: 480 seconds
1140094518 J * Doener doener@i5387CF5E.versanet.de
1140094967 J * VxJasonxV ~jason@ip68-110-115-17.ph.ph.cox.net
1140095359 P * k4mper 
1140095606 J * Smutje_ ~Smutje@xdsl-84-44-247-111.netcologne.de
1140095710 Q * Smutje Ping timeout: 480 seconds
1140095710 N * Smutje_ Smutje
1140095935 M * SiD3WiNDR is there a vs release for 2.6.15.3?
1140095943 M * SiD3WiNDR I don't want my boxen dead with an icmp packet :p
1140095994 M * brc_ didnt get it. Why dead with an icmp packet ?
1140096002 M * SiD3WiNDR http://www.securityfocus.com/bid/16532
1140096039 M * brc_ lemme check it
1140096075 M * brc_ hmm, didnt know about it
1140096077 M * brc_ that sux :(
1140096092 M * brc_ if i block icmp requests, will it be solved ?
1140096867 M * Doener SiD3WiNDR: that's a single line change, anything that applies to 2.6.15.2 should apply to 2.6.15.3 as well (except Makefile of course)
1140096896 M * Doener s/should/will/ 99,999%+ ;)
1140097005 N * Bertl_zZ Bertl
1140097010 M * Doener morning Bertl
1140097046 M * brc_ Doener, do you think that dropping ICMP packets on the firewall would stop the DDOS ?
1140097048 M * brc_ oops DOS
1140097068 M * Bertl morning folks!
1140097113 M * brc_ mroning bertl
1140097150 M * FaUl brc_: you may just change one line in the source -thats all the patch...
1140097189 M * brc_ the problem is that i have some servers, and a workaround before the patch would make me feel safer
1140097344 M * SiD3WiNDR brc_: yea should be
1140097357 M * SiD3WiNDR Doener: ah, I didn't know the patch to .3 was so small - thanks :)
1140097412 M * Bertl what's wrong with 2.6.15.4?
1140097470 M * SiD3WiNDR is there a .4 now too?! :p
1140097507 A * SiD3WiNDR wasnt very much paying attention
1140098990 M * SiD3WiNDR patch-2.6.15.4-vs2.0.2-rc5.diff <- this the latest non-dev for 2.6.15.4, right?
1140099490 M * Bertl yup
1140099816 M * VxJasonxV A vserver still needs things like udev in order to function, right?
1140099863 M * Doener you never _need_ udev, you can always do with a static /dev ... and _inside_ a vserver udev won't work at all
1140099888 M * Doener at least if it is configured to be safe, i.e. it can't create dev nodes
1140099895 M * VxJasonxV inside a vserver was what I was asking
1140099903 M * VxJasonxV ok, so I can mask that, and figure out if gentoo screwed up deps again
1140099968 Q * mikeb Remote host closed the connection
1140100022 M * Bertl VxJasonxV: IIRC there is a special guest layout for gentoo
1140100030 M * VxJasonxV baselayout-vserver :-)
1140100044 M * VxJasonxV which I'm using.  but I'm trying to update, and running into all sorts of dependancy crap
1140100052 M * VxJasonxV I don't know if there's a special udev.  I suppose I should look...
1140100992 J * meebey_ meebey@booster.qnetp.net
1140100993 Q * meebey Read error: Connection reset by peer
1140101241 J * iW|Rebel ~rebel@host103-239.pool8254.interbusiness.it
1140101358 M * Bertl welcome iW|Rebel!
1140101381 M * Roey hi all
1140101386 M * Roey doener, blitzen, etc.
1140101389 M * Roey bertl, vxjasonxv
1140101477 M * VxJasonxV ?
1140101506 M * Roey just saying hi
1140101591 M * VxJasonxV hi :-)
1140101614 M * Roey =)
1140101629 M * VxJasonxV garrr
1140101636 M * VxJasonxV these dependencies are very much bothering me
1140101697 P * VxJasonxV de/part'ing
1140101701 J * VxJasonxV ~jason@ip68-110-115-17.ph.ph.cox.net
1140101703 M * VxJasonxV oops...
1140101830 M * Roey they bother you to the point that you lose your connection.
1140101831 M * Roey ahem.
1140101832 M * Roey ;)
1140101834 M * VxJasonxV :P
1140102161 M * VxJasonxV is anyone in here a gentoo user?
1140102179 A * matti 
1140102195 M * matti And Hollow.
1140102201 M * matti And Hollow is an developer.
1140102202 M * matti ;]
1140102204 M * Bertl hey matti!
1140102205 M * VxJasonxV YEah, SuperLag named Hollow, but he's not around is he?
1140102210 M * matti Bertl: :*
1140102263 M * VxJasonxV I asked, and someone answered, but since I accidentally closed the window I can't scroll up and re-read what he said :(
1140102277 M * matti VxJasonxV: Asked about?
1140102282 M * VxJasonxV I'm trying to emerge world (in my vserver, on amd64) and it keeps whining about udev.  Do I, or don't I, need udev?
1140102301 M * VxJasonxV (udev wants baselayout, and not baselayout-vserver, which will be the problem if I need udev...)
1140102310 M * matti VxJasonxV: Wrong profile?
1140102312 M * VxJasonxV I'll rephrase it too: Do I WANT udev (i.e. will it be easier to use it)
1140102317 M * VxJasonxV shouldn't be using the wrong profile...
1140102319 M * VxJasonxV I'll check though
1140102367 M * VxJasonxV hmmm
1140102372 M * VxJasonxV I don't remember where to check >_>
1140102382 M * VxJasonxV I know where profiles are, but not to check which one I'm using
1140102435 M * matti VxJasonxV: Heh.
1140102444 M * matti VxJasonxV: So, this is not an vserver issue.
1140102469 M * VxJasonxV it might be
1140102473 M * matti VxJasonxV: And ya need better understanding of distribution y're using.
1140102480 M * matti VxJasonxV: /etc/make.profile.
1140102507 M * VxJasonxV lrwxrwxrwx  1 root root 58 Jan 27 22:19 /etc/make.profile -> ../usr/portage/profiles/default-linux/amd64/2005.1/vserver/
1140102515 M * VxJasonxV I totally forgot about that file
1140102519 M * matti ;-p
1140102523 M * iW|Rebel hi :)
1140102550 M * iW|Rebel I had a problem with klogd eating whole cpu
1140102564 M * VxJasonxV there's no vserver in 2006.0 yet, so I won't change it
1140102569 M * VxJasonxV so I'm using the most current vserver profile
1140102577 M * matti VxJasonxV: I see.
1140102593 M * iW|Rebel someone told me to try kernel 2.6.15.4, but problem's stil there
1140102612 M * iW|Rebel I'm compiling dev kernel now, maybe I'll be more lucky...
1140102638 M * matti VxJasonxV: Now, you can supress udev emerging.
1140102645 M * VxJasonxV YEah, I can
1140102666 M * matti VxJasonxV: By using /etc/portage/packages.mask
1140102673 M * VxJasonxV it's package.mask
1140102676 M * VxJasonxV and yes, I know :)
1140102678 M * Bertl iW|Rebel: you want to either:
1140102681 M * VxJasonxV I'm no strange to /etc/portage/*
1140102686 M * matti VxJasonxV: ;-)
1140102688 M * Bertl a) remove the entry for the klogd for your guest
1140102706 M * VxJasonxV The problem I run into is everything dep'ing on udev
1140102710 M * Bertl iW|Rebel: or, b) add the VXC_SYSLOG flag to your config
1140102719 M * VxJasonxV and udev dep'ing on baselayout (and not baselayout-vserver)
1140102764 M * matti VxJasonxV: Hmm...
1140102789 M * iW|Rebel <Bertl> iW|Rebel: or, b) add the VXC_SYSLOG flag to your config <--  where exactly? I'm quite new with vserver...
1140102794 M * matti VxJasonxV: emerge --info, emerge -pv world?
1140102833 M * VxJasonxV and I don't need a kernel either, rawr
1140102858 Q * Loki_muh Remote host closed the connection
1140102948 Q * meebey_ Remote host closed the connection
1140102978 J * meebey meebey@booster.qnetp.net
1140103055 Q * W0nka Remote host closed the connection
1140103057 J * Wonka produziert@chaos.in-kiel.de
1140103058 M * Bertl lag
1140103058 Q * Bertl Remote host closed the connection
1140103064 Q * eyck Remote host closed the connection
1140103087 J * Bertl herbert@212.16.62.52
1140103093 J * eyck eyck@81.219.64.71
1140103122 M * iW|Rebel excuse me berti, where do I have to put  VXC_SYSLOG in config?
1140103228 M * Bertl http://linux-vserver.org/Caps+And+Flags
1140103244 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1140103250 M * Hollow VxJasonxV: http://bugs.gentoo.org/show_bug.cgi?id=105616
1140103260 M * VxJasonxV good morning, Hollow :)
1140103265 M * Hollow morning
1140103273 M * Bertl iW|Rebel: so you add 'syslog' to ccapabilities
1140103287 M * VxJasonxV I put kexec into my kernel (talking desktop here, not vserver) and forgot to emerge the tools...
1140103288 M * VxJasonxV awesome
1140103292 M * Hollow the interesting question is, which package does depend on udev? because if it runs inside a vserver it probably won't need udev, and the deps are wrong
1140103310 M * matti Hollow: ;-)
1140103314 M * iW|Rebel kk, thx Bertl :D
1140103324 M * matti Hollow: Bug in RDEPEND?
1140103334 M * Hollow matti: well, that's the overall problem of this crappy udev bug... you normally don't need to depend on udev, it's completely crap.
1140103344 M * matti Yep.
1140103346 M * matti I know.
1140103351 J * Loki|muh loki@satanix.de
1140103382 M * Hollow VxJasonxV: look at comment 26 for a workaround
1140103475 M * VxJasonxV JUST read it :)
1140103475 M * VxJasonxV thanks
1140103498 M * Hollow you're welcome!
1140103500 M * VxJasonxV does .provided support version globbing? >=sys-fs/udev-0 ?
1140103510 M * Hollow no
1140103513 M * VxJasonxV drat
1140103517 M * VxJasonxV alright, thanks again
1140103541 M * Hollow just use -999 or so
1140103555 M * Hollow should prevent world update failures
1140103556 M * VxJasonxV ahhhh, good idea
1140103618 M * Doener Hollow: btw, what's the meaning of RDEPEND? Packages that depend on this?
1140103635 Q * iW|Rebel Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )
1140103663 M * VxJasonxV is /etc/portage/profile supposed to be a symlink?
1140103682 M * VxJasonxV because I don't have that dir on any of my systems
1140103686 M * Hollow DEPEND: build deps, RDEPEND: runtime deps, PDEPEND: post deps
1140103716 M * Hollow (though PDEPEND is just used in special-case uglyness afaik)
1140103749 M * Doener ok
1140103775 M * VxJasonxV weeee, I can finally update
1140103777 M * VxJasonxV gracias hollow
1140103811 M * matti Doener: http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=1
1140103817 M * Hollow and the distinction between DEPEND and RDEPEND is not persistent at all
1140103835 M * matti Doener: Especially http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=1#doc_chap5
1140103836 M * Hollow (i.e. wrong RDEPENDs all over the place because people just use DEPEND for both
1140103838 M * matti :)
1140103845 M * matti Ineed.
1140103851 M * matti Hollow h've 100% right.
1140103857 M * matti And that is a pain in the ass.
1140103879 M * matti Custom and private overlay is some solution for that... Heh...
1140103902 M * Hollow indeed, espcially for special profile cases like vserver (not necesarrily because RDEPEND, but because people don't care about deps if it just works)
1140103921 M * matti Yep ;/
1140103955 M * VxJasonxV now to kick groff 1.19 in the ass because it blocks man-pages-ja and I can finally up
1140104003 M * Hollow VxJasonxV: if you run more gentoo guests (and have a gentoo host) you should definitely take a look at vupdateworld, vdispatch-conf and vemerge
1140104015 M * VxJasonxV I'm not the server owner ;)
1140104020 M * Hollow ah, ok :)
1140104038 M * VxJasonxV I'll defintely see if the server owner knows/uses those though
1140104048 M * VxJasonxV because he uses vserver EVERYWHERE
1140104051 M * Hollow is it a private box or do you rent this gentoo vserver?
1140104052 M * VxJasonxV (everywhere = home and work)
1140104055 M * VxJasonxV private box
1140104098 M * Hollow so, are you happy with gentoo as vserver? ;)
1140104107 M * Hollow (besides the udev bug ;)
1140104112 M * VxJasonxV yeah, thus far I hadn't had any problems
1140104116 M * Hollow but this will definitely be solved for 2006.0
1140104122 M * VxJasonxV the fact that it's fixed now makes me not care too :)
1140104124 M * matti Hollow: Hehe I see the QA question ;]
1140104142 N * ebiederm_zZ ebiederm
1140104151 M * Hollow we will probably have to introduce a new-style baselayout virtual to solve these problems..
1140104182 M * Hollow because, technically speaking, only versioned baselayout deps are the main cause here
1140104192 M * matti Hollow: I am happy, and romke is IIRC also very greatful of ya work ;]
1140104203 M * Hollow thanks :)
1140104206 M * romke :>
1140104208 M * matti Hollow: In Portage 2.2?
1140104214 M * matti Oh, speaking of witch ;p
1140104218 M * matti romke: Hello mate.
1140104222 M * matti romke: What's up?
1140104239 M * Hollow hm, no 2.1 at least... but i guess it works with some recent version of 2.0 as well
1140104246 M * matti I see.
1140104249 M * romke matti: hi, just building new gentoo guest (for ksi;))
1140104261 M * matti romke: Sweet.
1140104275 Q * matti Remote host closed the connection
1140104320 M * Hollow matti: this was introduced at the end of last year, and moved to stable very soon, because versioned depends on packages which should be a virtual are a PITA all over the place, and these new-style virtuals (which can be versioned, and actually live in /usr/portage/virtual/ as normal ebuilds) solve these problems
1140104338 M * VxJasonxV he got the boot ;D
1140104447 J * matti matti@linux.gentoo.pl
1140104455 M * matti Oops.
1140104462 M * matti OOM kills my irssi ;-)
1140104470 M * SiD3WiNDR eheh
1140104472 M * SiD3WiNDR OOM sucks.
1140104510 M * matti Yee... I need to change its beaviour to return null.
1140104513 M * matti ;-)
1140104515 M * Hollow matti: this was introduced at the end of last year, and moved to stable very soon, because versioned depends on packages which should be a virtual are a PITA all over the place, and these new-style virtuals (which can be versioned, and actually live in /usr/portage/virtual/ as normal ebuilds) solve these problems
1140104550 M * matti Hollow: I hope so.
1140104620 M * matti Hollow: BTW, I also hope, that x86 team will be more in-touch - there's a lot of "version bump" and consider as stable packages... not moved to stable tree ;/
1140104621 M * Roey OOM.
1140104631 M * matti Roey: Out Of Memory.
1140104644 M * Roey Does the OOM Killer trigger more frequently when you run multiple vservers?
1140104647 M * Roey matti:  aye, thanks
1140104656 M * matti Oh, sorry ;-p
1140104713 Q * matti Remote host closed the connection
1140104919 J * matti matti@linux.gentoo.pl
1140104932 M * matti Heh, that was ain't funny ;-|
1140104939 M * matti Bleh.
1140104965 Q * matti Remote host closed the connection
1140104968 M * SiD3WiNDR :P
1140105337 J * cdrx ~legoater@pixpat.austin.ibm.com
1140105456 M * Bertl wb cdrx!
1140105516 M * cdrx hi bertl
1140105563 M * cdrx i'll have to ask you a few questions about what you call lightweight context without init process
1140105604 M * cdrx is there a page on the wiki ?
1140105799 M * mnemoc see the first element of the topic
1140105815 M * mnemoc but don't doubt to ask 
1140105829 J * shuri ~shuri@64.235.209.226
1140106029 M * ebiederm Bertl: Did you see my comment on mugwumps tutorial?
1140106300 M * Bertl not yet ...
1140106315 M * Bertl ebiederm: where did you comment?
1140106339 M * ebiederm Just in the irc log when everyone was asleep.
1140106352 M * Bertl ah, cool :)
1140106385 M * ebiederm The point I wanted to make is when fetching a branch from a remote repository mugwump described the hard way to do it.
1140106418 M * ebiederm Basically git-fetch <repository> <remote-branch>:<local-branch> doesn't require any setup ahead of time.
1140106429 M * ebiederm And then you can just do git-checkout <local-branch>
1140106460 M * ebiederm Using git-pull in that circumstance is not what you want if you are not ready to merge yet.
1140106509 M * ebiederm Since you had problems I figured I would mention there was a less error prone way.
1140107079 M * Bertl yes, thanks a lot!
1140107578 J * stefani ~stefani@superquan.apl.washington.edu
1140107609 J * matti matti@linux.gentoo.pl
1140108147 Q * shedi Quit: Leaving
1140109503 J * bonbons ~bonbons@83.222.39.180
1140110039 Q * prae Quit: Execute Order 69 !
1140110159 J * Viper0482 ~Viper0482@p54977A56.dip.t-dialin.net
1140110200 M * Bertl welcome bonbons, Viper0482!
1140110213 M * bonbons Hello Bertl!
1140110524 J * seljacina_ ~seljak@dlp-100.as3.sr.bih.net.ba
1140110533 P * seljacina_ 
1140111186 M * Bertl okay, off for dinner ... back later
1140111201 N * Bertl Bertl_oO
1140112199 N * Bertl_oO Bertl
1140112431 M * ebiederm Bertl: Quick question.  What kind of access permissions do you use for entering a guest?
1140112506 M * ebiederm I.e  Under what conditions do you currently allow entering a guest?
1140112509 M * Bertl currently it's two flags/caps
1140112517 M * Bertl one is CAP_CONTEXT (bcaps)
1140112545 M * ebiederm ok.
1140112562 M * Bertl and the other one marks a context as non-enterable
1140112562 M * Roey oh
1140112565 M * Roey there's a CAP_CONTEXT?
1140112566 M * Roey heh
1140112590 M * ebiederm I am beginning to digest all of this.
1140112606 M * ebiederm And it occurs to me there is precedent in bproc from the clustering world.
1140112619 M * Bertl VXF_INFO_PRIVATE
1140112633 M * ebiederm rfork isn't exactly the same a enter but it is close enough that comparisons are valid.
1140112700 M * ebiederm bproc currently uses normal unix permissions  (on the node in this case) to control that.
1140112719 M * ebiederm Normal unix permission run into problems when you get the same uid meaning different things.
1140112769 M * ebiederm So before I can talk security I probably need to look and see what CLONE_NEWUID really means :)
1140112774 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1140112788 M * matta hey bertl
1140112833 M * ebiederm Thanks for the info.
1140112836 M * Bertl hey matta! LTNS!
1140112869 M * Bertl ebiederm: np at all
1140112883 M * matta same
1140112887 M * matta hey
1140112888 M * matta question
1140112904 M * matta i'm being stupid and can't find this on the site
1140112920 M * matta what's the best patch I can find for 2.6.12.x (6)
1140112921 M * matta ?
1140112942 M * matta hrm, i might just try the 2.6.14 one
1140112947 M * Bertl hmm, why would you want one for 2.6.12?
1140112988 M * Bertl patches for the upcoming stable and devel release are available for 2.6.15 and 2.6.16
1140112997 M * matta because I need to test it under xen
1140113019 M * Bertl and xen has no guest patches for 2.6.14+?
1140113027 M * Bertl did they stop developing it?
1140113039 M * matta they have bleeding edge 2.6.16-rc3, but that is in their -unstable tree
1140113052 M * Bertl ah, what about 2.6.15.4 ?
1140113059 M * matta this is for users to be able to run vservers in xen VM's and the stable is 2.6.12.6
1140113078 M * matta plus I want to start experimenting with vserver again myself :)
1140113093 M * Bertl i.c., well let me see regarding 2.6.12.6
1140113115 M * matta i'm just wondering if the 2.5.14 pathc would apply
1140113121 M * matta i'm gonna try a dry run patch
1140113137 J * liquid3649_ ~Viper0482@p54977FBA.dip.t-dialin.net
1140113156 M * Bertl on 2.6.12? no way
1140113172 M * matta oof
1140113175 M * matta yeah, LOTS of rejects
1140113183 M * matta much more fuzz/offset
1140113190 M * matta can't find file to patch at input line 1963
1140113190 M * matta Perhaps you used the wrong -p or --strip option?
1140113190 M * matta The text leading up to this was:
1140113190 M * matta --------------------------
1140113190 M * matta |diff -NurpP --minimal linux-2.6.14.3/drivers/infiniband/core/uverbs_mem.c linux-2.6.14.3-vs2.01/drivers/infiniband/core/uverbs_mem.c
1140113190 M * matta |--- linux-2.6.14.3/drivers/infiniband/core/uverbs_mem.c        2005-10-28 20:49:23 +0200
1140113190 M * matta |+++ linux-2.6.14.3-vs2.01/drivers/infiniband/core/uverbs_mem.c 2005-12-13 18:08:07 +0100
1140113192 M * matta and then that
1140113193 M * Bertl http://www.13thfloor.at/vserver/s_rel26/v2.0/patch-2.6.12.4-vs2.0.diff.bz2
1140113206 M * Bertl that's the best I can suggest for 2.6.12
1140113213 N * ebiederm ebiederm_oO
1140113221 M * Bertl of course you can patch up with the fixes done since
1140113222 M * matta so it's missing the 2.0.1 fixes
1140113237 M * Bertl http://vserver.13thfloor.at/Devel/
1140113255 M * matta that was clean
1140113258 M * matta just a few offsets
1140113264 M * matta btw how is ngnet going?
1140113295 M * matta oh, and I saw the OVZ vs IBM/VServer thread on LKML... LOL
1140113304 M * matta it seems you don't mind working with the IBM guys
1140113325 M * matta but OVZ admitting their approach is wrong is a significant hit for them... who knows what else could be "done better"
1140113328 M * Bertl well, I don#t mind working with OpenVZ either, just there is no working with yet ...
1140113403 M * matta ngnet?
1140113411 M * matta still proof of concept or what?
1140113412 M * Bertl well, ngnet is on hold again
1140113431 M * Bertl we are trying to sync with what ebiederm is doing there
1140113454 M * matta gotcha
1140113482 M * Bertl neverthess a view steps in the ngnet direction already took place
1140113491 M * matta i'm reading the howto
1140113507 M * Bertl (mainly in the way how the existing networking was restructured)
1140113524 M * Bertl btw, the most recent improvement is the scheduler
1140113570 Q * Viper0482 Ping timeout: 480 seconds
1140113600 M * matta i think i read something about that
1140113625 M * matta makes it a 'fairshare'
1140113652 M * Bertl yeah, well, it is much more flexible than that, but yes, thats one feature
1140114152 J * dothebart ~willi@xdsl-213-196-255-48.netcologne.de
1140114595 Q * tudenbart Ping timeout: 480 seconds
1140114873 M * mef bertl: our tests of mapping a vserver to CFQv2 don't show the type of fair queuing we were after.
1140114883 M * mef bertl: maybe it is better with CFQv3.
1140114896 M * mef bertl: in the mean time, we've been trying to learn how Xen does fair disk i/o sharing.
1140114913 M * mef bertl: it might be in there somewhere, but we cannot find it right now.
1140114939 M * mef bertl: all that we see is that the virtual block device daemon batches incoming block r/w requests by 16 per domain.
1140114963 M * mef bertl: and the host domain uses the anticipatory I/O scheduler.
1140115013 M * mef bertl: unless there is some other fancy logic that we couldn't find, I think that is all that Xen (or more precisely the host domain) is doing.
1140115072 M * matta xen creates 2 threads per domain
1140115074 M * mef bertl: makes me wonder whether putting something like the new TB scheduler in front of the A-I/O scheduler will give us something similar or better.
1140115077 M * matta xvd thread
1140115119 M * matta you can manipulate the pid's of the xvd threads, it uses both the nice and ionice values for priority just like CFQ
1140115125 M * Bertl mef: well, IMHO Xen is cheating in several ways ...
1140115127 M * matta CFQ should be used as the scheduler in domain 0
1140115149 M * mef matta: by default Xen 2 and Xen 3 dom0 use A-IO.
1140115177 M * mef Or at least whatever we pulled down.
1140115181 M * matta specify elevator=cfq
1140115186 M * matta xen 2 has no disk I/O QoS
1140115188 M * Bertl mef: until you can explain why running a dd aside of your db tests speeds up xen, we are in the dark ...
1140115190 M * matta er, well, it does
1140115192 M * matta it's round robin
1140115200 M * mef bertl: agreed.  
1140115211 M * mef matta: RR is what we are seeing.
1140115225 M * matta specify elevator=cfq on kernel line under xen 3.0
1140115241 M * matta then get ionice and try changing the priority of your domains xvd threads
1140115246 M * mef bertl: I suspect that a single guest domain does not get the full capacity available to the disk.  Hence there is no difference when you add a second domain to the mix.
1140115252 M * Bertl mef: the CFQ tests remind me of early scheduler tests
1140115254 M * matta root     17045  0.0  0.0      0     0 ?        S<   Feb15   0:00 [xvd 153 fd:1d]
1140115254 M * matta root     17320  0.0  0.0      0     0 ?        S<   Feb15   0:00 [xvd 153 fd:1c]
1140115268 M * matta ie. that is a thread (per CPU) for domain id 153
1140115269 M * mef bertl: of course, we are using the A-I/O and not CFQ in our Xen setup.
1140115287 M * Bertl we certainly could add some 'hacks' to make the statistics better, but this will harm overall performance
1140115290 M * mef matta: dom0 creates separate threads for each client domain?
1140115303 M * matta that's what i've been saying
1140115310 M * matta you don't see the xvd threads in dom0?:
1140115319 M * matta might be because you didn't enable cfq as the default
1140115328 M * mef matta: don't have xen booted on the box right now.
1140115331 M * mef will check later.
1140115368 M * mef matta: we'll have to try CFQ with Xen, as the A-I/O results (as Bertl points out) are odd for sure.
1140115370 M * Bertl mef: what would be interesting is, which setup gets _more_ I/O done in total
1140115380 M * mef bertl: you know the answer to that! :)
1140115391 M * matta Bertl: who is mef?
1140115407 M * Bertl matta: mef is Marc from princeton
1140115411 M * matta ok
1140115412 M * mef matta: http://www.cs.princeton.edu/~mef
1140115414 M * matta just wondering
1140115445 M * mef matta: we use vserver on 600+ machines in 300+ sites in 30 countries as part of PlanetLab.
1140115452 M * matta ok so he's one of the princeton people who have taken some serious resources to assist in vserver development
1140115468 M * mef matta: bertl still does the bulk of the lifting.
1140115513 M * mef matta: while all of our work is freely available, we have not pushed it back to the vserver community (yet).
1140115529 M * matta do you have your work listed anywhere?
1140115531 M * Bertl the princeton guys are so kind to test all the variations and pathes I decide _not_ to take (like ckrm), just to prove me right :)
1140115542 M * mef bertl: touche
1140115558 M * matta still on the ckrm kick some people are?
1140115567 M * matta I thought that idea was ditched over a year ago :)
1140115573 M * Bertl but seriously, they invited me there and we had a jolly good (and very productive) time
1140115594 M * Bertl the new scheduler is the result of talks with Andy there
1140115597 M * mef matta: we used ckrm for about a year and, well, we no longer do.
1140115611 M * matta should've invited me, i'm in philly suburbs :)
1140115630 M * Bertl well, if we had known, we had visited you
1140115650 M * Bertl matta: see that's what you get when you do not pay a regular visit here
1140115658 M * matta i'm always on network
1140115662 M * matta this just isn't on auto-join
1140115671 M * Bertl well, my fault? :)
1140115688 M * matta ok, you're on auto-jopin
1140115797 M * mef have someone on the phone.. will be back in a few minutes
1140115802 M * Bertl np
1140115999 Q * bonbons Quit: Leaving
1140116059 J * dearaujo ~dan@cpe-66-25-189-193.austin.res.rr.com
1140116080 J * shedi ~siggi@inferno.lhi.is
1140116224 M * pflanze Hi Bertl
1140116294 M * Bertl hey pflanze! LTNS!
1140116307 M * pflanze Bertl: have you seen my question "Has anyone run UML inside vserver?" above? I've got problems getting the uml networking to run (are unix sockets protected by vserver?). And maybe even more problems (hanging during init startup process).
1140116312 A * Bertl .o( seems all the 'old' chaps come visit today :)
1140116330 M * pflanze "ah ja?" :)
1140116344 M * Bertl pflanze: vserver in UML should be fine (out of the box)
1140116359 M * pflanze No, UML in vserver.
1140116367 M * Bertl pflanze: UML in vserver should work as long as you do not use the kernel modules and/or networking
1140116377 M * pflanze I do not use skas.
1140116381 M * Bertl the networking will require TUN devices
1140116391 M * pflanze For networking I use the uml_switch daemon and tun/tap/
1140116405 M * pflanze my first attempt has been to run the uml_switch on the host.
1140116425 M * pflanze Which "works" insofar that uml_switch can create the tap0 device and successfully ~binding it.
1140116436 M * pflanze It listens on a unix socket.
1140116458 M * pflanze I bind mount the directory with that socket into my vserver, and then tell the uml linux process to use that socket. 
1140116471 M * pflanze Which works insofar that it seems to open that socket, but gets a permission denied.
1140116485 M * Bertl yup, security
1140116498 M * pflanze I can even do "cat > /path/to/that/socket" and am getting permission denied. 
1140116529 M * pflanze So that is very strange. Anything vserver is doing to protect unix domain sockets with a process from ctx 0 from being accessed by processes from other contexts?
1140116546 M * Bertl yep, that's called socket isolation
1140116550 M * pflanze ah.
1140116577 M * pflanze what is it good for..? considering that usually socket files are protected anyway from being accessed.
1140116587 M * pflanze And how can it be worked around?
1140116636 M * Bertl well, I'm currently checking the code ...
1140116691 M * pflanze I've read on the uml homepage that uml_switch is checking the uid of the connecting processes, as sort of authentication.
1140116707 M * pflanze But I've double checked to make sure the same uid's are used on the host and the guest.
1140116746 A * pflanze dares to question why uml_switch does not simply put it's socket file into a 0700 directory owned by the user in question..
1140116867 M * Bertl hmm, could you strace a write to the socket from inside?
1140116899 M * Bertl something like 'strace -fF dd if=/dev/null of=/socket
1140116916 M * pflanze ok, will do.
1140116926 M * Bertl (although this is probably not working on the host anyway)
1140117015 M * pflanze Since the box in question has just vanished from the network, I'll have to take on a journey there. So I'll report back later (hopefully soon..).
1140117045 M * Bertl k
1140117800 Q * mnemoc Read error: Connection reset by peer
1140118027 J * Smutje_ ~Smutje@xdsl-84-44-184-101.netcologne.de
1140118135 Q * Smutje Ping timeout: 480 seconds
1140118135 N * Smutje_ Smutje
1140118160 M * mugwump ebiederm_oO: thanks for the feedback re: git-pull vs git-fetch
1140119674 Q * pflanze Quit: later
1140119686 Q * mkhl Quit: 
1140120615 Q * hallyn Ping timeout: 480 seconds
1140120746 J * mkhl mkhl@200-148-40-184.dsl.telesp.net.br
1140120847 J * hallyn ~xa@c-24-11-243-196.hsd1.in.comcast.net
1140121147 M * Bertl wb mkhl! hallyn!
1140122473 M * hallyn hey
1140122490 M * hallyn dumb cable modem
1140122566 M * mugwump Hey Hallyn
1140122800 Q * gerrit Ping timeout: 480 seconds
1140123067 Q * ntrs Ping timeout: 480 seconds
1140123207 M * mugwump Bertl: did you get a chance to look at the patch stack I've been making?
1140123215 M * Bertl nope
1140123249 M * mugwump ok, well today I'm going to see if I can't make it work :)
1140123276 M * mugwump Then I think it might be ready for the first round of savaging
1140123542 J * Hatei ~only@kw-cpe-062249079079.kabelweb.at
1140123749 Q * Hatei Remote host closed the connection
1140124257 Q * SNy Read error: Connection reset by peer
1140124259 J * SNy b81fb74c94@bmx-chemnitz.de
1140124755 M * hallyn mugwump: savaging meaning it's going to lkml?
1140124759 M * mugwump :)
1140124845 M * mugwump that's the one.  Basically up to about here or so: http://vserver.utsl.gen.nz/gitweb/?p=vserver.git;a=shortlog;h=74311f5664c380048bd203bdbc9b8561c5bd927d
1140124863 M * mugwump except with an iso-8859-1 encoding instead of the utf-8 there that is broken
1140125030 J * gerrit ~gerrit@bi01p1.co.us.ibm.com
1140125159 M * hallyn bertl: very good points about the security issues.  thanks for pointing those out.
1140125331 M * Bertl my pleasure!
1140125669 M * brc_ bertl is that you on wth_linux_vserver_140 ?
1140125720 J * pflanze ~chris@unk-111.ethz.ch
1140125735 M * brc_ nice presentation, congratulations.
1140125921 Q * liquid3649_ Quit: bin raus, 
1140125991 J * Aiken ~james@tooax8-223.dialup.optusnet.com.au
1140126031 M * Bertl brc_: I guess so :)
1140126074 M * mugwump I found that file, but it's 400MB ... what's the bitrate?
1140126288 M * mugwump ah, 47:41 total time, so about 1.1mbit.  Anyone have a reencoding of it at a smaller bitrate, and perhaps deinterlaced?
1140126337 J * ntrs ~ntrs@68-188-37-15.dhcp.stls.mo.charter.com
1140126465 M * dhansen ebiederm_oO: Bertl: (and anybody else) will you be attending OLS?
1140127206 M * mugwump hmm, looks interesting.  I missed the last YAPC::NA in Ca
1140127307 Q * SNy Ping timeout: 480 seconds
1140127398 M * Bertl dhansen: unless somebody sponsors it, no :/
1140127572 M * Skram Hi all
1140127679 J * SNy fbe82c609c@bmx-chemnitz.de
1140127765 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk
1140127811 M * Bertl wb SNy! welcome vrwttnmtu!
1140127821 M * vrwttnmtu Hey Bertl 
1140128247 J * brc__ bruce@20151181056.user.veloxzone.com.br
1140128294 M * Skram how can one reboot a vps? when i type restart inside a vps it shutdowns and never comes back
1140128321 M * Bertl that's not expected behaviour
1140128351 M * Bertl kernel, patch and tool versions please
1140128363 M * brc__ BitchX-1.1-final+ by panasync - Linux 2.6.15.4
1140128366 M * brc__ patched :)
1140128405 Q * brc_ Ping timeout: 480 seconds
1140128460 N * brc__ brc
1140128530 M * matta bertl, still around?
1140128538 M * Bertl yup
1140128554 M * matta you had mentioned before someone completed the cowlinks 
1140128562 M * matta is this part of vserver now... or?
1140128601 M * Bertl hmm, well, yes, that was me, I guess :)
1140128625 M * matta is there documentation on it?
1140128640 M * Bertl the source code ...
1140128698 M * matta hah
1140128702 M * matta ok
1140128703 M * matta well
1140128708 M * matta let's do a short rundown
1140128729 M * matta by applying the latest vserver patch cowlinks are already 'installed' ?
1140128736 M * Bertl http://www.13thfloor.at/vserver/d_rel26/v2.1.0/split-2.6.14.4-vs2.1.0/37_2.6.14.4_cow.diff.hl
1140128822 M * matta hrmn
1140128827 M * dearaujo my apologies - what are cowlinks?
1140128835 M * matta so on a vxd enabled mount
1140128835 M * Skram can anyone help me with apache stuff?
1140128844 M * matta cowlinks will replace hardlinks ?
1140128887 J * LittleJ0hn ~joaopaulo@201-1-12-16.dsl.telesp.net.br
1140128890 M * Bertl dearaujo: well, that is how matta calls it .. the thing is, it's nothing special, CoWLB means that hard links (as used for unification) will get copied on write
1140128904 M * Bertl welcome LittleJ0hn!
1140128917 M * Bertl Skram: if it is linux-vserver related :)
1140128919 M * matta I am assuming vxd is the new name for tagxid ?
1140128939 M * LittleJ0hn tkx
1140128956 Q * Doener Quit: Leaving
1140128963 J * Doener doener@i5387CF5E.versanet.de
1140128969 M * Bertl matta: you assume wrong :)
1140129026 M * matta what is vxd then? this is new
1140129106 M * matta it seems to work well?  only a single inode is cached?
1140129118 M * pflanze hm, applying patch-2.6.14.3-vs2.01.diff.gz to 2.6.15.4 doesn't look easy.
1140129127 M * pflanze Is anything newer around?
1140129155 M * Skram Bertl: so do guests get access to swap or what?
1140129158 M * matta http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.0.1.3.diff ?
1140129169 M * Skram ?
1140129187 M * Bertl pflanze: http://vserver.13thfloor.at/Experimental/
1140129211 M * Bertl Skram: yes, they do ..
1140129212 P * dearaujo 
1140129223 M * Skram Bertl: Oh, Okay...
1140129232 M * Skram But if they have a hard limit and no soft limit, they dont?
1140129238 M * Skram because my kernel doesnt do soft limits, arg
1140129263 M * Bertl this is independant, but hard is hard :)
1140129275 M * Skram so if there are no limits, it uses swap/
1140129301 M * Bertl it might even use swap when there are limits
1140129307 M * Skram how does decide to use swap?
1140129314 M * Skram im having major memory shortages
1140129332 M * Bertl but once the limit is reached, then there is nothing to decide or swap, it's OOM
1140129345 M * harry oom sucks!
1140129373 M * harry it killed my sshd, so i couldn't choose what to kill anymore
1140129382 M * Bertl well, with a proper setup, you'll never reach that
1140129400 M * harry Bertl: with proper setup for your vservers
1140129416 M * harry but thta doesn't mean you can't have problems with your host when something goes berzerk
1140129446 M * Bertl that's why everybody aware of the vm-issues uses strict-no-overcommit
1140129460 M * Bertl (at least on servers you care about :)
1140129481 M * harry hehe, important sidenote ;)
1140129637 P * LittleJ0hn 
1140129771 Q * comfrey Remote host closed the connection
1140130008 Q * lilalinux Remote host closed the connection
1140130182 J * Roll Roll@host24.200-45-159.telecom.net.ar
1140130185 M * Roll hi
1140130258 P * Roll 
1140130426 M * Bertl hmm ...
1140130631 M * Doener no answer within 5 seconds, there can't be anyone alive here ;)
1140130640 M * SNy hrhr
1140130642 M * SNy indeed
1140130655 M * SNy he's gotta roll on, see
1140130658 M * SNy ;p
1140131918 M * Psy0rz_ :)
1140132009 M * dhansen Bertl: how well does vserver run on ppc64?  If you had a machine, would it help?
1140132207 M * daniel_hozac http://archives.linux-vserver.org/200601/0040.html
1140132295 M * dhansen hehe
1140132296 M * dhansen nice
1140132473 Q * ntrs Quit: Leaving
1140132520 J * matt1 ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1140132551 M * Bertl dhansen: a test machine would help, of course
1140132566 M * Bertl dhansen: a test machine with somebody testing even more :)
1140132820 M * Skram I want soft limits :(
1140132825 M * Skram But I dont want to touch the kernel :(
1140132855 Q * matta Ping timeout: 480 seconds
1140132899 M * Bertl Skram: set the hard limit high enough, and pretend the soft limit is lower :)
1140132913 M * daniel_hozac lol.
1140132914 M * dhansen Bertl: once we're using common code, it will get tons of testing :)
1140132924 M * Skram Bertl: what do you mean?
1140132939 M * Skram but see, i dont want someone to not be able to ssh in if their mem is maxed
1140132986 M * Bertl Skram: what you want is nice, but impossible ...
1140133010 M * pflanze Bertl: sorry for me crying for help: the problem with my uml socket has simply been the oversight that I was member of the group, but the socket was srwxr-xr-x.
1140133017 M * Bertl Skram: you want a glass, which, when full, can still take another bottle of wine :)
1140133028 M * Skram haha
1140133044 M * dhansen Bertl: I was thining about proposing a Containers/VServer/VPS BOF at OLS.  
1140133058 M * Bertl dhansen: yeah, I read that :)
1140133079 M * dhansen Bertl: I couldn't remember who I told :(
1140133095 M * SNy wow, Roll is Teh Man
1140133112 M * Bertl pflanze: so everything fine, not vserver related, yup?
1140133117 M * SNy he managed to come up for a few seconds in a handful of other channels
1140133142 M * pflanze Bertl: well no, the uml still hangs during bootup. Dunno why. I'll try to run it on the host now to check.
1140133162 M * Bertl strace might be your friend there too
1140133167 M * pflanze But at least networkig now works (with uml_switch running on the host).
1140133202 M * pflanze (I can boot the uml with init=/bin/bash and it works "just fine")
1140133609 M * Doener mugwump: just to see how to use mencoder, i recoded the wth video. progressive, 139MB, 384kbit video, 64kbit audio. interested?
1140133627 Q * wasser Ping timeout: 480 seconds
1140133637 M * Skram does ps aux show ram percents of the total ammount being used or the limit (eg. 100megs is used, but the vserver has lets say.. 156mb limit)
1140133649 M * mugwump Doener: actually my download has already finished, but thanks anyway :)
1140133656 M * Doener ok