1164499227 J * lilalinux_ ~plasma@dslb-084-058-198-224.pools.arcor-ip.net
1164499659 Q * lilalinux Ping timeout: 480 seconds
1164502724 M * Piet Bertl: could you help me with extending the size of a vservers' tmpfs without rebooting it?
1164502738 M * Piet s/extending/increasing/
1164502751 M * Bertl hmm, not sure that is possible, does tmpfs honor remount options there?
1164502786 M * Bertl if so, try: 'vnamespace -e <xid> -- mount -o remount,size=32M /path/to/vserver/tmp/
1164502790 M * Piet i tried to do "mount -o remount /tmp" in the vserver, but it wouldn't work
1164502805 M * daniel_hozac from the inside you'll need secure_remount
1164502931 M * Piet is secure_remount contained in some .deb ?
1164502950 M * Piet vserver-utils?
1164502976 M * doener that's a ccap ;)
1164503015 M * Bertl Piet: just try the line I gave above, adjust the options
1164503189 M * Piet mount: can't find /var/lib/vservers/vserver1/tmp in /etc/fstab or /etc/mtab
1164503210 M * Piet woudl it help to put it in there?
1164503273 M * Piet the /var/lib/vservers/vserver1/tmp directory does exist if that's of any help
1164503324 M * daniel_hozac try mount -n.
1164503379 M * Bertl ah, yeah, still forget that mount is trying to be smart
1164503496 M * Piet just running "mount -n" gives a list, but does not list the tmpfs of the vserver
1164503521 M * daniel_hozac and it shouldn't.
1164503546 M * daniel_hozac -n suppresses the mtab modification.
1164503566 M * Piet "vnamespace -e 6667 -- mount -n -o remount,size=32M /var/lib/vservers/vserver1/tmp" gives the same error as before
1164503571 M * daniel_hozac my mount seems to require the path specified twice though.
1164503698 M * daniel_hozac (i.e. vnamespace -e <guest> mount -n -o remount,size=64m /path/to/<guest>/tmp{,} works fine)
1164503788 M * Piet that gives no error, but doesn't change it
1164503820 M * ntrs About the mysql problem with 2.6.18.X. It seems it's fixed by "export LD_ASSUME_KERNEL=2.2.5". Confirmed on several different redhat based virtual servers.
1164503824 M * Piet within the vserver "mount | grep /tmp" gives "none on /tmp type tmpfs (size=16m,mode=1777)"
1164503834 M * Piet 16m is the old size
1164503857 Q * Borg- Ping timeout: 480 seconds
1164503881 M * ntrs Also, it seems it affects all redhat based distros, RH 7.3, 8, 9, FC1-5, CentOS 3 and 4
1164503979 Q * ||Cobra|| Server closed connection
1164503992 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1164504025 M * Bertl Piet: check with grep /tmp /proc/self/mounts (inside)
1164504047 M * Bertl Piet: the rpoblem always is that mount relies on /etc/mtab too much
1164504061 M * Bertl ntrs: that might point to PTL vs Linux Threads
1164504079 M * Bertl ntrs: and possibly the redhat glibcs
1164504101 M * Bertl daniel_hozac, doener: what do you think of that?
1164504109 J * Piet_ hiddenserv@tor.noreply.org
1164504110 M * ntrs yes it could be but it was working fine until 2.6.17
1164504153 M * Piet_ Bertl: none /tmp tmpfs rw 0 0
1164504164 M * Piet_ last line i saw was this: <Bertl> Piet: check with grep /tmp /proc/self/mounts (inside)
1164504179 M * ntrs Also, it seems that distro supplied mysql and mainline mysql are different. The mainline mysql RPMs do not show this problem.
1164504183 Q * Piet Killed (NickServ (GHOST command used by Piet_))
1164504196 M * ntrs Maybe they are compiled with different threading.
1164504201 N * Piet_ Piet
1164504221 M * Bertl ntrs: do you, by any chance, know which kernels the mentioned distros use?
1164504228 M * Bertl (by default, I mean)
1164504239 M * ntrs I have no idea for all of them
1164504275 M * ntrs the latest CentOS 4 kernel is vmlinuz-2.6.9-42.0.3.ELsmp
1164504291 M * daniel_hozac RHL8 is when NPTL was introduced.
1164504326 M * doener there were some userspace NPTL bugs in glibc, but I wonder how they could eat a socket
1164504349 M * doener or did I misinterpret the "disappearing scoket"?
1164504382 M * Bertl I'm still clueless what actually happens
1164504388 M * daniel_hozac Piet: check with df -h inside.
1164504390 J * jack ~jack@xplr-ts-t11-208-114-141-115.barrettxplore.com
1164504399 M * Bertl heya!
1164504404 M * jack hello
1164504407 M * Bertl welcome jack!
1164504421 M * jack thanks!
1164504439 M * Bertl let's welcome the original founder of linux-vserver :)
1164504455 M * doener ntrs: do you have any data concerning the problem?
1164504460 M * Piet daniel_hozac: oh, right, that's 64 m now
1164504462 M * jack It is pretty late for you
1164504479 M * Bertl jack: np, my work time is almost us time
1164504488 M * Piet daniel_hozac: so i just ignore the 'mount' output?
1164504502 M * daniel_hozac Piet: yes, as Bertl said, mount lies all the time.
1164504502 M * jack what can I do for you ?
1164504510 M * Bertl jack: as you probably figured, we have some issues regarding the domain
1164504537 M * Bertl jack: about two or three months ago, the domain itself expired, and we could not reach you
1164504561 M * Bertl jack: powerfox (a community member) payed the fee, IIRC
1164504591 M * Bertl jack: a week after that, linux-vserver.org (the vps) disappeared
1164504615 M * jack whois claim it is active until 2007
1164504629 M * Piet ah, good to know :)
1164504635 M * Bertl jack: we assumed you had some severe issues, but that wasn't a big problem
1164504651 M * Bertl jack: yes, as I said, powerfox payed the fee for 2006
1164504666 M * Piet now, is there a known issue with postfix + sasl in vservers?
1164504672 M * Bertl jack: we were moving the wiki to mediawiki anyway
1164504694 M * Bertl jack: but we still have issues with the DNS
1164504701 M * jack I saw you did a lot of rework on the web site
1164504715 M * Bertl jack: your both nameservers do not accept name service information from the master
1164504735 M * Bertl also we would like to list at least five nameservers there (glue)
1164504765 M * Bertl currently we are in a state where only one listed nameserver shows the correct info
1164504777 M * Bertl (which is bad :)
1164504838 M * Bertl jack: so we would like to get some administrative record for the domain, or, maybe, transfer the domain
1164504856 M * jack I have no problem doing that (transfer)
1164504885 M * Bertl okay, ntrs, IIRC you said you are willing to take the domain, right?
1164504892 M * jack I can check why the current secondary are not working though. I have administrative rights there
1164504895 M * ntrs Sure, yes
1164504938 M * doener ntrs: no data? :(
1164504951 M * Bertl jack: okay, can you adjust the nameserver data too?
1164504965 M * Bertl ajcor is it easier to transfer and do that then?
1164504973 M * ntrs doener, no data, but basically mysql stops responding after a while. all connections are in a socket wait state or something similar
1164504978 M * Bertl s/ajcor/or
1164504994 M * ntrs we can transfer right now if it's ok with Jack
1164505003 M * jack No problem
1164505012 M * doener ntrs: how easy is it to reproduce that? did you try to attach strace to the mysql threads?
1164505029 M * ntrs doener, no, this is just what customers have reported.
1164505031 M * Bertl okay, jack, ntrs, please exchange the necessary info
1164505050 M * Bertl jack: all those issues aside, how is it going for you?
1164505065 M * ntrs jack, how do you want to proceed?
1164505103 M * jack I have been promoted IT manager with 17 people working for me. Since january. Very difficult at first. Very very (at least for me)
1164505139 M * jack We are still using lots and lots of vservers :-)
1164505141 M * ntrs jack, should I just initiate a transfer?
1164505174 M * jack Yes. do so. I guess all I have to do is acknowledge an email from gandi
1164505188 M * Bertl jack: great! btw, you added some dhcp config stuff to your tools, IIRC?
1164505237 M * jack Yes. (dhcp) works wonderful. Our QA departement is creating 5-10 new vservers per week. It helps a lot
1164505240 M * Bertl jack: btw, would be really nice to see you every now and then here on the channel
1164505245 M * ntrs jack, yes, you should only need to approve the transfer
1164505287 M * Bertl jack: are there any special things to consider there? i.e. it might be itneresting to integrate that in util-vserver ...
1164505288 M * jack Since one year, I have a good internet connectivity (satellite) from home. Before that, it was only modem. Not fun with irc (hanging up often)
1164505334 M * Bertl jack: I was on dial up modem until 3 months ago, I know ...
1164505335 M * jack For dhcp, there is a special recipy. I even works if the master itself is doing dhcp. Quite cool.
1164505362 M * daniel_hozac jack: how do you handle IP address changes?
1164505365 M * Bertl jack: okay, would be great if you could have a chat with daniel_hozac or ensc 
1164505366 M * jack Note that the later vserver utility from me is now on subversion at svn.solucorp.qc.ca/repos/solucorp/vserver
1164505380 M * ntrs jack, I initiated a transfer
1164505390 M * ntrs Is your jack@solucorp.qc.ca still ok?
1164505419 M * jack We don't handle IP change. We did not address that.
1164505435 M * daniel_hozac ah.
1164505443 M * jack To handle that, we would need to hack the vserver stuff in the kernel.
1164505453 M * daniel_hozac hmm?
1164505458 M * jack Note that IP address change when using dhcp is not that common
1164505465 M * daniel_hozac recent kernels should support that just fine.
1164505475 M * daniel_hozac util-vserver 0.30.212-rc1 even supports changing the addresses assigned.
1164505485 M * daniel_hozac (naddress --nid ... --set --ip ...)
1164505493 M * ntrs jack, I guess you will need to acknowledge some email at jack@solucorp.qc.ca
1164505509 M * jack I toucht that once a vserver was up, it was kind of locked configuration wise
1164505520 M * daniel_hozac not at all, at least not for 2.6 kernels.
1164505524 M * jack indeed (I will check my emails)
1164505552 M * daniel_hozac i realize that IP address changes aren't very common, but i still think they'd need to be handled.
1164505554 M * jack for 2.6 kernel, I remember that once you have complete the context setup, it is locked after that.
1164505561 M * doener daniel_hozac: hm, what about the single ip bind to 0.0.0.0 special case?
1164505576 M * daniel_hozac doener: right, that's still a problem.
1164505583 M * Bertl doener: will go away, at least partially
1164505590 M * daniel_hozac yes, single_ip is supported in nflags ;)
1164505608 M * jack Quite frankly, the idea of having a vserver is to provide service. If your IP change in the middle, bye bye tcp connection
1164505624 M * daniel_hozac jack: not at all, you can change most things while it's still running.
1164505625 M * Bertl doener: but fixing that up would be possible too, as the sockets are tagged :)
1164505641 M * doener Bertl: yeah, just wanted to point that out, as daniel_hozac said that the 2.6 series would just work ;)
1164505653 M * jack Yes but the clients are unaware of IP changes
1164505654 M * Bertl jack: you have _no_ idea what vserver are used for nowadays :)
1164505658 M * daniel_hozac true, i forgot that detail ;)
1164505695 M * doener Bertl: going back a few days... you asked for something to track patches, right?
1164505718 M * Bertl jack: we had a setup called moreubuntu, which provides multiple desktop places on a single machine, with Linux-VServer :)
1164505779 M * daniel_hozac doener: yes please! you found/made something?
1164505792 M * jack I see (dhcp).. Our implementation has a script called by dh-client. I guess this script can handle reassignement
1164505798 M * Bertl doener: yep
1164505818 M * doener daniel_hozac: well, I just finished learning python basics, and would try to write something... 
1164505836 M * doener need to do something with python before I start to forget everything ;)
1164505842 M * daniel_hozac ah ;)
1164505878 M * daniel_hozac doener: mod_python, or would it be some non-HTTP thing?
1164505926 M * doener choose one, I'm open for ideas... non-http, soap, xmlrpc, mod_python, cgi, whatever ;)
1164505941 M * daniel_hozac as long as there is at least a web interface, i'm happy :)
1164505984 M * jack I have received the domain transfer request. I am processing it now
1164506001 M * Bertl jack: what kernels do you use at work?
1164506016 M * doener daniel_hozac: you... like... webinterfaces? o.O
1164506085 M * daniel_hozac doener: for tracking things, sure. it's simple and thus quite easy to get an overview.
1164506130 M * jack 2.6.10 and up to 2.6.17 with the latest vs2.0.x. On amd64
1164506152 M * Bertl great!
1164506233 M * daniel_hozac eyck_: btw, are you using the alpha utils at all? if so, could you give util-vserver 0.30.212-rc1 a spin and see if i broke chbind (or anything else for that matter)?
1164506283 M * doener Bertl, daniel_hozac: so, what would we need? Simple tracking of branch <-> patch with state (applied, not applied), description and maybe patch grouping (for patches that have versions for different branches)
1164506310 M * daniel_hozac sounds good to me.
1164506314 M * Bertl doener: yes, and in any case some kind of comments
1164506334 M * Bertl who provided the aptch, when, and why
1164506360 M * doener who and when is easy, why would go into the description I guess?
1164506366 M * Bertl and it would be great if we could move them between new, pending and applied
1164506390 M * Bertl applied would basically build the changelog then
1164506416 M * Bertl also, I'd like to have a commandline interface to add patches :)
1164506425 M * daniel_hozac yeah, i was thinking that too.
1164506432 M * doener sure. Is Python ok for you?
1164506438 M * Bertl definitely
1164506460 M * trippeh (Disclaimer: I havn't read the entire discussion) Are you guys reimplementing quilt? :)
1164506482 M * doener trippeh: more like distributed-quilt (ie. git ;)
1164506484 M * Bertl trippeh: kind of web quilt, yeah ...
1164506494 M * doener ehrm, stgit
1164506520 M * doener with the "st" part also being distributed... *goes hacking instead of blabbering*
1164506588 M * Bertl ntrs: how do you set the LD_ASSUME_KERNEL=2.2.5
1164506620 M * Bertl ntrs: it might be, that I can now observer the issue on something different than mysql
1164506625 M * ntrs export LD_ASSUME_KERNEL=2.2.5 in /etc/rc.d/rc.local
1164506640 M * Bertl okay, I'll try that for a service, sec
1164506641 M * ntrs or export LD_ASSUME_KERNEL=2.2.5 at the command line
1164506678 M * daniel_hozac Bertl: your named?
1164506688 M * Bertl seems so ...
1164506706 M * doener I had php-fgci (via lighttpd) hanging in a futex some time ago... LD_ASSUME_KERNEL fixed that as well
1164506720 M * daniel_hozac so 2.6.18 broke some glibc assumption(s)?
1164506729 M * doener but I've seen php hanging on vanilla kernels in the same way
1164506758 M * doener and both were 2.6.16 and earlier (never tried to remove the LD_ASSUME_KERNEL ;)
1164506863 M * daniel_hozac i guess it's time to find out what LD_ASSUME_KERNEL actually does...
1164506873 M * Bertl hmm, no, that was a false positive ... it works in both cases now .. no idea ...
1164506881 M * daniel_hozac hehe.
1164506891 M * Bertl daniel_hozac: IIRC, it changes the NPTL behaviour in glibc
1164506899 M * doener daniel_hozac: assume that the running kernel is that version and not use newer features?
1164506937 M * jack About the domain transfer. I am not a godaddy customer, so I can't perform the domain transfer this way
1164506938 M * daniel_hozac right, but what does that actually mean? i mean, what things does it actually disable, other than NPTL.
1164506948 M * doener http://people.redhat.com/drepper/assumekernel.html
1164506964 M * daniel_hozac ah, should've known there was a document there :)
1164507082 M * ntrs Jack which way do you think will work better?
1164507094 M * ntrs jack, what does the email say?
1164507136 M * jack It says "Dear godaddy customer"
1164507151 M * jack It points me to a web page where I must log as a customer
1164507153 M * daniel_hozac ntrs: have you tried LD_ASSUME_KERNEL=2.4.19?
1164507193 M * ntrs jack, I don't think you have to be a customer. You could simply register as a user I think.
1164507200 M * ntrs daniel_hozac, no I have not tried that
1164507262 J * MrX ~urk@219.95.238.23
1164507324 M * jack So I have to create an account ?
1164507389 M * ntrs jack, I don't know. I think you need a transfer Authorization Code from your domain registrar. Who is the current domain registrar?
1164507400 M * Bertl wb MrX!
1164507407 M * Bertl ntrs: gandi.net IIRC
1164507474 M * ntrs This seems to be the process and gandi.net
1164507474 M * ntrs http://www.gandi.net/faq/view_category/en/1.6/transferring_a_domain_name_from_gandi
1164507481 M * jack OK I think I got it. I have received an email with some transaction ID and security code. All I have to do is to forward this to the new domain owner, already godaddy customer
1164507524 M * jack So all I need is your email address. I will forward the email to you and you will complete the transfer. I guess that gandi will also contact me for confirmation later
1164507530 M * ntrs JAck, first you need to remove the Transfer Prohibited flag.
1164507582 M * ntrs jack, this is also the process at godaddy.com
1164507583 M * ntrs https://www.securepaynet.net/gdshop/pdf/TransferInsider.pdf
1164507910 M * ntrs jack, there seems to be some auth code that you will get from gandi.net that you will need to enter it at the web URL shown in the email you got from godaddy.com
1164508020 M * marcfiu hello
1164508023 M * marcfiu back for a few minutes
1164508048 M * marcfiu Bertl: the kick-fix-hack for __enter_* __leave_* seems to have done the trick on my system.
1164508066 M * daniel_hozac marcfiu: but it's likely to break some other things.
1164508082 M * Bertl marcfiu: yes, it was just a test, but sounds promising
1164508105 M * marcfiu daniel_hozac:  which is why it presumably is called the quick-fix-hack. :)
1164508137 M * marcfiu Bertl: so what are you trying to accomplish with the __enter* __leave* stuff?
1164508314 M * Bertl marcfiu: we have the problem that certain actions are taken from within an irq
1164508331 M * ntrs jack, I found the ultimate document about the transfer.
1164508332 M * ntrs http://products.secureserver.net/products/domain_transfers/transfer_domain_gandi.pdf
1164508349 M * ntrs It explains the domain transfer from gandi to godaddy
1164508359 M * Bertl marcfiu: some of them, should not happen inside a guest context, while others actually require the guest context
1164508415 M * marcfiu like deducting tokens from a guest upon clock intr. :)
1164508415 J * _dmax ~semaj@81.193.57.42
1164508427 J * s0undt3ch_ ~s0undt3ch@81.193.57.42
1164508448 M * Bertl marcfiu: precisely
1164508449 M * ntrs jack, did you remove the Transfer Prohibited flag from the domain?
1164508463 M * jack I just did few seconds ago
1164508485 M * ntrs Ok. It seems I will have to request the transfer again
1164508520 Q * sid3windr Ping timeout: 480 seconds
1164508542 M * ntrs jack, I reinitiated the transfer
1164508637 M * marcfiu Bertl: haven't mucked around at this low level in a while.
1164508671 M * marcfiu Bertl: is it safe to assump that irq=0 if its a clock intr. and you shouldn't be doing the __enter_* __leave_* calls?
1164508718 M * Bertl marcfiu: nah, what we actually have to do is eliminate the need for __enter/__leave in this context
1164508732 M * Bertl (and thus preserve the context tagging)
1164508760 M * Bertl we planned to do so for a while, but postponed it for whatever reason
1164508764 Q * dmax Ping timeout: 480 seconds
1164508772 N * _dmax dmax
1164508814 M * marcfiu I'm confused, I thought this was a recent bug.
1164508835 M * daniel_hozac the untagging is rather recent.
1164508845 J * sid3windr luser@bastard-operator.from-hell.be
1164508859 Q * s0undt3ch Ping timeout: 480 seconds
1164508859 N * s0undt3ch_ s0undt3ch
1164508903 M * jack The pdf document claim godaddy may take up to 24 hours to notice the unliocking of a domain transfer from gandi
1164508946 M * marcfiu daniel_hozac: that's what I thought
1164508948 M * ntrs jack, yes, that is correct. So I guess we can continue tomorrow. Can you come back to the channel sometime tomorrow after 6PM Central Time?
1164508951 M * jack I have received the new request from godaddy. Same thing as before. My understanding is that I must forward the email to you
1164508989 M * ntrs No, actually, when you get the auth code from gandi you will need to enter that info somewhere. There should be a URL in that email I think.
1164509007 M * jack Yes I can reconnect tomorrow. 
1164509044 M * ntrs jack, is there a transaction ID and security code in that email?
1164509046 M * jack The email was sent by godaddy. So the auth code must be entered there. I think gandi will perform their own request as well
1164509058 M * jack Yes there is a transaction ID and security code
1164509077 M * ntrs Ok. I need those.
1164509090 M * ntrs I need to enter them here.
1164509091 M * jack what is your email ?
1164509113 M * marcfiu jack and ntrs, what are you guys chatting about?
1164509123 M * daniel_hozac linux-vserver.org domain transfer.
1164509127 M * marcfiu oh
1164509129 M * marcfiu ok
1164509144 M * Bertl marcfiu: /whois jack
1164509150 M * ntrs jack, see in private.
1164509166 M * ntrs I sent it to you there.
1164509191 M * ntrs jack, did you see it there?
1164509193 M * marcfiu bertl: ok... let me know if need me to test something or whatever.
1164509208 M * Bertl marcfiu: will do so, we should have a few changes there shortly
1164509217 M * ntrs jack, or just send both values in private
1164509254 M * Bertl ntrs, jack: ctcp/dcc
1164509275 M * jack I have sent the whole email to ntrs
1164509279 M * ntrs ok
1164509285 M * Bertl okay, great!
1164509316 M * Bertl jack: maybe just update the 'master' for the dns servers for now, they should get their data from either ns1.13thfloor.at or ns1.linux-vserver.at
1164509341 M * ntrs jack, ok now I need the authorization code you get from Gandi.net
1164509548 M * ntrs jack, Log in to your Gandi account at www.gandi.net/admin/domain/change?l=en.
1164509562 M * ntrs On the account administration page, open the “Modify parameters of a domain
1164509563 M * ntrs name” tool.
1164509570 M * ntrs On the domain-modification page, you will find the authorization code on the
1164509570 M * ntrs right, just above your name and address.
1164509935 M * jack Ok, it is not exactly presented this way, but I have the auth code. Do you want it by email or in the private chat ?
1164510400 M * ntrs jack, it does not matter, either way is fine
1164510551 M * jack All secondary for linux-vserver.org have been updated to point to the new primary. They are now up2date
1164510637 M * ntrs jack, strange, I get "errors were detected" when I use that code.
1164510646 J * yarihm ~yarihm@84-74-17-6.dclient.hispeed.ch
1164510707 M * ntrs jack, actually it seems it worked now. My session timed out.
1164510730 M * jack good!
1164510734 M * ntrs jack, the status here now says "Pending Current Registrar Approval"
1164510752 M * ntrs I guess you will now get an email from gandi about this
1164510756 M * jack I guess a new email sits in may mail box. Let me check
1164510866 Q * yarihm 
1164510947 M * ntrs ok
1164510999 M * jack no mail yet
1164511040 Q * lilalinux_ Remote host closed the connection
1164511116 J * lilalinux ~plasma@dslb-084-058-198-224.pools.arcor-ip.net
1164511170 M * marcfiu Bertl: let me know via email if there is anything you' d like me to test.
1164511191 M * marcfiu Bertl: also, let me know whether the serial console worked for you from vici to pluster-20.
1164511194 M * Bertl marcfiu: okay, regarding the serial console, where is it connected to?
1164511204 M * marcfiu vici.cs.princeton.edu
1164511238 M * marcfiu hmph....
1164511249 M * marcfiu can you still get onto that machine?
1164511255 M * marcfiu I don't see your ssh key
1164511258 M * Bertl maybe, trying now
1164511290 M * Bertl nope, doesn't look like
1164511344 M * marcfiu searching for ssh key you've sent me.
1164513128 Q * kir Server closed connection
1164513139 J * kir ~kir@swsoft-mipt-nat.sw.ru
1164513163 M * Bertl wb kir!
1164513853 Q * lilalinux Remote host closed the connection
1164514674 Q * jack Quit: Leaving
1164514850 Q * vasko Server closed connection
1164514851 J * vasko ~vasko@unreal.rainside.sk
1164514851 A * vasko is gone. Gone since Wed Sep 20 15:44:00 2006
1164519754 Q * Piet Quit: Piet
1164521341 Q * Zaki Ping timeout: 480 seconds
1164522217 P * thunder1 
1164522289 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164522359 J * thunder11 ~thu@tor-irc.dnsbl.oftc.net
1164522360 P * thunder1 
1164522442 N * thunder11 thunder1
1164523769 P * thunder1 
1164523798 J * thunder1 ~thu@68.167.210.150
1164524544 J * Aiken ~james@tooax6-124.dialup.optusnet.com.au
1164524565 M * Bertl hey Aiken!
1164524587 M * Aiken hi Bertl
1164524682 M * Bertl I give you an early warning, we will have some testing to do in the next few days
1164524729 M * Aiken ok
1164525425 Q * m4z Read error: Operation timed out
1164525438 J * m4z ~m4z@bastard-operator.from-hell.net
1164525794 Q * m4z Read error: Connection reset by peer
1164525801 J * m4z_ ~m4z@bastard-operator.from-hell.net
1164525804 N * m4z_ m4z
1164525901 Q * thunder1 Remote host closed the connection
1164525909 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164526241 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at
1164526287 M * Bertl morning DavidS!
1164526563 J * Zaki ~Zaki@88.213.57.34
1164526938 M * DavidS Hi Bertl! I'm doing a little talk about puppet, cfengine-done-right, at debienna tomorrow, i'd think you'd might be interested in it ...
1164527123 M * micah daniel_hozac: so it is frowned apon to upload new upstream releases now, but maybe we can identify specific fixes that are worth getting in?
1164527967 M * Bertl DavidS: will see, maybe I find the time
1164527987 M * Hollow morning
1164528017 M * DavidS Bertl: I'll start around 20:00 (local time ;)
1164528868 Q * DreamerC_ Quit: leaving
1164528885 J * DreamerC ~dreamerc@61-217-227-9.dynamic.hinet.net
1164530614 J * ouhena ~ouhena@ouhena69.net2.nerim.net
1164530616 J * moki ~moki@chello212186163177.32.11.vie.surfer.at
1164530627 M * moki good morning
1164530703 M * moki i am reading the xen dokumentation right now. has anyone have experience with combining xen and vserver? linux vserver enabled kernel as domU ?
1164530798 M * moki i used linux vserver a few times... xen is new to me. but the ability to have a openbsd oder netbsd host besides some vserver contexts does really sound ineresting to me.
1164530806 J * dna_ ~naucki@208-213-dsl.kielnet.net
1164530853 M * moki hi dna
1164530970 M * Bertl welcome moki!
1164530977 M * moki servas bertl :)
1164530992 M * Bertl moki: most important, linux-vserver and xen are easily combined
1164531001 M * moki bertl: i thought so :)
1164531005 Q * thunder1 Remote host closed the connection
1164531022 M * Bertl moki: debian even provides kernels which can be used in all combinations
1164531024 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164531026 M * moki bertl: anyone using it in the real world?
1164531046 M * Bertl yes, I know at least two people :)
1164531060 M * moki bertl: fine, im a debianiac
1164531073 M * moki bertl: fine, lets see if i can make it 3
1164531095 M * moki thats how a nice relaxing sunday should start :)
1164531118 M * DavidS moki: aye, I am .. works very fine :)
1164531141 M * moki DavidS: smooth, what oses do you run parallel ?
1164531203 M * DavidS I currently have only debian and windows XP ... I will be using XEN primarily to separate customers, while i use vserver to keep services clean
1164531406 M * moki DavidS: ic. i think about sticking a obsdfirewall infront of linuxdomUs with vservers in it to seperate services.
1164531467 M * moki i eant to proof that "personalfirewalls" - softwarefirewalls, can do good work to ;)
1164531473 M * DavidS i am dubious about the value of too much firewalling within the xen
1164531515 M * moki davids: isnt there strict separation possible?
1164531556 M * DavidS a) it significantly hurts performance pushing packets through b) whatever you do, the xen kernel is always exposed to the network, if you don't start to hand through the actual pci devices
1164531561 M * moki davids: as far as i read i can attach one domU to pci devices, cpus... all sticky like the new cpusetfeature on linux.
1164531616 M * DavidS not the least: if you can't access you dom0 without going through some intermediate, you can't (savely) reboot this intermediate from the dom0 because it'd kill your connection
1164531618 M * moki DavidS: hmm, ok, i read about a 5% performance penalty... ok, heavy networking may not be included in here.
1164531666 M * DavidS it isn't noticable if you just bridge the domU to you're external network, but i tried firewalling on a linux domU and it hurt pretty much
1164531666 M * moki davids: that sounds sane
1164531712 M * moki DavidS: what cpus do you se?
1164531758 M * DavidS model name      : Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz
1164531761 M * DavidS only one ;)
1164531788 M * moki DavidS: for i am going to to utilyze a quad pentiumII box back from '98 :) .. with netbsd and linux. openbsd doesnt run on those cpus in xen.
1164531838 M * DavidS slick :)
1164531891 M * moki DavidS: yeah, a dell poweredge 6300, 4_*400mhz, 4g ram.. all hotswap...   in my opinion a little like a sun e450, but with intelcpus and 32bit
1164531950 M * moki DavidS: but this box for shure performs much slower than your setup :)
1164532038 M * moki anyways, ill give it a try. and if it doesnt work. i buy a soekris ;)
1164532127 M * moki another thing
1164532234 M * moki i just read about suspending/freezing - keeping the state and later restart the Vhost. is soemthing like that thought about to be integrated into vserver ? for migration purposes? ... dont understand me wrong i know that i can stop my vserver, copy the content to some place else. copy the config and start it there.
1164532691 M * moki just athought :)
1164534057 Q * DavidS Ping timeout: 480 seconds
1164534146 Q * sladen Ping timeout: 480 seconds
1164534183 J * sladen paul@starsky.19inch.net
1164535713 Q * michal` Ping timeout: 480 seconds
1164536146 J * michal` ~michal@www.rsbac.org
1164536742 M * Bertl okay, I'm off to bed now .. have a good one everyone! cya!
1164536751 N * Bertl Bertl_zZ
1164536796 Q * Aiken Remote host closed the connection
1164536817 J * Aiken ~james@tooax6-124.dialup.optusnet.com.au
1164537429 M * moki cu later
1164537430 M * moki bye
1164537432 P * moki Leaving
1164537607 Q * thunder1 Remote host closed the connection
1164537633 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164538397 Q * nebuchadnezzar Quit: ERC Version 5.1.4 (IRC client for Emacs)
1164539275 J * bonbons ~bonbons@83.222.39.117
1164539293 Q * besonen Read error: Connection reset by peer
1164539298 J * besonen_ ~besonen@dsl-db.pacinfo.com
1164539422 J * nebuchadnezzar ~nebu@zion.asgardr.info
1164540666 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1164540676 J * ensc ~irc-ensc@p54B4D3AB.dip.t-dialin.net
1164541815 Q * Aiken Quit: Leaving
1164543321 Q * shedi Remote host closed the connection
1164543637 J * shedi ~siggi@inferno.lhi.is
1164543824 P * ouhena 
1164544935 Q * rob-84x^ Quit: That's it for today
1164544967 J * rob-84x^ rob@submarine.ath.cx
1164546172 Q * marcfiu Quit: Download Gaim: http://gaim.sourceforge.net/
1164547127 M * daniel_hozac micah: well, 0.30.212 is mostly a feature release, not a bugfix release.
1164547287 M * daniel_hozac i guess changeset 2396, 2395, 2368, 2358, and 2357 are the most important ones for you.
1164547435 M * daniel_hozac personally i think it's too bad Debian users will be missing out again :)
1164547773 M * daniel_hozac oh, and changeset 2402 should fix 383794.
1164549554 M * sid3windr why are debian users missing out?
1164549709 M * Loki|muh missing out means what?
1164549736 M * daniel_hozac 0.30.212 has quite a few new features.
1164549902 M * daniel_hozac build -m rsync, naddress, vastly improved Gentoo support, 2.1 scheduler configuration support, support for all the limits/flags/etc. from 2.1, and probably some more things i forgot.
1164550256 M * daniel_hozac ah yes, an improved namespace cleanup is enabled by default.
1164550323 M * eyck_  
1164550341 Q * eyck_ Quit: leaving
1164550354 J * eyck zI6f2wLk@nat.nowanet.pl
1164550564 J * cuerva ~soatola42@82.153.18.114
1164550985 Q * soatolaEspera Ping timeout: 480 seconds
1164551810 J * Johnnie ~jdlewis@jdlewis.org
1164553301 Q * meandtheshell Quit: Leaving.
1164553564 J * meandtheshell ~markus@85-124-207-182.dynamic.xdsl-line.inode.at
1164553574 J * lilalinux ~plasma@dslb-084-058-198-224.pools.arcor-ip.net
1164554500 J * Borg- borg@cube.benet.uu3.net
1164556013 P * click [IRSSI]
1164557061 Q * FloodServ osmosis.oftc.net helium.oftc.net
1164558778 N * weasel Guest0
1164558780 N * Guest0 weasel
1164558821 J * FloodServ services@services.oftc.net
1164558923 Q * FloodServ hydrogen.oftc.net helium.oftc.net
1164558945 J * FloodServ services@services.oftc.net
1164558945 Q * weasel Killed (lithium.oftc.net (Nick collision (new)))
1164558951 J * weasel weasel@asteria.debian.or.at
1164559512 J * click click@ti511110a080-5677.bb.online.no
1164561709 Q * eyck Remote host closed the connection
1164561714 J * eyck jaeA0U1p@nat.nowanet.pl
1164562644 Q * eyck Ping timeout: 480 seconds
1164562733 J * SoftIce ~newbie@vc-196-207-45-253.3g.vodacom.co.za
1164562806 M * SoftIce hi, i'm getting an issue now that I have upgraded to fc6
1164562808 M * SoftIce http://rafb.net/paste/results/sP7fBh35.html
1164562930 J * eyck e70DW9cL@nat.nowanet.pl
1164563565 M * SoftIce does anyone have a fc6 wiki guide
1164563626 J * blues^ blues@blysk.ds.pg.gda.pl
1164563630 Q * blues Read error: Connection reset by peer
1164563633 J * bubulak_ ~bubulak@whisky.pendo.sk
1164563656 Q * bubulak Read error: Connection reset by peer
1164563679 Q * bon Read error: Connection reset by peer
1164563684 J * bon bon@blij.in-de.eu
1164564331 M * daniel_hozac SoftIce: yeah, i broke vyum-worker slightly for FC6.
1164564372 M * SoftIce oh ;)
1164564380 M * SoftIce any workarounds for the mean time ?
1164564392 M * daniel_hozac http://svn.linux-vserver.org/projects/util-vserver/changeset/2356 http://svn.linux-vserver.org/projects/util-vserver/changeset/2361
1164564441 M * micah daniel_hozac: what is a svn URI I can use to check out the repository? I can only seem to get at it via the web
1164564507 M * daniel_hozac http://svn.linux-vserver.org/svn/util-vserver/trunk/
1164564564 M * SoftIce err, is that a diff patch?
1164564573 M * daniel_hozac SoftIce: what?
1164564595 M * daniel_hozac SoftIce: there's a unified diff link at the bottom.
1164564682 N * Bertl_zZ Bertl
1164564686 M * Bertl morning folks!
1164564691 M * daniel_hozac morning Bertl½
1164564701 M * daniel_hozac s/½/!/
1164564771 M * Bertl hey, how's going? do you have patches for the sched.h issue?
1164564801 M * daniel_hozac no... i gave it a try, but it seemed to be a bit more complicated than i expected.
1164564807 M * SoftIce daniel_hozac: what about just replacing it with this code
1164564810 M * SoftIce http://svn.linux-vserver.org/projects/util-vserver/browser/trunk/scripts/vyum-worker?rev=2361
1164564816 M * SoftIce or is that the broken code?
1164564840 M * Bertl daniel_hozac: :(
1164564847 M * daniel_hozac SoftIce: that's the current version, so it should work.
1164564887 Q * derjohn Quit: by(t)e
1164565001 Q * thunder1 Ping timeout: 480 seconds
1164565001 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1164565008 M * micah daniel_hozac: huh, I could've sworn I tried that URL, guess I was confused
1164565031 M * daniel_hozac heh
1164565055 M * daniel_hozac Bertl: hmm, actually... couldn't we include vserver/context.h from sched.h still?
1164565087 M * SoftIce daniel_hozac: works thanks, listen I forgot how I got the right yum version for this error
1164565115 M * daniel_hozac SoftIce: what?
1164565119 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net
1164565124 M * SoftIce You are using a version of yum which is insecure and broken in chroot.....
1164565129 P * marcfiu 
1164565130 M * daniel_hozac right.
1164565137 M * daniel_hozac that message tells you how to fix it.
1164565144 M * SoftIce oh, sorry
1164565166 M * SoftIce or ask the author of yum to apply them
1164565167 M * SoftIce  (preferred).
1164565168 M * SoftIce ;0
1164565169 M * SoftIce heh
1164565173 M * SoftIce kidding
1164565179 M * daniel_hozac well, that's not going to happen...
1164565190 M * daniel_hozac skvidal is very much against it.
1164565200 M * daniel_hozac (at least the last time i asked)
1164565215 M * SoftIce why is that? showing them up? ;)
1164565238 M * daniel_hozac what?
1164565267 M * SoftIce well you pointing out a flaw in their yum utility
1164565287 M * daniel_hozac it's more of an improvement.
1164565325 Q * MrX Ping timeout: 480 seconds
1164565368 M * SoftIce I guess if he gets enough mails about it, he might do something about it.
1164565394 M * daniel_hozac i doubt it.
1164565442 M * daniel_hozac i can understand why he doesn't want to apply it.
1164565442 M * SoftIce any idea why? does he see it causing more harm than good?
1164565447 M * SoftIce why is that
1164565463 M * daniel_hozac (chroot|host)fs aren't specific enough names.
1164565582 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1164565952 M * SoftIce ye, but some people do, and that would be helpfull for the people that support FC
1164566078 P * marcfiu 
1164566810 M * SoftIce hmm, so say a standard fc5 kernel
1164566815 M * SoftIce what modifications do you guys make to it?
1164566838 M * SoftIce do you guys put added modules to hardware
1164566846 M * SoftIce or is it just related to chroot, etc?
1164566850 M * SoftIce userland stuff
1164566917 M * daniel_hozac we add the Linux-VServer patch, of course.
1164566940 M * SoftIce and thats all?
1164566949 M * SoftIce as iv done a fc4/fc5 upgrade
1164566954 M * SoftIce and getting a kernel panic now after upgrade
1164566960 M * SoftIce thinking if you kernel would fix it
1164566962 M * SoftIce ?
1164566974 M * daniel_hozac very unlikely.
1164566978 M * daniel_hozac what does the panic say?
1164566979 M * micah daniel_hozac: i've got all those in... i think it'll be ok to not have the 2.1 features as the etch kernel will be the 2.6.18 kernel, which has 2.0.2
1164566993 M * SoftIce daniel_hozac: thats what sucks, its at a data centre
1164566995 M * SoftIce they didn't say
1164567052 M * SoftIce daniel_hozac: what about if I upgrade again from fc5 / fc6
1164567054 M * daniel_hozac micah: sure, but at least naddress, build -m rsync and namespace cleanup by default are very useful things even on 2.0.
1164567066 M * SoftIce then use the fc6 kernel, is that completely different to fc5 kernel
1164567070 M * SoftIce maybe I would have better luck with that
1164567070 M * SoftIce ?
1164567086 M * daniel_hozac FC5 and FC6 aren't that different.
1164567113 M * SoftIce but I cant see why 2.6.17 /18 are that much different in the kernel config they have done
1164567127 M * micah daniel_hozac: what is naddress? I haven't followed that
1164567140 M * SoftIce daniel_hozac: do yoou think it could be different enough to fix the kernel issue? :)
1164567141 M * daniel_hozac micah: you can add/remove addresses on the fly.
1164567158 M * daniel_hozac micah: if your network context is setup for it (i.e. has more than one address)
1164567183 M * daniel_hozac SoftIce: without having the slightest idea of what the problem is, i can't say.
1164567224 M * micah daniel_hozac: oh nice, so the 'bug' that I was using to add addresses was fixed, and then a proper feature was added :)
1164567235 M * daniel_hozac micah: right ;)
1164567248 M * SoftIce daniel_hozac: The monitor showed that the kernel panic'd..The machine was rebooted with the default kernel
1164567251 M * SoftIce heh how helpfull was that
1164567258 M * SoftIce thats alll they told me..
1164567260 J * Carpenter ~gt-blacks@p5086F784.dip.t-dialin.net
1164567263 M * SoftIce then they finished with saying this
1164567280 M * SoftIce You should be able to upgrade to FC 6 or FC 5 via yum. You might want to google for some information on how to do so. 
1164567302 M * SoftIce heh ;) like I have not done a million upgrades
1164567344 M * micah daniel_hozac: well... humm. the changesets you pointed out do the namespace cleanup, maybe we can isolate naddress changes and get them in. I'm still not clear what build -m rsync gets you, but thats probably easily isolated
1164567368 M * daniel_hozac micah: that's just the improved namespace cleanup, no? or did i give you the change of default too? :)
1164567394 M * daniel_hozac micah: but with all of those things... you have basically 0.30.212.
1164567409 M * daniel_hozac 0.30.212 is just some cleanups in addition to those features.
1164567412 M * micah daniel_hozac: you just gave me 2357, the improve namespace cleanup changeset
1164567487 M * daniel_hozac that's what i thought.
1164567507 M * daniel_hozac 2379 enables it by default.
1164567595 M * daniel_hozac and build -m rsync gives you cloning capabilities, even from another host.
1164567736 J * Piet hiddenserv@tor.noreply.org
1164567742 M * daniel_hozac seems easier to just take 0.30.212-rc1 and call it 0.30.211? :)
1164567754 M * micah it might :)
1164567766 M * daniel_hozac (http://people.linux-vserver.org/~dhozac/p/uv/experimental/delta-0.30.211-0.30.212-rc1.diff is the entire patch, FWIW)
1164567771 M * micah thanks
1164568063 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1164568161 Q * Piet Remote host closed the connection
1164568373 M * SoftIce hmm, I see the SMP side of FC now has changed
1164568456 M * Bertl okay, will take a shot at the base.h approach ...
1164568486 M * daniel_hozac Bertl: if we keep vserver/context.h in sched.h, it should get a lot easier... that should be possible, right?
1164568519 Q * sid3windr Ping timeout: 480 seconds
1164568829 J * sid3windr luser@bastard-operator.from-hell.be
1164568932 M * Bertl daniel_hozac: let's see how far I get without that 'trick' :)
1164568934 M * SoftIce daniel_hozac: your latest kernel for FC is that a FC6 kernel or just a FC5 2.6.18 kernel patched?
1164568997 M * daniel_hozac SoftIce: the FC6 kernel is an FC6 kernel, and the FC5 kernel is an FC5 kernel.
1164569130 M * daniel_hozac Bertl: http://people.linux-vserver.org/~dhozac/p/k/delta-base-dirty01.diff is how far i got.
1164569233 M * Bertl daniel_hozac: do you ahve numbers for the memory -> mm.h move?
1164569251 M * SoftIce daniel_hozac: yes but i'm saying the vserver kernel
1164569265 M * SoftIce what is that based on, is it based on FC6 kernel + vserver patches?
1164569276 M * daniel_hozac the FC6 one, yes.
1164569279 J * Piet hiddenserv@tor.noreply.org
1164569294 M * daniel_hozac Bertl: damn, i must've lost those... i had to restart X yesterday.
1164569297 M * SoftIce ahh, thanks
1164569334 M * micah daniel_hozac: hmm I get a link error when building with those changes
1164569343 M * micah src/chroot-sh.c:178: undefined reference to `isNumberUnsigned'
1164569396 M * daniel_hozac micah: did you run autoreconf -fi? or make -f Makefile.svn?
1164569448 M * daniel_hozac Bertl: for direct includes it was something like 1000+ files.
1164569502 M * Bertl so that would add quite some dependacies, right?
1164569523 M * micah daniel_hozac: no, I haven't had to in the past
1164569533 M * daniel_hozac micah: yes you have :)
1164569552 M * micah i haven't altered my build scripts, which worked before
1164569555 M * daniel_hozac Bertl: yeah.
1164569564 M * daniel_hozac micah: not when building from SVN
1164569581 M * micah well, I haven't built from svn before :)
1164569958 M * micah need to adjust my build-deps to have libtool now
1164570185 M * daniel_hozac not autoconf and automake?
1164570917 M * micah they already where there
1164570941 M * micah although I did adjust automake to require 1.9
1164570953 M * micah everything builds and works fine now, btw. :)
1164570976 M * micah hopefully this will get some more testers
1164571184 M * daniel_hozac you use dietlibc everywhere, right?
1164571191 M * Bertl micah: the folks from debian.at are testing a lot of the stuff
1164571340 J * ptryk ~chatzilla@h248n2fls34o284.telia.com
1164571351 M * Bertl welcome ptryk!
1164571390 M * ptryk hello
1164571411 M * ptryk and.. thanks .. wow this place is busy o_O
1164571470 M * Bertl what's up? can we help you?
1164571493 M * ptryk i am actually here hoping to gain some insight.. i am new with building kernels and such ..
1164571522 M * ptryk i tried earlier to build the latest stable to put into kubuntu.. 
1164571522 M * Bertl well, there are a bunch of howtos out there, any specific issues/questions?
1164571547 M * ptryk when it was installed it wouldn't give me more than console login?
1164571576 M * Bertl the Linux-VServer kernels look and feel like the normal kernel
1164571584 M * ptryk i was wondering if anyone else maybe had experinced it?
1164571595 M * Bertl except for the fact that ther is more functionality available
1164571633 M * Bertl i.e. if you switch from a vanilla (kernel.org) kernel to a Linux-VServer patched one, nothing obvious will change
1164571648 M * micah daniel_hozac: yeah dietlibc is used, although I'm not sure how I would confirm that it is used 'everywhere'
1164571691 M * micah Bertl: cool, new version should be available in the next day with the 212-rc1 fixes
1164571691 M * ptryk that is what I thought, which is why I thought it forcing me into console mode was a bit unruly.
1164571719 M * Bertl console mode means, you do not get X or what?
1164571740 M * ptryk yeah sorry, no GUI.
1164571762 M * Bertl what kind of X/graphics card/setup do you use?
1164571817 M * ptryk not sure about the x/graphics.. the card is nividia  .. ubuntu uses gdm.. kubuntu is kde based.
1164571844 Q * lilalinux Remote host closed the connection
1164571847 M * Bertl nvidia with maybe proprietary drivers?
1164571856 M * ptryk bertl.. true.
1164571867 M * ptryk need to reload them?
1164571870 M * Bertl so you will need to reinstall them after each kernel change
1164571881 M * ptryk ahhh.. of course! doh.
1164571911 M * ptryk will give it a go.. thanks so much for your help! :)
1164571924 M * Bertl you#re welcome! feel free to hang around!
1164571945 M * ptryk thanks, will do ;)
1164571946 M * Bertl s/#/'/
1164572324 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at
1164572455 M * Bertl welcome DavidS!
1164572490 Q * SoftIce 
1164572745 M * DavidS 'lo Bertl :)
1164572767 M * daniel_hozac micah: on all arches you build for, i mean.
1164572769 M * Bertl daniel_hozac: micah has a new util-vserver version to test, if you like ...
1164572777 M * Bertl *sorry*
1164572805 M * Bertl DavidS: micah has a new util-vserver version to test, if you like ...
1164572806 M * daniel_hozac micah: glibc builds will need changeset 2405.
1164572808 A * DavidS has to finish his puppets for the talk/demo tomorrow
1164572896 M * DavidS btw, i can provide you a xen-vserver domU for testing on Debian/i386 if you want one ..
1164573032 M * micah daniel_hozac: is 2405 in -rc1?
1164573055 M * micah i guess I can look
1164573065 M * Bertl DavidS: cool!
1164573084 M * daniel_hozac micah: no.
1164573102 M * daniel_hozac micah: -rc1 was built after 2402.
1164573110 M * micah i see it now
1164573149 M * micah well, I'm pretty sure we are dietlibc on all arches, so that shouldn't impact us
1164573220 M * DavidS obviously, it will only have a private ip and ssh on a high port forwarded, but it sits behind a unlimited 4Mbit pipe and since i don't expect too much cpu load by you, I'd allow it usage of the whole cpu.
1164575757 J * crypt ice@p54A3A7CD.dip0.t-ipconnect.de
1164575759 M * crypt Hi all
1164575766 M * Bertl hey crypt!
1164575782 M * crypt Bertl, one question what does the setugid util do?
1164575823 M * daniel_hozac sets the uid/gid and executes a program? :)
1164575826 M * Bertl it sets the user and/or group id for a process
1164575856 M * crypt ok you said to me that this programm would be useful for the testing with quota
1164575885 M * daniel_hozac yes, as you'll need to use a specific uid/gid (the one you set the quota for)
1164575918 M * crypt ok so i have to set the vserver's root dir to a specific uid and gid with that tool?
1164575949 M * daniel_hozac for the quota tests, you shouldn't think of vservers at all.
1164575978 M * daniel_hozac all you have is the uids/gids (initially, later xid will be added).
1164575987 M * crypt ok I think i have a general understanding problem off what to do exactly for the tests ;)
1164576139 M * crypt do I have to create afew dirs in the shared partiotn set some uid gids on that dir's and have to test wether the quota i added for these dirs is handeld correct?
1164576152 M * Bertl you need not only test for uid=gid=0 (root) but more important for uid!=0 and gid!=0 cases as well
1164576187 M * Bertl as testing with different users and groups becomes complicated, as it usually depends on a passwd file :) this tool helps you to simplify testsw
1164576190 M * Bertl -w
1164576274 M * crypt so i create quotas on dir's where the users are not in passwd
1164576279 M * crypt file
1164576302 M * daniel_hozac the lack of passwd entries helps with the reproducibility of the tests.
1164576306 M * Bertl quotas are not related to dirs, that seems to be amisconception on your side
1164576330 M * crypt ahhh ok
1164576340 M * Bertl on the contrary, a directory will also add to quota (in this case inode quota not block quota)
1164576424 M * crypt ok so let's say: on the shared partition i create a quota for a user that doesn't exist like uid and gid 7000
1164576433 M * crypt inode quota
1164576435 M * Bertl as I said last time, play around alittle with quota first
1164576452 M * crypt yeah i tried to do that ;) and that works as i thought so ;)
1164576457 M * Bertl get a feeling for how quota works, when it changes and who is affected from quota limits in what ways
1164576460 Q * dna_ Quit: Verlassend
1164576496 M * crypt ok ;) i'll read a lot more about quota
1164576525 M * daniel_hozac crypt: btw, did i understand correctly before that you are adding quota tests to testfs?
1164576551 M * crypt daniel_hozac, yes
1164576782 M * crypt i'm using the ext3 filesystem
1164576867 M * crypt cerated by testfs.sh
1164576872 M * daniel_hozac testfs should create lots of different filesystems.
1164576901 M * crypt yes but i have limit it to ext3 after seeing, that the rest works as well ;)
1164576931 M * crypt or is that not correct?
1164576955 M * daniel_hozac sure, creating the tests on a known-good filesystem seems like a good idea.
1164576974 M * crypt ok thanks ;)
1164576979 J * Aiken ~james@tooax8-232.dialup.optusnet.com.au
1164577038 M * Bertl morning Aiken!
1164577075 M * Aiken hi
1164577122 M * crypt so next steps for me (informing more about how quota is working), using the setugid tool to set limits for user's who doesn't exist and after that testing the quota stuff again in this nvironement
1164577130 M * crypt environment
1164577161 M * Bertl hmm, no .. you can set limits for all kind of uid/gid which do not exists easily
1164577179 M * crypt hmm but for what using xour tool?
1164577181 M * Bertl but you will have a hard time to test them without a tool like the setugid :)
1164577183 M * crypt -x+y
1164577204 M * Bertl i.e. you cannot easily extend a file as uid=7 without it
1164577220 M * Bertl and you cannot query quota for uid=7 (as user) without it
1164577255 M * crypt so i create/expand files through your util for the user who does not exist
1164577295 M * Bertl for example, to verify that he is accounted properly for, and quota follows/limits
1164577321 M * Bertl look, let me give an example here:
1164577335 M * Bertl you want to test quota as non root, okay?
1164577340 M * crypt yes
1164577350 M * Bertl so you ahve to become a user, right?
1164577354 M * crypt yes
1164577361 M * Bertl what user will you become?
1164577389 M * crypt the user who is limited through quota
1164577404 M * Bertl okay, so you need to create a user (which does not own files)
1164577413 M * crypt right
1164577423 M * Bertl that is something which will be kind of hard to do in a test script, right?
1164577437 M * crypt yes this is right
1164577459 M * Bertl instead of doing that, and possibly destroying a setup, you can simply 'switch' to some uid
1164577473 M * Bertl that's what the tool is for
1164577473 M * crypt a nice :)
1164577519 M * crypt thanks a lot for that full description!
1164577534 M * Bertl you're welcome!
1164577571 M * crypt ;) I'll write all what i'm learning down, for maybe some other user having the same questions
1164577598 Q * bonbons Quit: Leaving
1164577627 M * Bertl that can't hurt ...
1164577634 M * crypt yes really
1164577684 M * waldi "it hurts if i do this"
1164577872 M * Bertl waldi: how so?
1164577902 M * waldi hihi :)
1164578068 J * DuckMaster ~Duck@tox.dyndns.org
1164578105 A * doener .oO( a quota tester returned? )
1164578117 Q * DuckMaster Remote host closed the connection
1164578119 M * doener that's... new :)
1164578137 J * DuckMaster ~Duck@tox.dyndns.org
1164578187 Q * Blissex Remote host closed the connection
1164578192 Q * duckx Read error: Connection reset by peer
1164578243 M * daniel_hozac doener: indeed :)
1164578644 Q * michal` Ping timeout: 480 seconds
1164579128 M * daniel_hozac micah: where could i get the files?
1164579143 J * michal` ~michal@www.rsbac.org
1164579144 M * daniel_hozac ah, incoming.debian.org, right.
1164579685 J * mariilusii hkjhg@port-83-236-50-8.dynamic.qsc.de
1164579776 M * Bertl welcome mariilusii!
1164580186 M * crypt Bertl, what filesystem user id do i have to use for the quota tests? same as user id?
1164580261 M * Bertl make yourself more comfortable with the term fsuid, effective uid and uid first, same goes for the guid, and test which one actually affects the filesystem
1164580288 M * crypt thanks
1164580289 M * Bertl also note the difference between 'echo test >a' and 'touch b'
1164580303 M * crypt k
1164580333 M * crypt thanks again
1164580361 M * Bertl np
1164580390 M * crypt on this level of files i have never played so I have to learn a lot
1164580704 J * lilalinux ~plasma@dslb-084-058-198-224.pools.arcor-ip.net
1164581149 Q * crypt Quit: Leaving
1164581938 J * Osgiliath ~osgiliath@kurzweil.no-ip.org
1164581950 M * Osgiliath hi :)
1164582159 M * Bertl hey Osgiliath!
1164582474 M * daniel_hozac Bertl: ah, now that i remember, could we go through some of your bugs in savannah?
1164582525 M * Bertl sure
1164582537 M * doener Bertl: just to check, "echo test >a" recreates "a" with content "test" and "touch b" create "b" if it does not exist and changes mtime otherwise. right?
1164582538 M * Bertl btw, I'll have some new work for you shortly
1164582553 M * Bertl (checking for duplicate/missing includes)
1164582577 M * daniel_hozac doener: and echo test >a requires you use a bash -c "echo test >a" in your test scripts rather than just setugid ... echo test > a.
1164582584 M * Bertl doener: yes, but it gets even more interesting when used in a typical shell setup :)
1164582621 M * doener daniel_hozac: and even that failed for me when used directly with ssh ;)
1164582639 M * doener i.e. ssh somehost bash -c "echo foo > file"
1164582647 M * doener file ended up being empty
1164582648 M * doener :(
1164582652 M * daniel_hozac Bertl: https://savannah.nongnu.org/bugs/?13844 this is actually a bug in debootstrap, right?
1164582666 M * daniel_hozac doener: hmm, that sounds strange.
1164582681 M * Bertl doener: but for quota, it has another aspect, the touch will not allocate any block
1164582700 M * doener just a dentry
1164582720 M * daniel_hozac dentries aren't limited, are they? i thought it was inodes and blocks.
1164582755 M * Bertl we have dentry limits :)
1164582763 M * daniel_hozac right, but quota doesn't ;)
1164582767 M * doener ehrm yeah, an inode as well (dentry without inode would be hard I guess)
1164582777 M * daniel_hozac yeah ;)
1164582779 M * Bertl doener: no, no inode
1164582785 M * doener not?
1164582791 M * Bertl the inode is already there, it's the dir inode
1164582791 M * daniel_hozac hmm, now you lost me too.
1164582803 M * daniel_hozac shouldn't a get an inode?
1164582810 M * daniel_hozac or do we assume a already exists?
1164582830 M * Bertl yes, it will get an inode if it doesnt exist
1164582864 M * doener that's what I meant... it won't get another dentry either if it already exists, will it?
1164582867 M * Bertl but the file entry itself doesn't cause an inode to be allocated
1164582881 M * daniel_hozac right.
1164582882 M * Bertl just take ln as example
1164582908 M * doener yeah, hardlinks are different, but touch doesn't create them ;)
1164582920 M * Bertl no, hardlinks are not different :)
1164582936 M * Bertl touch actually does two things
1164582938 M * daniel_hozac depends on what we're talking about, i suppose :)
1164582956 M * Bertl - it creates a new inode (without block extents)
1164582969 M * Bertl - it adds a directory entry to the existing dir inode
1164582989 M * Bertl the entry goes without any quota allocation
1164582990 M * doener a hardlink just takes a dentry, the inode must exist... touch may create/creates a complete file with inode and a dentry...
1164583035 M * doener guess I managed to confuse everyone :)
1164583049 M * daniel_hozac hehe :)
1164583146 M * Bertl doener: do something useful ... check my code :)
1164583153 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.18.3-vs2.1.1.2.1.diff
1164583163 M * daniel_hozac haha.
1164583182 M * Bertl daniel_hozac: please run your magic scripts on that
1164583196 M * daniel_hozac what should i be adding to them?
1164583217 M * Bertl we are looking for duplicate/missing includes
1164583235 M * Bertl where the missing part is probably hard to accomplish
1164583252 M * daniel_hozac missing is a lot easier than duplicate :)
1164583253 M * Bertl I'll submit that to PLM to get a check run
1164583316 M * Bertl ah, what would be interesting too are excessive includes too
1164583326 M * Bertl (i.e. unnecessary ones)
1164583333 M * daniel_hozac that's quite easy as well.
1164583337 M * daniel_hozac for which headers?
1164583349 M * Bertl vs_* and vserver/*
1164583366 M * Bertl all of them got slightly modified
1164583384 M * daniel_hozac ah, ok.
1164583420 M * doener Bertl: actually I went back to ptrack... is that acceptable? ;)
1164583420 P * anonc adios
1164583438 M * Bertl doener: what's that?
1164583453 M * doener well, p stands for "patch" ;)
1164583473 M * daniel_hozac ah, and the RSS stuff is fixed?
1164583483 M * daniel_hozac or on it's way to being fixed?
1164583534 M * Bertl yes, that should work now :)
1164583571 M * Bertl have to break that down into smaller hunks, but now I want to know if it works ...
1164583878 M * daniel_hozac i'll run the previous version of my script while i'm investigating what changes i need to do...
1164584015 M * Bertl okay
1164584449 M * Osgiliath Bertl, stupid question, but, i compiled utils-vserver 0.30.211 with ./configure --prefix=/ --with-vrootdir=/vservers
1164584458 M * Osgiliath and each time, i get this :
1164584467 M * Osgiliath Paths:
1164584467 M * Osgiliath                    prefix: /usr/local
1164584467 M * Osgiliath         sysconf-Directory: ${prefix}/etc
1164584467 M * Osgiliath             cfg-Directory: ${prefix}/etc/vservers
1164584467 M * Osgiliath          initrd-Directory: $(sysconfdir)/init.d
1164584468 M * Osgiliath        pkgstate-Directory: ${prefix}/var/run/vservers
1164584468 M * Osgiliath           vserver-Rootdir: /vservers
1164584487 M * Osgiliath with /usr/local as prefix .... :x
1164584489 M * Bertl I doubt that, what I think actually happens
1164584497 M * Bertl is that you ahve two versions installed now
1164584512 M * Bertl one which is in /usr/local/* (the one you are constantly calling)
1164584521 M * Bertl and another one in the places you expect
1164584534 M * Osgiliath hum ok i see
1164584602 M * Osgiliath thanks :)
1164584631 M * Bertl np
1164584721 M * Bertl daniel_hozac, doener: hmm, it seems they are constantly deproving PLM :(
1164584743 M * Bertl or what is the oposite of 'improving'?
1164584765 M * daniel_hozac hehe, good question.
1164584810 M * daniel_hozac worsen is what the dictionaries suggest.
1164584854 M * doener deteriorate?
1164584860 M * DavidS erode
1164584869 M * DavidS ruin
1164584888 M * Bertl webster/thesaurus sais worsen .. so I'll stick to that for now :)
1164584892 M * DavidS "verschlimmbessern", my grandmother used to say
1164584906 M * Bertl yeah, that is what I had in mind
1164584916 M * Bertl but that is not really translateable
1164584944 M * DavidS sadly ;)
1164584991 M * daniel_hozac Bertl: what did they do to PLM now?
1164585097 Q * ptryk Quit: exit stage left..
1164585169 M * doener Bertl: webster knows "disimprove" :)
1164585182 M * Bertl great, that sounds useable ...
1164585200 M * Bertl daniel_hozac: well, it seems (I only have two results yet) that they stop after the first error
1164585201 M * doener so you were pretty close ;)
1164585214 M * daniel_hozac wow, that would be bad.
1164585228 M * Bertl http://plm.testing.osdl.org/filter_requests/output/45427
1164585259 M * Bertl ah, seems to be an ia32 oddity
1164585306 M * Bertl btw, ia32 .. does that mean they do not run on amd?
1164585308 M * DavidS hail the days of batch processing and turn-around times counted in days!