1371170205 J * thierryp ~thierry@home.parmentelat.net 1371170690 Q * thierryp Ping timeout: 480 seconds 1371171429 M * Bertl_oO off to bed now ... have a good one everyone! 1371171437 N * Bertl_oO Bertl_zZ 1371173836 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:e501:2f0d:2825:7a35 1371174321 Q * thierryp Ping timeout: 480 seconds 1371177475 J * thierryp ~thierry@home.parmentelat.net 1371177501 Q * FireEgl Read error: No route to host 1371177959 Q * thierryp Ping timeout: 480 seconds 1371178271 J * FireEgl ~FireEgl@173-25-83-57.client.mchsi.com 1371181099 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:fda2:462a:27d3:c044 1371181582 Q * thierryp Ping timeout: 480 seconds 1371182678 Q * ircuser-1 Max SendQ exceeded 1371182700 J * ircuser-1 ~ircuser-1@35.222-62-69.ftth.swbr.surewest.net 1371184730 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:b4a2:f0a8:256e:16d6 1371185212 Q * thierryp Ping timeout: 480 seconds 1371186204 Q * DLange Server closed connection 1371186219 J * DLange ~DLange@dlange.user.oftc.net 1371188363 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:7066:1e87:d242:eb01 1371188847 Q * thierryp Ping timeout: 480 seconds 1371190531 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:718c:a47b:6bdd:ad2d 1371190905 J * Ghislain ~aqueos@adsl1.aqueos.com 1371194742 J * SteeleNivenson ~SteeleNiv@pool-96-224-241-140.nycmny.fios.verizon.net 1371194962 M * SteeleNivenson hello all. I have a question about the output of vserver-stat 1371195004 M * SteeleNivenson I have a few dozen hosts running different versions of vserver. some of the hosts output a line for security context 0 from vserver-stat 1371195021 M * SteeleNivenson the remaining hosts do not output a line for that context. 1371195047 M * SteeleNivenson the only similarity I have found so far is the ones that do print CTX 0 have i386 hosts 1371195091 M * SteeleNivenson the ones that do not are x86_64 architecture 1371195133 M * SteeleNivenson my client is concerned that the ones that show security context 0 may have been compromised. 1371196045 M * Ghislain hum i do not show ctx 0 in the stats either 1371196059 M * Ghislain they have all the same kernel and util vserver version ? 1371196216 M * SteeleNivenson good question. lemme script up something to check. 1371196329 M * SteeleNivenson fwiw, the vserver documentation references my situation as expected output from vserver-stat 1371196330 M * SteeleNivenson http://linux-vserver.org/Frequently_Asked_Questions#Which_guest_vservers_are_running.3F 1371197017 M * Ghislain yes it seems but none of mines show them 1371197075 M * SteeleNivenson almost done with the script...2 minutes. 1371197186 M * SteeleNivenson alright, the trend for util-vserver on all the hosts with that output is they have a version <= 0.30.210 1371197239 M * SteeleNivenson Ghislain, what version is your host that doesn't show it? 1371197255 M * Ghislain hum beware if you have a recent kernel and utils that are not the isolation willnot be effective 1371197272 M * Ghislain 3.4.45-vs2.3.3.9 1371197277 M * Ghislain util-vserver: 0.30.216-pre3038 1371197305 M * SteeleNivenson my hosts are very old, highest kernel is 2.6.22 1371197324 M * SteeleNivenson my hosts that don't show it are > 0.30.210 1371197343 M * SteeleNivenson this looks like a version difference. 1371197441 M * Ghislain i know what kernel maintenance is a pain but you should start a program of upgrade starting by the utils and then the kernels and a reboot :) 1371197465 M * SteeleNivenson that's why I have this client 1371197474 M * SteeleNivenson I'm doing this security audit before upgrading. 1371197499 M * SteeleNivenson they are concerned about this suspicious output. Now I have to convince them it's nothing to worry about. 1371197550 M * SteeleNivenson Ghislain, thanks for your help. 1371197858 M * daniel_hozac on old kernels and util-vservers, vserver-stat just summed the ps output. 1371197883 M * SteeleNivenson ah, that would explain the inclusion of context zero. 1371197906 M * daniel_hozac on modern kernels with modern util-vserver (last 5 years or so), it uses vserver APIs to gather its data, and as such doesn't list the host or the spectator. 1371197912 M * SteeleNivenson the coup de grace would be a reference of the source code to show the client as proof. 1371197925 M * SteeleNivenson brillz. that's great news. 1371198055 M * SteeleNivenson daniel_hozac, I'm new to vserver. What is the role of the spectator context? 1371198121 M * Ghislain spectator sees all context 1371198125 M * Ghislain read only 1371198134 M * SteeleNivenson makes sense. 1371198168 M * Ghislain it is restricted if the kernel has privacy guest option 1371198194 M * Ghislain if not it permit from the host to "see" everything going on the different guest 1371199972 M * Ghislain quick stupid question, vmount can umount also ? 1371202947 N * Bertl_zZ Bertl 1371202952 M * Bertl morning folks! 1371203262 M * SteeleNivenson good morning 1371203369 M * Bertl so you're new to Linux-VServer? how do you like it so far? 1371203594 M * SteeleNivenson it's good. I have more experience with Xen but a lot of the terminology is the same "host/guest OS" and "contexts" 1371203656 M * SteeleNivenson my client has a few racks of servers running a very old version of vserver that came along with an acquisition. They've been putting off an upgrade for a long time. 1371203679 M * Bertl just don't think 'xen' think more 'chroot/jail on steroids' 1371203690 M * SteeleNivenson I haven't built a guest vserver from scratch yet, only done analysis on the existing servers. 1371203784 M * SteeleNivenson I read about the architectural differences from emulation. Most of this client's servers are running kvm. They offer Windows hosts in another POP, so they needed real emulation for those. 1371203851 M * SteeleNivenson one question: does vserver support "live" migrations? I assumed I will have to shut down the guests to upgrade but I don't know if that is required. 1371203945 M * Bertl it doesn't support live migration because we consider that a feature which would unnecessarily complicate things 1371203953 M * Bertl but you can do almost live migration 1371203982 M * Bertl or alternatively wrap the host system in a VM to accomplish that 1371203992 M * Ghislain the strength of vserver is that it follow the most efficient KISS principle ^^ 1371204050 M * Bertl i.e. what you normally do with guests if you want to 'migrate' them is to simply rsync the data to a different maching, stop the guest, optionally rsync over again (takes only a few seconds) and bring up the guest on the other host 1371204139 M * SteeleNivenson that makes sense. It'll be great to get some of these guests off Fedora core 4! 1371204147 M * Ghislain add to that arp refresh of the cache or you will have to wait that the router update it 1371204166 M * Bertl SteeleNivenson: guest or host distro? 1371204211 M * SteeleNivenson guest 1371204839 J * distemper ~user@2001:470:50b6:0:ac66:3182:4f74:f680 1371205047 Q * distemper_ Ping timeout: 480 seconds 1371205120 N * l0kit Guest345 1371205128 J * l0kit ~1oxT@0001b54e.user.oftc.net 1371205463 Q * Guest345 Ping timeout: 480 seconds 1371207887 Q * ircuser-1 Read error: Operation timed out 1371208392 Q * Jb_boin Ping timeout: 480 seconds 1371209577 Q * tokkee Server closed connection 1371209581 J * tokkee tokkee@osprey.tokkee.org 1371210185 J * Jb_boin ~dedior@proxad.eu 1371214192 J * BenG ~bengreen@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com 1371216080 M * Bertl off for now ... bbl 1371216084 N * Bertl Bertl_oO 1371218660 M * disposable i would like to run a particular script on a base server when a particular vserver gets started. is there a place to put this script somewhere in /etc/vservers/xyz/ ? 1371218688 M * disposable similarly, i'd like another script to run when somebody shuts that vserver down 1371218702 M * daniel_hozac look at the scripts directory on the great flower page 1371218711 M * daniel_hozac there are several options there. 1371218787 M * disposable daniel_hozac: thank you. 1371218923 J * ircuser-1 ~ircuser-1@35.222-62-69.ftth.swbr.surewest.net 1371221203 Q * thierryp Remote host closed the connection 1371221224 Q * Kabaka Remote host closed the connection 1371221608 J * thierryp ~thierry@home.parmentelat.net 1371221897 J * Kabaka ~Kabaka@04ZAABJ80.tor-irc.dnsbl.oftc.net 1371223502 Q * FireEgl Quit: Leaving... 1371226599 J * hijacker_ ~hijacker@cable-84-43-134-121.mnet.bg 1371227849 M * Ghislain i have a little hashification problem with messages like "Invalid cross-device link", the root FS is an XFS drive 1371227904 M * Ghislain http://paste.linux-vserver.org/23845 1371227919 M * Ghislain seems the file and the hash directory are on the same partition 1371228093 M * Bertl_oO kernel/patch version? 1371228106 M * Ghislain i sue 3.4.45-vs2.3.3.9 1371228162 M * Ghislain s/sue/use/ 1371228189 M * Ghislain stat confirm the hash and file are on the same device 1371228231 Q * BenG Quit: I Leave 1371228236 M * Bertl_oO yeah, I've seen similar on non xfs filesystem, seems that some userspace tools become too smart .. 1371228326 M * Ghislain dam i switched to xfs because quotatools failed on ext4 and now on xfs this is hashification..:( 1371228334 M * Ghislain dam those smart things :D 1371228392 M * Bertl_oO but in this case, the problem originates from the kernel 1371228432 M * Bertl_oO the interesting part is right before line 1 though :) 1371228440 M * Ghislain impossible the kernel cannot be smart 1371228442 M * Ghislain oh 1371228447 M * Ghislain let me give you more then 1371228674 M * Ghislain perhaps this one : http://paste.linux-vserver.org/23846 1371228676 M * daniel_hozac is it checking mounts too? 1371228734 Q * thierryp Remote host closed the connection 1371228892 M * Ghislain does not seems to as far as i can see 1371229031 M * Bertl_oO daniel_hozac: it looks to me like the 'link' is done in the cwd, and I wonder, how that plays with symlinks or bind mounts ... 1371229692 M * Ghislain same kernel on a ext4 it works 1371229729 M * daniel_hozac curious. 1371229730 M * Ghislain differences i can see: one is using ext4 with dlimits, the other xfs with project quota 1371229750 M * Ghislain yes same kernel/utils 1371229807 M * daniel_hozac ah yes 1371229813 M * daniel_hozac project quota will cause that IIRC 1371229845 M * Ghislain no way around this ? 1371229851 M * daniel_hozac different projects can't be hardlinked 1371229858 M * Ghislain dam it 1371229942 M * Ghislain i am screwed again :D 1371229953 M * Ghislain not the tool that try to be too smart, just me 1371230303 M * Bertl_oO :) 1371230332 M * Bertl_oO what was the problem with ext3 and quota again? 1371230338 M * Bertl_oO i.e. disk limits 1371230396 M * Ghislain those disapeared with ext4 but in ext3 the quotatools where being smart and messed the vroot so much i never got the quota to work on the /home this was not on dlimits 1371230434 M * Ghislain the dlimits issue i have i was not able to track down my problem so i switched next install to xfs to have project quota do the job 1371230453 M * Ghislain of course i do not tested hashify on this as i had only one test guest 1371230467 M * Ghislain stupid incomplete test procedure 1371230814 M * Ghislain note that now you can remove ext3 driver to use the ext4 even for ext3 one so perhaps it works with this one 1371230833 M * Ghislain even in ext3 mode, not tested but you can hope so 1371230884 M * Ghislain i must afk thanks for the help 1371230894 M * Ghislain i put a note on the faq about it 1371231709 M * Bertl_oO okay, np, note that ext3 and ext4 are still different codebases 1371231914 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:9d50:a2d8:931a:f7eb 1371232397 Q * thierryp Ping timeout: 480 seconds 1371232955 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:d84c:49ff:f541:1cba 1371233437 Q * thierryp Ping timeout: 480 seconds 1371237008 J * deekeff ~sim@dizko.enavtio.se 1371239009 J * bonbons ~bonbons@2001:a18:20b:a301:6970:42d0:fb26:e1c3 1371240061 Q * hijacker_ Remote host closed the connection 1371242996 J * thierryp ~thierry@2a01:e35:2e2b:e2c0:781a:863a:bbd2:90 1371248309 Q * thierryp Remote host closed the connection 1371252531 J * thierryp ~thierry@home.parmentelat.net 1371252535 Q * bonbons Quit: Leaving 1371253014 Q * thierryp Ping timeout: 480 seconds