1273626834 M * daniel_hozac eja: yes 1273626872 M * daniel_hozac it's network isolation, meaning guests are just restricted in which of the host's IP addresses they can use. 1273627546 J * dowdle_ ~dowdle@scott.coe.montana.edu 1273627656 Q * dowdle_ Remote host closed the connection 1273629306 Q * cehteh Remote host closed the connection 1273630242 N * Bertl_zZ Bertl_oO 1273632494 J * jrklein ~jrklein@2001:0:53aa:64c:0:4a4d:b97d:df43 1273638180 Q * dowdle Remote host closed the connection 1273640857 J * DreamerC_ ~DreamerC@122-116-181-118.HINET-IP.hinet.net 1273640962 Q * DreamerC Ping timeout: 480 seconds 1273641047 J * mtg ~mtg@port-87-193-189-26.static.qsc.de 1273641061 J * ghislain ~AQUEOS@adsl2.aqueos.com 1273641547 Q * balbir Ping timeout: 480 seconds 1273641927 J * derjohn_foo ~aj@d062180.adsl.hansenet.de 1273642358 Q * derjohn_mob Ping timeout: 480 seconds 1273643646 J * ntrs ~ntrs@77.28.7.181 1273644377 J * ghislain1 ~AQUEOS@adsl2.aqueos.com 1273644608 Q * derjohn_foo Ping timeout: 480 seconds 1273644664 Q * ghislain Ping timeout: 480 seconds 1273644787 Q * ncopa Ping timeout: 480 seconds 1273645102 N * DreamerC_ DreamerC 1273645856 Q * Romster Ping timeout: 480 seconds 1273646059 J * dna ~dna@122-213-103-86.dynamic.dsl.tng.de 1273646733 J * petzsch ~markus@dslb-092-078-238-227.pools.arcor-ip.net 1273647570 A * ard sighs.... 1273647588 A * ard now notices that he asked on the wrong channel... 1273647943 J * ncopa ~ncopa@180.40.189.109.customer.cdi.no 1273648528 J * derjohn_foo ~aj@213.238.45.2 1273648704 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1273649010 Q * petzsch Quit: Leaving. 1273649321 J * balbir ~balbir@122.248.163.1 1273649404 J * petzsch ~markus@dslb-092-078-238-227.pools.arcor-ip.net 1273650779 Q * balbir Ping timeout: 480 seconds 1273650861 J * sharkjaw ~gab@90.149.128.29 1273652025 Q * petzsch Quit: Leaving. 1273652878 J * kir ~kir@swsoft-msk-nat.sw.ru 1273653620 Q * sharkjaw Remote host closed the connection 1273654094 J * Romster ~romster@202.168.100.149.dynamic.rev.eftel.com 1273656159 J * balbir ~balbir@122.248.161.59 1273657089 Q * balbir Ping timeout: 480 seconds 1273657985 Q * ghislain1 Quit: Leaving. 1273661084 T * * http://linux-vserver.org/ |stable 2.2.0.7, exp 2.3.0.36.28, grsec 2.3.0.36.28|util-vserver-0.30.216-pre2864| 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 forget about the minute. 1273661084 T * Bertl - 1273661363 J * sharkjaw ~gab@90.149.128.29 1273661403 Q * Bertl_oO Read error: Connection reset by peer 1273661404 J * Bertl_oO herbert@IRC.13thfloor.at 1273661433 J * petzsch ~markus@dslb-092-078-238-227.pools.arcor-ip.net 1273661501 J * BobR_zZ odie@IRC.13thfloor.at 1273662303 J * balbir ~balbir@122.248.163.1 1273662708 Q * BenG Quit: I Leave 1273665015 Q * ntrs Read error: Connection reset by peer 1273665026 J * ntrs ~ntrs@77.28.30.182 1273665431 Q * balbir Ping timeout: 480 seconds 1273665753 Q * sharkjaw Quit: Leaving 1273666722 J * balbir ~balbir@122.248.163.1 1273667937 J * cehteh ~ct@pipapo.org 1273668124 Q * nkukard Ping timeout: 480 seconds 1273668560 J * imcsk8 ~ichavero@221.pool85-58-0.dynamic.orange.es 1273668875 J * Loki_muh ~loki@213.133.125.66 1273668875 Q * Loki|muh Read error: Connection reset by peer 1273668887 N * Loki_muh Loki|muh 1273668893 Q * jrklein Quit: jrklein 1273668945 J * jrklein ~jrklein@2001:0:53aa:64c:0:4a4d:b97d:df43 1273669065 J * ghislain ~AQUEOS@adsl2.aqueos.com 1273669428 Q * jrklein Ping timeout: 480 seconds 1273670843 Q * petzsch Quit: Leaving. 1273670843 J * jrklein ~jrklein@2001:0:53aa:64c:0:1212:b957:8a5b 1273671692 Q * imcsk8 Quit: This computer has gone to sleep 1273671989 Q * balbir Ping timeout: 480 seconds 1273672089 J * thierryp ~thierry@dsl092-234-022.phl1.dsl.speakeasy.net 1273673320 N * Bertl_oO Bertl 1273673347 M * Bertl morning folks! 1273674657 J * barismetin ~barismeti@tvwna-ip-c-158.princeton.org 1273674698 J * imcsk8 ~ichavero@221.pool85-58-0.dynamic.orange.es 1273675618 Q * derjohn_foo Ping timeout: 480 seconds 1273675714 Q * imcsk8 Quit: This computer has gone to sleep 1273676254 J * vserver_guy ~vserver@mirror.sonassi.com 1273676346 Q * vserver_guy 1273676535 J * vServer_User_Zz ~vserver@mirror.sonassi.com 1273677129 Q * vServer_User_Zz Remote host closed the connection 1273678360 J * nkukard ~nkukard@groupware-0-1.virt.iitsp.net 1273679188 J * derjohn_foo ~aj@88.128.81.171 1273679235 J * dowdle ~dowdle@scott.coe.montana.edu 1273680362 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1273680434 M * eja daniel_hozac: but the IP still resides on the host right? 1273680463 M * Bertl by default, IPs, routing, etc all happens on the host 1273680481 M * Bertl the only thing done for the guest is limiting the available set of IPs 1273680732 M * eja is it possible to have the IPs not set on the host? i'd prefer they reside on the guest and all the host sees is a tun interface. 1273680775 M * Bertl nope 1273680793 M * Bertl you can switch to network namespaces, which will give you guest private interfaces 1273680804 M * Bertl but that requires a routing or bridging setup on the host 1273680828 M * eja i already have a bridging setup for the virtualbox guests. 1273680995 M * ard or a vlan for just that guest ;-) 1273681028 A * ard prefers the vserver way for plain webservers and stuff 1273681054 M * ard but I was thinking about a network namespace for a bunch of vservers on a specific vlan 1273681069 M * ard and another network namespace for another group of vservers ;-) 1273681100 M * eja well i have a tiered setup now with core, management etc vlans and i'd like to not violate that by adding a bunch of IPs to the core machines. 1273681128 M * eja it'll mess up my routing and firewall policies. they could always be redone though if necessary. 1273681136 M * ard if you are going to route through the core machine there is no real difference 1273681147 M * eja are network namespaces a linux-vserver construct or part of the linux kernel? 1273681184 M * eja if there's no IP on the host then it's essentially unreachable. i can always just firewall it but it would still mess up routing. 1273681186 M * ard I have a pretty good firewalling setup on a plain vserver system that controls traffic between the vservers and out and ingoing 1273681200 J * vserver_guy ~vserver@mirror.sonassi.com 1273681205 M * vserver_guy hi guys 1273681223 M * ard network namespace is part of linux and you can enable it in util-vserver 1273681226 M * vserver_guy what advantages are there to using a newer kernel for vserver (2.27 vs 2.27) ? 1273681233 M * ard since 2008 util vserer vserion *215 1273681234 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg 1273681236 M * eja at this point i'm really just trying to evaluate vserver, so i'd prefer to get it up and running w/ minimal changes to everything else. 1273681240 M * ard as daniel_hozac told me :-) 1273681257 M * ard allright: touch spaces/net in your vserver config directory 1273681264 M * Bertl vserver_guy: (2.27 vs 2.27)? 1273681276 M * vserver_guy oh, I meant 2.26 1273681280 M * ard start the vserver, and find some way to transplant your virtual interface to your namespace 1273681295 M * ard ip link set netns 1273681301 M * Bertl vserver_guy: what are you referring to with those numbers? 1273681307 M * vserver_guy kernel 1273681318 M * Bertl so you mean, 2.6.26 vs 2.6.27? 1273681322 M * vserver_guy yes, sorry 1273681336 M * ard eja: be sure to add linux netadmin capabilities to your setup. 1273681347 M * Bertl vserver_guy: both are broken and outdated, so doesn't matter which one you choose 1273681351 M * vserver_guy lol 1273681355 Q * Romster Read error: No route to host 1273681378 M * ard eja: I am busy with exact the same setup, so bare with me, friday latest I have a working solution (else my girlfriend kills me) 1273681379 M * vserver_guy what would you reccomend 1273681388 M * Bertl 2.6.33 should be fine 1273681416 J * Romster ~romster@202.168.100.149.dynamic.rev.eftel.com 1273681426 M * ard if you do bridging vlan on bonds on a bnx2 driver don't :-) but that's a kernel bug, nothing to do with vserver 1273681429 J * petzsch ~markus@p4FF47B2B.dip.t-dialin.net 1273681485 M * ard (bugs in probably the bnx2 driver, but could also be bond or bridge) 1273682296 M * Bertl off for now .. bbl 1273682303 N * Bertl Bertl_oO 1273682437 Q * ntrs Ping timeout: 480 seconds 1273682713 Q * jrklein Ping timeout: 480 seconds 1273682853 J * manana ~mayday090@84.17.25.149 1273684314 Q * vserver_guy Ping timeout: 480 seconds 1273684336 Q * dna Quit: Verlassend 1273684381 Q * derjohn_foo Ping timeout: 480 seconds 1273684507 J * derjohn_mob ~aj@88.128.81.171 1273684516 J * vServer_User_Zz ~vserver@mirror.sonassi.com 1273684831 Q * mtg Quit: Verlassend 1273685114 Q * vServer_User_Zz Remote host closed the connection 1273685121 M * eja so the supported method of networking for vserver is to add an alias to an existing interface on the host and then that's shared/routed to the guest? 1273685128 J * vserver_guy ~vserver@mirror.sonassi.com 1273685192 M * ard no... it's not routed to... 1273685206 M * ard the guest has a filtered view of all ip's 1273685218 M * ard I usually put all ip's on lo with a /32 1273685233 M * ard --interface blaat=lo:x.y.z.b/32 1273685242 M * ard then you get lo:blaat ... 1273685259 M * ard the guest will only see the two ip's on device lo 1273685275 M * ard the guest is not allowed to alter add/change ip's or routing 1273685321 M * ard it's very restrictive, so if your vserver get's rooted it can't harm your network by taking over identities 1273685337 J * jrklein ~jrklein@2001:0:53aa:64c:0:1212:63e5:6731 1273685393 M * ard It will only hurt you if you intend to run programs like dhcpd... (I have a version that runs well within a restricted vserver, you just need to run dhcrelay on the routers) 1273685417 M * ard But the usual LAMP is perfectly fine and secure with vservers 1273685476 M * ard but it might be that you also want network seperation between multiple vservers (on different vlans f.i.), then you need the network namespace 1273685545 M * ard basically: if your host only has a single ip, and your vservers are in the same network as the host, you can just install util-vserver with the right kernel, and do a vserver build -m --interface 1273685553 M * ard and you have jailed a fresh installation 1273685635 M * eja i have an ubuntu 8.04 debootstrap installed. just struggling with getting the networking functioning. 1273685728 Q * vserver_guy Remote host closed the connection 1273685742 J * vserver_guy ~vserver@mirror.sonassi.com 1273686059 M * _Shiva__ ard: do you have a pointer on the bnx2/vlan/bond/bridge thing? 1273686342 Q * vserver_guy Remote host closed the connection 1273686356 J * vserver_guy ~vserver@mirror.sonassi.com 1273687222 Q * kir Quit: Leaving. 1273687526 J * balbir ~balbir@122.172.157.52 1273687940 J * ntrs ~ntrs@77.28.171.129 1273688396 Q * petzsch Quit: Leaving. 1273688444 M * ard yeah 1273688610 M * ard http203://paste.linux-vserver.org/15152 1273688715 M * ard http://paste.linux-vserver.org/15152 1273689168 Q * jrklein Ping timeout: 480 seconds 1273689169 J * jrklein ~jrklein@2001:0:53aa:64c:0:1212:63e5:6731 1273689653 J * crushedhat ~chatzilla@216.57.76.186 1273690416 Q * derjohn_mob Ping timeout: 480 seconds 1273690578 Q * jrklein Quit: jrklein 1273690582 J * derjohn_mob ~aj@88.128.81.171 1273690859 J * jrklein ~jrklein@2001:0:53aa:64c:0:1212:63e5:6731 1273691782 Q * thierryp Remote host closed the connection 1273692604 M * _Shiva__ ard: ah - tnx.. but i mean, is there any (more) /official/ bug @kernel.org or @lkml ot the like..? :-) 1273692653 M * _Shiva__ s/ot/or/ 1273692743 Q * hijacker_ Quit: Leaving 1273692795 Q * ignaz Remote host closed the connection 1273693070 Q * derjohn_mob Remote host closed the connection 1273693263 J * petzsch ~markus@p4FF47B2B.dip.t-dialin.net 1273696345 Q * manana Remote host closed the connection 1273696364 Q * FireEgl Quit: Leaving... 1273696666 Q * jrklein Quit: jrklein 1273696715 J * ntrs_ ~ntrs@77.28.160.136 1273697158 Q * ntrs Ping timeout: 480 seconds 1273697487 J * jrklein ~jrklein@2001:0:53aa:64c:0:1212:63e5:f7b0 1273698616 Q * ghislain Quit: Leaving. 1273698819 M * crushedhat I'd like to ask a question about the VS init process. Anyone currently online who knows this area? 1273698834 M * Bertl_oO nope :) 1273698863 N * Bertl_oO Bertl 1273698889 M * crushedhat hi Bertl, nice to see you 1273698907 M * Bertl crushedhat: just ask, don't ask to ask or if somebody is here to answer, you'll get an answer if you simply wait 1273698914 M * crushedhat I’m running VS on Fedora 12, with a kernel built from Fedora’s latest (2.6.32.11-99.fc12.x86_64) and VS patch 2.6.32.11-vs2.3.0.36.29.4, and utils 0.30.216-pre2883. 1273698940 M * crushedhat When I stop or start a guest I get several “initctl: Unable to send message: Connection refused” messages. Also “shutdown: Unable to send message: Connection refused” when I attempt to shutdown or reboot from inside the guest. strace shows that these are both trying to send to a an abstract Unix net port “/com/ubuntu/upstart”. 1273698990 M * Bertl which means you are running the guest in sysv mode, without a guest init 1273698992 M * crushedhat As I understand it, the init process should be listening on this, but apparently isn't. This seems to have something to do with VS guests not having Unix domain abstract sockets in a separate namespace 1273699000 M * crushedhat plain mode 1273699022 M * Bertl kernel/patch/util-vserver version? 1273699033 M * crushedhat above... 1273699065 M * Bertl okay, so a distro kernel? 1273699084 M * crushedhat yes, plus the VS patch 1273699107 M * Bertl did you run testme.sh to do basic tests? 1273699141 M * crushedhat yes, but I'll try it again... 1273699170 Q * petzsch Quit: Leaving. 1273699181 M * Bertl guest distro is? 1273699195 M * Bertl ubuntu karmic? 1273699243 M * daniel_hozac if you are using plain initstyle, you wouldn't see messages on stop. 1273699244 M * crushedhat Linux-VServer Test [V0.17] Copyright (C) 2003-2006 H.Poetzl 1273699246 M * crushedhat chcontext is working. 1273699248 M * crushedhat chbind is working. 1273699249 M * crushedhat Linux 2.6.32.11-99.nuMetra.1.fc12.x86_64 #1 SMP Tue Apr 27 17:26:36 PDT 2010 x86_64 1273699251 M * crushedhat Ea 0.30.216 236/glibc (DSa) 1273699252 M * crushedhat VCI: 0002:0305 236 13010fb1 (TbsPHIWD) 1273699254 M * crushedhat --- 1273699255 M * crushedhat [000]# succeeded. 1273699256 M * Bertl (please use paste.linux-vserver.org for everything longer than 3 lines) 1273699257 M * crushedhat [001]# succeeded. 1273699258 M * crushedhat [011]# succeeded. 1273699260 M * crushedhat [031]# succeeded. 1273699261 M * crushedhat [101]# succeeded. 1273699263 M * crushedhat [102]# succeeded. 1273699264 M * crushedhat [201]# succeeded. 1273699266 M * crushedhat [202]# succeeded. 1273699267 M * Wonka O.o 1273699270 M * mnemoc /o\ 1273699289 M * crushedhat sorry... don't use IRC much. Wasn't expecting the separate msgs 1273699325 M * mnemoc what did you expect? a huge single line concatenated? 1273699329 M * crushedhat guest distro is also Fedora 12 1273699349 M * Bertl f12 uses upstart? 1273699373 M * daniel_hozac yes 1273699414 M * crushedhat kinda. The infrastructure is upstart, but most of the real work is still done via sysv-style scripts. Not native like Ubuntu 1273699475 M * crushedhat for example, if I do a simple cmd like "initctl list", it gets all the info from the /com/ubuntu/upstart abstract Unix socket 1273699487 M * Wonka mnemoc: Jabber has multiline messages... 1273699496 M * Bertl auto single IP configured? 1273699515 M * mnemoc Wonka: sure, but that still uses several lines, even if the "nick" is not repeated on each 1273699528 M * crushedhat yes. Actually I've got one IPv4 and one IPv6 1273699598 M * crushedhat Does the IP networking interact with the Unix domain sockets? 1273699609 M * Bertl nope 1273699630 M * Bertl but the 'permission denied' part could be from a misconfiguration 1273699712 J * petzsch ~markus@p4FF47B2B.dip.t-dialin.net 1273699796 M * crushedhat I get lots of "kernel: imklog: Cannot open proc file system, 1." in /var/log/messages on the guest. This seems to be from rsyslog, that doesn't have perm on /proc/kmsg. I don't know whether this is related. Should guests have permission on /proc/kmsg? Wouldn't this make every guest log all kernel messages? 1273699835 M * Bertl nope, by default, guests do not have access to /proc/kmsg 1273699854 M * Bertl you can enable a virtualization which gives you a dummy /proc/kmsg 1273699872 M * Bertl but the better approach is to configure syslog correctly 1273699874 M * crushedhat when you asked about "permission denied" are you referring to "connection refused", or something else? 1273699905 M * crushedhat configure rsyslog to not try to open /proc/kmsg? 1273699927 M * Bertl correct, the default guest install scripts should fix that for you 1273700060 M * crushedhat looks like I'd need to comment out "$ModLoad imklog.so" in /etc/rsyslog.conf. Default scripts didn't do this 1273700170 M * crushedhat That seems to have fixed the imklog messages. All this is unrelated to the /com/ubuntu/upstart issue, I assume. 1273700212 M * Bertl probably, what is the guest configuration? 1273700445 M * crushedhat I created using "vserver build -m yum -- -d f12" and other options for IP, rootdir, etc. Don't remember the exact cmdline now. The f12 distrib files were copied from the f11 ones, with minor changes to adjust the URLs of yum repos. 1273700457 M * crushedhat Is that what you're asking? 1273700507 J * thierryp ~thierry@tvwna-ip-b-13.princeton.org 1273700528 Q * petzsch Quit: Leaving. 1273700544 M * Bertl daniel_hozac: doesn't recent util-vserver contain f12 configs? 1273700577 M * crushedhat not the latest I saw, unless they got added in the last few days 1273700599 M * daniel_hozac no, i haven't had any system to test them on until now. 1273700627 M * Bertl i.c. okay, so probably some manual cleanup is required 1273700629 M * daniel_hozac special rpm needed and all. 1273700648 M * crushedhat special rpm? 1273700681 M * daniel_hozac for non-F12 systems. 1273700744 J * imcsk8 ~ichavero@221.pool85-58-0.dynamic.orange.es 1273700902 M * crushedhat do you know of a significant change F11 -> F12 that would affect the Unix networking? 1273700903 M * Bertl crushedhat: so you can trigger the 'connection refused' with a simple command (from inside the guest)? 1273700938 M * daniel_hozac we want /proc/virtual//{info,status} too 1273701004 M * crushedhat yes, the "initctl list" is a good example. From an strace: 1273701006 M * crushedhat sendmsg(3, {msg_name(22)={sa_family=AF_FILE, path=@"/com/ubuntu/upstart"}, msg_iov(1)=[{"upstart\n\0\0\1\0S", 13}], msg_controllen=0, msg_flags=0}, 0) = -1 ECONNREFUSED (Connection refused) 1273701065 M * crushedhat netstat shows that there is no process listening on the socket. On the host, init is listening on it. 1273701213 M * Bertl so, you're sure upstart is running inside the guest? 1273701268 M * crushedhat http://paste.linux-vserver.org/15159 1273701328 M * daniel_hozac that guest doesn't have its own init. 1273701430 M * crushedhat no, there is no command called upstart, but the init process is upstart. /sbin/init is actually in the upstart package on Fedora 12 1273701458 Q * thierryp Remote host closed the connection 1273701460 M * crushedhat there is an init process running in the guest, since it's using plain init 1273701489 Q * barismetin Remote host closed the connection 1273701497 M * daniel_hozac no, there's not. 1273701512 M * daniel_hozac it is not using plain init. 1273701620 M * crushedhat I mean that it is using the plain init style for vserver, not sysv, gentoo, etc. 1273701646 M * crushedhat Unless I'm confused and it's really running a different style 1273701677 M * Bertl you're probably confused 1273701703 J * derjohn_mob ~aj@e180213139.adsl.alicedsl.de 1273701859 M * daniel_hozac that guest is definitely not plain. 1273702099 Q * bonbons Quit: Leaving 1273702256 Q * ntrs_ Ping timeout: 480 seconds 1273702257 Q * jrklein Quit: jrklein 1273703020 Q * imcsk8 Quit: This computer has gone to sleep 1273703035 M * crushedhat How can you tell? From the status flags? 1273703077 M * crushedhat The guest definitely has its own init process 1273703096 M * Bertl we seriously doubt that :) 1273703101 Q * derjohn_mob Ping timeout: 480 seconds 1273703256 M * crushedhat Are you saying that ps is showing me something fake, that's really the host's init? I attached guest ps output to the paste. 1273703269 Q * dowdle Remote host closed the connection 1273703270 M * Bertl correct, that is the blend through init 1273703279 J * dowdle ~dowdle@scott.coe.montana.edu 1273703293 M * Bertl it is there to make processes like pstree or ps happy 1273703419 M * crushedhat So I guess I am in sysv mode, since that is the default. 1273703442 M * Bertl yep, we assume so too 1273703535 M * crushedhat Ok, so is there a way to make the /com/ubuntu/upstart socket work in sysv init mode? 1273703555 M * crushedhat Or do you suggest I try plain mode instead? 1273703597 M * crushedhat It looks like sysv mode is simpler and less overhead, so I think I'd prefer to run that if I can. 1273703602 M * Bertl well, if you can convice upstart to run with a different pid, started from a sysv script, I don't see a reason why not 1273703666 M * Bertl but IIRC, upstart is quite immature and doesn't handle anything outside the 'expected' init 1273703722 M * crushedhat I guess what I'm asking is: how is my config different from what most others are running? How does this work for everyone else? Is there something about F12 that's not quite supported correctly yet? 1273703769 M * Bertl I'd say most folks don't use f12 guests yet 1273703799 M * Bertl and those who do, most likely use the plain init style to have upstart running as init 1273703862 M * crushedhat What is the default for Ubuntu guests, that depend on upstart more? Is it plain init style? 1273703887 M * Bertl yes, recent ubuntu uses plain init style 1273703907 M * crushedhat I guess something must have changed between F11 and F12 related to upstart. 1273703959 J * selim ~chatzilla@e181086244.adsl.alicedsl.de 1273703961 M * Bertl f11 didn't use upstart, or? 1273704047 M * crushedhat I just tried "initctl list" on an F11 box, and it also sends messages to /com/ubuntu/upstart abstract Unix domain socket. 1273704054 P * selim 1273704094 M * daniel_hozac upstart has been in Fedora since Fedora 9. 1273704176 M * crushedhat I thought it had been there awhile. So do F9, F10, and F11 users also have these issues with VS when using sysv init? 1273704236 M * daniel_hozac "issues"? the post install script tries to silence the spew, but it might not work for F12 yet. 1273704366 M * crushedhat ah, I see. I get lots of these msgs every time I start and stop a guest. Also, stopping hangs for several seconds. I'll paste output. How is the spew suppressed? Or is it only output from build that is suppressed? 1273704498 M * crushedhat pasted 1273704726 M * crushedhat If I want to try plain init style, can I just change it in the config, or do I have to rebuild the guest? That is, are there large changes in what's in the filesystem between the two styles? 1273704782 M * daniel_hozac none 1273704863 M * daniel_hozac just do echo plain > /etc/vservers//apps/init/style 1273705518 M * crushedhat If I use plain init style, there are no initctl warnings during start/stop. However, start doesn't start any daemons such as sshd or rsyslog, even though these are enabled (chkconfig). Do I need to run runit for the startup scripts to run correctly? 1273705780 M * crushedhat The util-vserver:InitStyles wiki page isn't very clear about much of this. I'd offer to update it, but I don't understand it enough myself yet. 1273705786 M * Bertl I'd say the init process, together with inittab shand a runlevel script is supposed to handle that 1273705798 M * Bertl s/shand/and/ 1273705815 M * Bertl i.e. if you choose sysv, then util-vserver will execute the runlevel scripts 1273705838 M * Bertl if you select plain init style, then init will be started, and is supposed to deal with the rest 1273705854 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1273706152 M * crushedhat In sysv init style, the daemons start but I get all the initctl errors. In plain init style, no initctl errors but no daemons either. I can start them ok manually. I don't know where the log is for the startup process. It's not the guest's /var/log/messages--since rsyslog doesn't start, there's nothing logged there. 1273706246 M * Bertl well, if you use upstart as init, you have to check with the upstart documentation how to get debug output 1273706304 M * Bertl putting a pty or tty into the guest as console might be a quick workaround 1273706320 Q * BenG Quit: I Leave 1273706429 M * crushedhat Ok, I'm willing to do this debugging if necessary. Does this all work on F11 (or even F10 or F9), or are there similar problems? It might help if I can focus on what changed that broke something. 1273706577 M * daniel_hozac you don't usually use plain with Fedora. 1273706693 M * daniel_hozac you might want to look at the debian initpost script for some hints on upstart though. 1273706740 M * crushedhat ok, I started out with sysv style, and saw the initctl errors. Maybe I should use sysv style, and just ignore them? How are they suppressed in F11 and older? 1273706823 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com 1273706845 M * daniel_hozac ignoring them works. 1273706893 M * daniel_hozac /sbin/initctl.vserver is used instead. 1273707063 M * crushedhat Are they suppressed in F11 somehow? I thought I saw an earlier chat log where someone said this "should be fixed". 1273707235 M * daniel_hozac you tell me 1273707352 M * crushedhat It would take me a while to get VS running on F11 or older. I was wondering if you know of others using this successfully. 1273707383 M * daniel_hozac the host has no relevance. 1273707465 M * crushedhat oh, ok. I'll try building some F11 and older guests. I guess they won't mind the F12 kernel. 1273707761 M * crushedhat What do you see as the long-term solution to this? The "right" way to fix it? 1273708475 Q * dowdle Remote host closed the connection 1273708673 M * Bertl off to bed now .. have a good one everyone! 1273708687 N * Bertl Bertl_zZ