1319587439 Q * bonbons Quit: Leaving 1319588548 Q * treaki Ping timeout: 480 seconds 1319592912 Q * clopez Ping timeout: 480 seconds 1319594042 J * nicholi ~nicholi@rrcs-76-79-196-34.west.biz.rr.com 1319596243 Q * ksn Ping timeout: 480 seconds 1319596451 J * mib_flny22 44cde00a@ircip2.mibbit.com 1319596749 P * mib_flny22 1319596833 Q * hparker Ping timeout: 480 seconds 1319597102 J * puck ~puck@leibniz.catalyst.net.nz 1319597376 M * puck hmmm, I'm trying to update a pre-start script that looks in the directory /usr/local/share/fai (on the host) for every directory in there, makes a directory in the guest, and then bind mounts it. 1319597408 M * puck If run the logic by hand as root, it works, but in the pre-start it sees an empty /usr/local/share/fai. /usr/local/share/fai is a mount point. 1319597430 M * puck It seems as though I'm not seeing the mounted filesystem 1319597480 J * ksn ~ksn@197.168.199.196 1319597492 M * puck Yes, that is the case. I just unmounted the filesystem, touch a file, then remounted it. the pre-start script sees the file, not the files within the mount point 1319597512 M * puck This is ubuntu lucid 1319597560 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1319599089 Q * ksn Ping timeout: 480 seconds 1319603185 J * black ~black@02d910f8.bb.sky.com 1319603718 Q * black Ping timeout: 480 seconds 1319604269 J * sannes1 ~ace@cm-84.209.106.118.getinternet.no 1319607387 J * derjohn_mob ~aj@213.238.45.2 1319608917 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1319610484 Q * ncopa Quit: Leaving 1319610595 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1319611960 J * jeroen__ ~jeroen@095-097-051-172.static.chello.nl 1319612286 J * ghislain ~AQUEOS@adsl2.aqueos.com 1319614010 M * daniel_hozac puck: add the path to the cleanup-skip file. 1319614597 N * bertl_zZ Bertl 1319614606 M * Bertl morning folks! 1319616188 Q * brc Ping timeout: 480 seconds 1319616668 M * hijacker morning 1319618216 M * Bertl off for now ... bbl 1319618220 N * Bertl Bertl_oO 1319618386 J * bonbons ~bonbons@2001:960:7ab:0:ad8c:4806:e214:ff13 1319618848 Q * fisted Ping timeout: 480 seconds 1319618850 J * fisted ~fisted@xdsl-87-78-221-210.netcologne.de 1319618867 Q * mcp Remote host closed the connection 1319619495 J * mcp ~mcp@wolk-project.de 1319619873 J * treaki ~treaki@p5B0335E9.dip.t-dialin.net 1319620472 J * clopez ~clopez@155.99.117.91.static.mundo-r.com 1319622905 J * brc ~bruce@72.20.27.65 1319624897 Q * Aiken Ping timeout: 480 seconds 1319626000 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1319627263 J * ksn ~ksn@197.109.25.238 1319627516 M * ncopa hi 1319627522 M * ncopa any patch for 3.0.8? 1319628689 M * daniel_hozac 3.0.7 doesn't apply? 1319629448 Q * hparker Quit: Quit 1319629778 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1319631924 N * Bertl_oO Bertl 1319631929 M * Bertl back now ... 1319633071 J * black ~black@02d910f8.bb.sky.com 1319634189 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1319634308 J * thierryp ~thierry@home.parmentelat.net 1319634468 Q * ksn Ping timeout: 480 seconds 1319635819 A * arekm looks at new patches 1319635917 M * arekm hm, but there are no new patches :) 1319636537 Q * BenG Quit: I Leave 1319636639 J * alpha_one_x86 ~kvirc@201.222.115.4 1319636694 M * alpha_one_x86 Hello, on the host with: -A OUTPUT -s 127.0.0.0/8 -d 127.0.0.0/8 -j ACCEPT the guest can't connect to the 127.0.0.1 but work with -A OUTPUT -j ACCEPT 1319636728 M * alpha_one_x86 I'm with the 2.6.38 vs kernel and util vserver 0.30.216_pre2955 1319636763 M * Bertl well, does the guest use 127.* for the connection? 1319636827 M * alpha_one_x86 Yes, telnet 127.0.0.1 3306 (mysql) 1319636877 M * daniel_hozac does the guest have single_ip set? 1319636888 M * alpha_one_x86 how check that's? 1319636898 M * daniel_hozac nattribute --get --nid 1319636904 M * Bertl what I meant was does the guest _use_ 127.* as source ip 1319636907 J * ksn ~ksn@197.169.222.67 1319636949 M * alpha_one_x86 nflags: single_ip yes 1319637226 Q * Aiken Remote host closed the connection 1319637853 M * Bertl off for a nap ... bbl 1319637859 N * Bertl Bertl_zZ 1319637941 M * daniel_hozac so you have told the guest to use just one IP. it's using its first IP as the source address regardless of where it connects. 1319638993 Q * ksn Ping timeout: 480 seconds 1319639513 Q * FireEgl Ping timeout: 480 seconds 1319640446 M * alpha_one_x86 mmm, ok 1319641960 J * tetr ~x@fh212n1-g-ml-a11.ias.bredband.telia.com 1319642058 M * tetr hi! is there any functionality for moving/copying/migrating linux-vserver guests to another server, while the guest still being online? 1319642127 Q * ncopa Quit: Leaving 1319644395 J * dowdle ~dowdle@scott.coe.montana.edu 1319644762 M * tetr at least it's possible with openvz to migrate a vz to another host node without even shutting it down - it copies all used memory as well - all live 1319644779 M * tetr but i would rather use linux-vserver than openvz.. 1319645355 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1319645449 M * ser no, rather not 1319645488 M * tetr why? :) 1319645537 M * ser because it is not a full virtualization solution like openvz 1319645615 M * ser but a delay needed to move it is very short, just one restart 1319645628 M * ser for me it is completely enough :) 1319646023 M * tetr avoiding restarts of a guest is a must for me im afraid 1319646093 Q * FireEgl Read error: Connection reset by peer 1319646360 M * ser /j #openvz ;-) 1319646392 M * alpha_one_x86 +1, down time is < 1min in my case with vserver 1319646896 J * FireEgl FireEgl@2001:470:e056:1:787a:ae0a:f281:3071 1319647268 M * ser kworker/0:1: page allocation failure: order:1 - completely new oops, after upgrading to 3.0.4 1319647741 M * tetr yeah but a restart resets connections etc 1319647749 M * tetr simply annoying 1319647751 M * tetr :) 1319647828 M * alpha_one_x86 ser: warning the kernel 3.0.6 crash very often, is not stable for server production 1319647864 M * ser but as far as i can understand it mnay be related to e1000 driver :( 1319648504 Q * alpha_one_x86 Quit: KVIrc KVIrc Equilibrium 4.1.1, revision: 5816, sources date: 20110403, built on: 2011-06-11 13:01:00 UTC http://www.kvirc.net/ 1319648608 Q * derjohn_mob Ping timeout: 480 seconds 1319649042 Q * guerby Ping timeout: 480 seconds 1319649221 J * guerby ~guerby@nc10d.tetaneutral.net 1319649602 Q * thierryp Remote host closed the connection 1319650118 Q * clopez Ping timeout: 480 seconds 1319650183 Q * FireEgl Remote host closed the connection 1319650584 M * daniel_hozac tetr: you can always layer something like KVM underneath, since i assume you're doing it for hardware failover. 1319650940 J * FireEgl FireEgl@2001:470:e056:1:c93f:b0f8:b63a:172 1319651235 Q * FireEgl Read error: Connection reset by peer 1319651451 M * tetr currently im looking at a setup of openvz with drbd in primary-primary configuration, using ocfs2 as fs on the drbd device.. both of the machines (host nodes) run their own set of guests, but in case of i.e. a kernel upgrade, or other planned maintenance for one of the machines, all guests are easily migrated to the other machine, without any downtime 1319651555 M * tetr for unplanned downtime (i.e. crashes) all guests obviously will have to be restarted on the other machine - in that situation linux-vserver would also be able to handle it i guess 1319651835 M * ser tetr: please notice you need ~two times more expensive hardware for that solution 1319651854 M * ser just to handle all these extra layers 1319651896 M * ser linux-vserver is cheap, very cheap, and designed to get the most of your servers, but of course is not a solution for everyone 1319652036 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1319652100 M * tetr well it already works with openvz. i'm sure it would work with linux-vserver also, if it just had the online migration of the guests 1319652143 M * arekm can openvz do migration without having shared fs on both machines? 1319652221 M * tetr yeah 1319652259 M * tetr 1. rsync 2. freeze guest 3. rsync 4. copy memory 5. unfreeze on other server 1319652527 N * Bertl_zZ Bertl 1319652564 M * Bertl tetr: you can simply put a kvm (or xen if you prefer) layer around Linux-VServer and tada! you have live migration 1319652607 M * Bertl also cgroups and mainline are working on at least a snapshot cgroup system (freezer) which will, once completed, allow to do that without a kvm/xen wrapper 1319652725 M * tetr dont want another layer like that tho, the latter sounds nice.. i've heard more and more about cgroups recently 1319652740 M * tetr do you think linux-vserver will use cgroups? 1319652778 Q * FireEgl Ping timeout: 480 seconds 1319652823 M * Bertl tetr: Linux-VServer uses cgroups since the beginning 1319652839 M * Bertl i.e. since they stopped crashing every 5 minutes :) 1319652884 A * arekm would love to see nice net namespace support ;P 1319652936 M * tetr oh :) 1319654404 M * tetr Bertl: Freezer cgroup subsystem seems to be in 3.0.8 btw 1319654795 M * Bertl well, if it works, then you can probably already use it with Linux-VServer 1319654856 M * daniel_hozac the freezer just pauses the processes. 1319654890 M * daniel_hozac C/R isn't in the kernel yet, i don't think. 1319654896 M * Bertl well, userspace could then do the migration 1319654921 M * Bertl IIRC, there was a userspace process migration/snapshot project some time ago 1319655040 M * tetr whats C/R? 1319655057 M * daniel_hozac checkpoint restart 1319655151 M * Bertl I remember the time (a few years ago) when IBM announced to add that within a few months, because they absolutely need it :) 1319655198 M * daniel_hozac yeah... isn't that the original reason cgroups were created? 1319655259 M * Bertl yes, indeed 1319655360 M * Bertl daniel_hozac: did you get around looking into the CoW code for 3.1? 1319655389 M * daniel_hozac not yet... 1319655404 M * Bertl no problem, just wanted to know ... 1319655450 Q * nicholi Remote host closed the connection 1319655593 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1319657188 Q * FireEgl Remote host closed the connection 1319658619 M * puck daniel_hozac: woohoo! namespace-cleanup-skip works! 1319660291 Q * sannes1 Remote host closed the connection 1319661394 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1319661568 J * ksn ~ksn@197.169.236.24 1319662452 J * cuba33ci_ ~cuba33ci@111-240-164-149.dynamic.hinet.net 1319662803 Q * cuba33ci Ping timeout: 480 seconds 1319662811 N * cuba33ci_ cuba33ci 1319664334 J * FireEgl FireEgl@2001:470:e056:1:dc5a:d00b:18c8:385d 1319664763 Q * FireEgl Remote host closed the connection 1319665503 Q * fisted Ping timeout: 480 seconds 1319665555 J * fisted ~fisted@xdsl-87-78-211-188.netcologne.de 1319665793 Q * black Read error: No route to host 1319666733 Q * treaki Ping timeout: 480 seconds 1319667520 J * treaki ~treaki@91.3.53.233 1319667624 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1319667812 M * ser i have not still resolved this magic IO issues with htop, htop sources are here: http://htop.svn.sourceforge.net/viewvc/htop/trunk/ 1319667831 M * ser maybe you can have a second to have a look :) 1319667887 M * ser the only process which IO is displayed, is htop itself :) 1319667913 M * ser of course in "chcontext --xid 1 -- htop" 1319668022 Q * bonbons Quit: Leaving 1319668697 M * Bertl we are talking about IO_*RATE? 1319668748 M * ser Rate and IO_RBYTES (WBYTES) as well 1319668758 M * ser imho it is the same source of info 1319669100 M * ser it is extremely important to have a such overview, as disk IO is usually the most essential possible bottleneck 1319669875 M * Bertl seems to be a side effect of the ptrace check, which looks a little weird to me anyway 1319669900 M * ser :-( 1319669910 M * Bertl daniel_hozac: still around? 1319670200 J * clopez ~clopez@238.10.117.91.dynamic.mundo-r.com 1319670307 M * Bertl ser: we'll look into that tomorrow 1319670319 M * ser Bertl: thanks a lot, mate! 1319670338 M * Bertl np, thanks for repeatedly reporting the issue :) 1319670372 M * ser i am on 3.0.7 now, and it works pretty well 1319670950 J * nicholi ~nicholi@rrcs-76-79-196-34.west.biz.rr.com 1319671185 M * Bertl glad to hear!