[00:17] frewsxcv: I was not aware of uwsgi at the time of configuration. Maybe I will look into changing that. I also did not know of nginx, but I am going to give it a try on aspare [00:17] nice xapian packages weren't compiled with python [00:17] awesome [00:17] *a spare arch server I have just set up. [01:30] i've got a ubuntu 10.10-server virtual machine running inside virtualbox 4.1 headles on a ubuntu 64bit 10.10-server host. it uses bridged networking. when i use ssh to do stuff on the 10.10 vm guest the terminal repsonse to typing is __reallllllyy__ slow, it uses 2gb of ram and has two cores. the server is an i7 3ghz with 8gb of ram. any ideas? [01:36] http://ubuntuforums.org/showthread.php?t=1752906 === smw_ is now known as smw [03:04] gahhh that did nothing [03:04] still slow as a muckraker [03:04] type a letter, wait 2secs, letter appears [03:06] airtonix: what diagnostics have you already tried? [03:09] not even sure what to do [03:10] my other two vm guests ( also ubuntu-server 10.10) don't have this delay response with ssh sessions i have running. i have all three running at the same time and all have ssh sessions going [03:12] Well, for example, have you looked for suspicious resource usage in top? Have you tcpdumped? Have you compared the /etc/ trees for different settings? [03:13] Just to be clear, you're sshing *from* a normal machine *to* the VM, right? [03:13] Your terminal emulator isn't running inside an emulated system [03:15] Just to be clear, you're sshing *from* a normal machine *to* the VM, right? <<<< yes [03:17] The particular failure case you're getting sounds to me like swap-thrashing or similar [03:18] It's not likely to be something like vda being slow because it's a qcow2 when the others are raw, because that would only have an effect after you hit the enter key and sent the command, whereas you're saying this is happening on each keypress [03:18] It could also be that the ssh line is flooded, hence tcpdump. [03:19] Could be something obscure like ip n caching isn't happening so every TCP packet needs an ARP handshake first. [03:29] hi [03:41] twb yeah i don't even know how to deal with any of that [03:48] how to download and burn all required packages on CD/DVD and reuse that time to time without internet [03:48] ? [03:50] linocisco, every package you install from the internet gets downloaded to /var/cache/apt/archives/ [03:51] unless you manually delete them or run apt-get clean, they will still be there [03:51] so if you have a system set up the way you want, just copy all the files from there [03:57] create_connection_server_info failed: NT_STATUS_ACCESS_DENIED [03:57] I'm getting that when trying to point Windows Backup at my home share. [03:57] oops, wait, fixed that, but now: [03:57] elitebook (192.168.1.143) connect to service dana initially as user dana (uid=1000, gid=1000) (pid 15214) [03:57] Nothing after that. [03:59] you have a samba share setup and have set everything appropriately? [04:00] I can write to files fine, but the "permissions" interface in Windows shows funky results (owner only has "special"). [04:01] It may very well be Windows failing, but I can't tell that. I can create files, but not "change permissions" via samba, as I assume the backups thing wants to do. [04:01] change the permissions on the samba share [04:03] yeah [04:03] samba creates the 'share' permissions that windows expects [04:04] and approximates NTFS security [04:05] I couldn't tell you off hand what option precisely you need [04:05] but there's a lot more of them than the stock configuration lets on [04:07] I try to change the permissions, I get "access denied". [04:09] How do I get more, but not horrendously much more, debug info? [04:21] Hmm, "access mask" bits are all unset on the server's SMB response. [04:21] er, "read attributes" and "sync" are set. [04:22] So, looks like it's a permissions issue. [04:22] Not quite "access denied", but instead, "you didn't give me the permissions I asked for." [04:23] er, wait... they're unset on the REQUEST, as well. [04:24] Create options: .... .... .... .... .1.. .... .... .... = Backup Intent: This is a create with BACKUP INTENT [04:26] twb: through experimentation i think i reducedthe problem but not eliminated it by sudo service stop a bunch of stuff including byobu (which was spiking now and then) [04:29] how to burn files in /var/cache/apt/archives from server ? [04:33] okay, so it's samba being funky. [04:37] okay, so it seems things owned dana:staff are erroring, but things owned dana:dana are not. [04:39] So, my permissions are funky. [04:43] Should I try Samba 4? [04:44] oh, and it's on ZFS, by the way. chacl not supported. [04:44] Homes are on zfs, that is. [04:44] Because btfs is usesless without a fsck that can actually fix. [04:47] To heck with it; I'm trying samba4. [04:52] GNUTLS failed to initialise - Error while reading file. [04:53] Unable to save certificate in /var/lib/samba/private/tls/cert.pem parent dir exists ? [04:53] That's a copy-and-paste. [04:58] okay, scratch samba4. [05:00] weird... I uninstalled 4 and reinstalled 3, and did smbpasswd, and now it works. Go figure. [05:00] airtonix: byobu is just a wrapper around screen [05:00] airtonix: it shouldn't be doing anything interesting === jussi01_ is now known as jussi [07:19] anybody knows why adding -net user,hostfwd directive makes vm use 100% cpu and make starts to choke whole server? [07:19] (kvm, 11.04) [07:43] lambda_x: that's for punching holes throught the userspace NAT? [07:43] Yeah [07:43] What was the full command line? [07:44] one sec [07:45] twb: /usr/bin/kvm -monitor unix:/var/run/kvm/dummy.socket,server,nowait -vnc 127.0.0.1:5908 -name dummy,process=dummy -rtc base=localtime -M pc-0.12 -balloon virtio -m 512 -drive file=/dev/vol0/dummy-boot,if=virtio,index=0,media=disk,boot=on,cache=none -drive file=/dev/vol0/dummy,if=virtio,index=1,media=disk,boot=off,cache=none -vga cirrus -S -daemonize -pidfile /var/run/kvm/dummy.pid -boot c -net nic,vlan=0,model=virtio,macaddr=fe:16:3e:75:89:cd -net tap,vlan=0 [07:46] Well, unless you are adding another -net nic, I guess it's confused as shit trying to both DNAT and TUN the one virtual NIC. [07:46] Er, s/TUN/TAP/ [07:47] well, so it is impossible? [07:47] You can't have both a userspace NAT and a tun on the same virtual NIC [07:47] to have tapped interface and redirect port? [07:47] s/tun/tap again [07:47] If you want to do port redirection of a TAP you do it in the firewall of the host OS [07:48] Because that's where the NAT, if any, will be [07:48] userspace NAT is in kvm/qemu for when you are running it as a non-root user and don't have permission to TAP [07:51] well, that is good idea... :) [07:52] tyvm twb [07:53] You could've worked if you if you stared at -net manpage entries for a few more hours [07:53] That's what I did [07:53] It is, sadly, not very obvious [07:55] true === kim0|holiday is now known as kim0 === smb` is now known as smb [08:16] I was hoping somebody might help me... I'm having issues with disk usage quotas.... [08:17] Alan: you forgot to mount it -ousrquota [08:17] no I didn't :P [08:17] the quota is working, it's just not correctly [08:18] so according to dumpe2fs, the block size is 4k [08:18] Wrong kind of block [08:18] (IIRC) [08:18] ... seriously? [08:19] what kind of block is a filesystem quota going to use if not a filesystem block? [08:19] http://paste.debian.net/128041/ [08:19] That's a working hard quota of 1GiB per user on /home, on ext4 on lucid [08:19] (The inode limits are currently way too high; ignore them.) [08:19] ok, well thanks for that [08:20] but I totally want to slap whoever decided that "block" was the correct terminology there [08:20] It's block limit as in blocks vs. inodes [08:21] Daviey: want me to pickup the iscsi tests? still have my test rig intact [08:21] when you say "block limits" in relation to a filesystem, how is it supposed to mean anything other than filesystem blocks? [08:21] twb: i understand the difference, but it's still ambiguous terminology [08:21] i.e. it implies the wrong units [08:21] Alan: the way you should have worked this out, is to set it to a number like 1k or 1M, then dd /dev/zero into a file and see where it chokes [08:21] especially since the inode limit is the number of inodes [08:21] the block limit is intuitively the number of blocks... [08:21] FWIW not all filesystems are block-oriented in the first place [08:22] twb: i thought i had done that, effectively... [08:22] Alan: maybe your test was bogus because quotaon wasn't on? [08:23] it was on [08:23] it was an unscientific test though [08:24] Now you know better [08:24] I set the quota to 1024, and started SFTPing stuff until it errored [08:24] which it did [08:24] I didn't actually check exactly how much space i'd used though... [08:26] so... are the M/G suffixes valid when specifying limits? [08:26] Probably not [08:26] or just there for convenience in the output? [08:26] damn :( [08:26] quotatools is from the dark ages [08:28] so... those are 1k blocks then? [08:31] I think they are 1b blocks === alexy_ is now known as alexy [10:11] hi [10:11] i have only P3 with 2*128MB SDRAM, can I install ubuntu server on that Compaq Ex815 computer ? [10:11] linocisco: I think 10.04 LTS should install on that. I have installed it on machines about that age and capabiity. But the best way to find out is to try it :) [10:12] jmarsden: thanks man [10:13] linocisco: You're welcome. have fun :) [10:13] jmarsden: that is obsolete in office . so i got it as present [10:14] Yes, it's not much use in a normal office environment as a workstation any more :) [10:14] 3am here, I need to sleep ... goodnight. [10:14] jmarsden: I am trying to make the most out of it. with pure CLI server. I want to install mail server on it [10:15] jmarsden: ok. good nite [10:29] how can I disable the xorg server in a livecd to not start at boot time? [10:30] trying to create a custom livecd [10:36] Daviey: just looking at the bug queue - does bug 835625 need to be noted for beta-1 [10:36] Launchpad bug 835625 in cyrus-sasl2 "package libsasl2-2 2.1.24~rc1.dfsg1+cvs2011-05-23-4ubuntu2 failed to install/upgrade: libsasl2-2:i386 2.1.24~rc1.dfsg1+cvs2011-05-23-4ubuntu2 (Multi-Arch" [Undecided,New] https://launchpad.net/bugs/835625 [10:36] ? [11:02] I installed a virtual machine with ubuntu 11.04 server. I can not get the 70-persistent-net.rules to be generated. What am I doing wrong? [11:19] join #ubuntu-server-ru [11:22] Hi people, i have some problem with UEC. How i can make failover cloud controller? [11:35] blinkiz: Check http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=562211 if it applies. [11:35] Debian bug 562211 in udev "udev: 70-persistent-net.rules is not generated" [Important,Open] [11:53] smb: so those two patches from xen hg will fix your hypervisor bug? [11:54] zul, Yes, first one fixes the disk, second one the NIC [11:54] smb: cool ill get it in after the beta [11:55] zul, Excellent. With those we can at least boot kernels before Oneiric. [11:55] in hvm mode [11:57] hi [11:57] how to check if my desired domain name is taken or free? [11:57] morning btw [12:03] I am setting up ubunt mail server with dpkg-reconfigure postfix [12:03] please help in choosing some options [12:05] should I choose Procmail for local delievery ? Yes or No ??? [12:06] hi [12:11] hi all [12:11] i got error in postfix [12:11] what do I do ? [12:11] Posting the error would be a good start. Use a pasetbin though. [12:11] telnet mail.example.org 25 = could not resolve .xxxxxx.org/25:Temporary failure in name resolution [12:11] what do I do ? [12:13] thats not really a postfix error, it sounds like either your DNS is malfunctioning or there is an issue with your domain registration. [12:15] Pici: I have no config on network cards and no DNS server. that server is now standalone without internet [12:15] Pici: how can I make it work? [12:15] Pici: giving IP address can solve ? [12:16] linocisco: is your server's hostname '.xxxxxx.org' or whatever? [12:16] Pici: yes. let say example.org [12:16] Pici: how can I check host name ? [12:16] linocisco: the command is 'hostname' [12:17] yes. [12:17] Pici: yes. I see [12:17] Pici: it is hostname=ubuntuserver [12:17] linocisco: so use that name instead of '.xxxxxx.org' [12:17] zul, Daviey, RoAkSoAx: any of you guys have a bit of time to sponsor a few jenkins related merge proposals for me? [12:18] how many beers are we talking about? ;) [12:18] Pici: I got it [12:19] Pici: i got telnet = fine. ehlo xxxx.org = fine [12:20] Pici: btw, can I install dovecot from ubuntu server CD ? [12:20] :| [12:20] zul: beer (>=4) | beer (< 5) [12:21] :-) [12:21] In about two hours, we will be beering at the office \o/ [12:21] jamespage: oh hell ok....give them here [12:21] Thanks for remining me :-) [12:21] zul: ta [12:26] jamespage: oh joy, multiarch fallout. [12:26] Daviey: well maybe [12:27] the last change was a multiarch change [12:27] but I can't reproduce ATM [12:27] :/ [12:29] hi [12:29] how can I add apt proxy on 10.04 LTS server? [12:30] i found no apt.conf file under /etc/apt [12:34] hi [12:36] New bug: #838768 in nova (universe) "Should default to --use_deprecated_auth" [Undecided,New] https://launchpad.net/bugs/838768 [12:37] ttx: ^^ seen that? [12:37] i found no apt.conf file under /etc/apt [12:37] i found no apt.conf file under /etc/apt [12:38] Daviey: ack [12:38] linocisco: Just add the CD into /etc/apt/sources.list [12:39] ersi: uncommenting ?? [12:40] Yes. And if it's wrong, correcting it. [12:42] ersi: why would I need CD while my server is connected to internet ? [12:42] I'm sorry, I can't help you. I'm about to lose it when I'm trying [12:42] Hmm.. i had better get the TechnicalOverview for Beta 1 done, and known bugs. Anything jumping out at people for stuff to inlclude. [12:42] ersi: lose what? [12:43] RoAkSoAx: ensemble api work, all works now? Is that something i can add? [12:44] Control of my mind, and my ability to be helpful. I'm gonna go do something else now. Good luck. [12:44] ersi: ok [12:45] hi all. I lost apt.conf file under /etc/apt [12:45] what do I do ?? [12:50] hi all. I lost apt.conf file under /etc/apt [12:50] shocking [12:55] hi all please help me === lenios__ is now known as lenios_ [12:57] hi all [12:58] How can you lose a file? [12:58] soren: I dont know . I found apt.conf.d folder instead of apt.conf file [12:59] soren: because I need to point correct apt proxy to download files with apt-get install command [12:59] soren: actually I didn't configure apt and network during installation [12:59] There is no apt.conf by default. [12:59] Just create it. [13:00] soren: that's all ? [13:01] Well... [13:01] You have to put your configuration in there. [13:01] Or put it in a file in apt.conf.d. [13:03] soren: Acquire::http::Proxy myproxy.net:8080 ?? [13:03] Have you tried? [13:03] Try before you ask. [13:04] soren: thank alot. I put outside of apt.conf.d folder. it is working now [13:05] soren: you are long term ubuntu server geek. I saw your name in mailing list for such a long time. [13:05] linocisco: Yes. I'm old. [13:05] It's true. [13:06] Daviey: we both forgot. I can't create an ubuntu-virt ppa. [13:06] I can. [13:06] soren: you are decreped [13:06] soren: I am now updating with apt-get update. then I will try to install (which package) to work with microsoft outlook [13:06] zul: I'm almost sure that's not a word. [13:06] soren: some said. I need to install dovecot-common [13:07] No idea. I live in the land of fairies and unicorns where there are no Microsoft clients. [13:07] soren: you are right its spelled decrepit :) [13:08] hallyn: Do you want me to? [13:08] hallyn: Wait, hang on. [13:08] * hallyn hangs on [13:09] hallyn: There. You can now. [13:09] hallyn: Have fun. [13:10] soren: May i have that foo aswell please ? :) [13:10] soren: do you know what packages we need to install to work with microsoft outlook ? I am setting up mail server to be used by microsoft outlook client [13:10] Daviey: I may be bribed into bestowing such foo upon you, yes. [13:11] linocisco: Nope, no idea. [13:11] soren: that is real difficult thing I even can' t find on youtube [13:11] On... youtube? [13:11] soren: I'll let you buy me a drink at ODS> [13:12] Daviey: You know how to make me feel special. [13:12] Reverse psychology at work right there. [13:12] soren: There is $50 in my drawer, feel free to take that on your way out. [13:13] Um. [13:13] No. [13:13] Just.. no :) [13:14] soren: thanks :) [13:15] Certainly. [13:15] if I'm on an Ubuntu LTS running as a server, no network manager, should I have my DNS servers within the interface file AS WELL AS etc resolv? Or just etc resolv? [13:15] Because I'm having some DNS issues on my ubuntu thin client LTSP server and I can't trace if it's network config or the server itself. [13:15] Daviey: so do you still think i should call it qemu-0.15, or should i make a 'upstream-candidate' ppa for more general use? [13:15] i fjeer to have ubuntu-virt end up like my acct with too many undeletable ppas... [13:15] roasted: Using DHCP? [13:16] if you call it upstream-candidate then you can reuse it for other things later [13:16] soren, DHCP is handled by windows. DHCP is disabled on the server. The server has a static IP for both interfaces. [13:16] roasted: Then put it in resolv.conf [13:16] soren, it is... :( [13:16] hallyn: upstream-candidate sounds ok to me [13:16] soren, I really can't figure this one out. [13:16] fishy-software-is-it-crack also sounds OK [13:16] soren, now all of the sudden its working. what the... [13:16] roasted: Whatever is in resolv.conf is what counts. [13:18] soren, It's strange. The clients themselves cannot ping by host name, however my laptop (plugged into a line that feeds into the Ubuntu server) can ping by hostname fine. [13:18] Er.. [13:18] BAck up a bit. [13:18] What are you trying to do, exactly? This doesn't sound like DNS at all. [13:21] soren, basically I have a thin client server here. IT's 10.04 64 bit. It has 2 network ports. each one is subnetted into a different IP range from the main network. [13:21] soren, both switches are Dell 3548's. One network port to one switch. Then the switch connects to 30 systems in the lab. [13:21] NIC 1 - 3548 - 30 thin clients that PXE boot to it. 10.52.18.X network on triple 255 subnet. [13:22] NIC 2 - 3548 - 30 thin clients that PXE boot to it. 10.52.17.X network on triple 255 subnet. [13:22] The main network is 10.52.X.X (third octet being whatever building it is) on 255.255.252.0. [13:22] I'm just kind of confused on where it could be. [13:23] The LTSP guys are also telling me I could put a DNS entry in the LTS.conf file, which controls how LTSP operates. [13:23] one of the LTSP devs told me to do that, and then said: I have found that in some of the versions, we don't have the initramfs code to update resolv.conf on the client from dhcp [13:23] perhaps I should try it quick [13:24] You're confused where *what* could be? [13:28] soren, the issue. [13:28] soren, sec, Im trying the devs suggested fix. [13:29] And I'm asking what the issue is. [13:30] yeah, no dice. [13:30] on the clients, I cant ping by hostname. [13:30] however, I can plug in my laptop to a port the client was plugged into, and I can ping by hostname then [13:31] It sounds like there may be an issue with your dhcp server. [13:31] You can't ping what by hostname? [13:31] Can you ping google? [13:32] nm, that line of questioning is probably better. [13:37] oh wow [13:37] I can ping google. I can't ping our file server. [13:40] There's your problem [13:46] Daviey: ensemble api? [13:47] hallyn: Does this look familiar? error: Unable to create cgroup for instance-00000002: No such file or directory [13:48] looks like this may be it [13:48] http://ubuntuforums.org/showthread.php?t=1626167 [13:48] soren: yes [13:48] soren: stop libvirt-bin; start libbvirt-bin [13:48] hallyn: What gives? [13:48] well lemme know if that fixes it first :) [13:48] It did. [13:48] Thanks :) [13:48] bingo. fixed. [13:49] Daviey: all the orchestra stuff is in ensemble trunk now [13:49] soren: bug 828061 [13:49] Launchpad bug 828061 in libcgroup "cgroup-bin prevents libvirt from starting" [Critical,In progress] https://launchpad.net/bugs/828061 [13:49] soren: best workaround is swap cgroup-lite for cgroup-bin [13:49] hallyn: Otherwise I'll have to restart libvirt-bin every time I.... what? [13:49] I need to insert a different barrier between cggconfig and libvirt upstart jobs [13:49] every time you boot [13:49] Oh. [13:50] libvirt-bin just needs to start a bit later [13:51] It's because cgroups-bin reclassifies the existing libvirtd or something? [13:51] right, [13:51] Hmm. [13:51] and i think most often, cgconfig gets the list of pids, [13:51] then libvirt spawns off 8 threads, [13:52] then cgconfig reclassifies the pids it has [13:52] whcih do not include the newly spawned threads [13:52] so really while cgconfig starts, the whole system should be frozen [13:52] the libcgroup stuff is going to cease to be developed, with functionality going into systemd, [13:53] and i think that's the right move. [13:53] Oh. [13:53] the run-time daemon thing is inherently racy. there are unsolvable races requiring a different kernel-user api [13:53] Do you have a plan for fixing this? [13:54] cgroup-lite :) [13:54] but no, i could stand to have a chat for ideas if you're up for it [13:55] could help to pull in speakman [13:55] uh, SpamapS i meant [13:55] soren: to be more precise, the libvirt one i can fix, but the general case is problematic [13:56] hm, maybe the libcgroup-libvirt-waiter.conf should turn into a more generic libcgroup-pause-all-daemons-using-libcgroup.conf [13:58] :p [13:59] speakman: apologies :) [13:59] hallyn: no worries :D [14:03] hallyn: That was what I was going to suggest :) [14:05] hi guys, I have a new acct that is not getting any emails. How to proper set that acct inbox? using postfix here [14:06] hi utlemming, Daviey has asked me to sort out the qemu arm image - I understand you built the binary image for it? [14:06] rbasak: yup [14:07] I can't get it to run :-) [14:07] soren: a more generic waiter job? I'm not sure of the best way to get that done though [14:07] rbasak: well, your my new best friend [14:07] Daviey wasn't sure if it had ever worked [14:07] lol [14:07] I'd like to start by reproducing your build [14:07] rbasak: the image works -- I used qemu-system-arm to get the image created [14:08] rbasak: the problem is the kernel [14:08] lol [14:08] well that's what I was trying to get to run :) [14:08] I'd said to Daviey before that the userspace will probably all work and the issue is the kernel :-P [14:08] rbasak: initially we were told to use the linario-vexpress kernel, but that kernel is not compiled by Ubuntu [14:09] so we got told to use the OMAP kernel. The OMAP kernel requires a bootloader (u-boot) I believe to be installed on the image [14:09] soren: i'll send an email to ubuntu-server with the problem and what we're dong so far [14:09] I think we need a versatileab kernel [14:09] qemu-system-arm doesn't appear to support OMAP [14:10] rbasak: :) [14:10] rbasak: hang out in #ubuntu-arm and they'll tell you otherwise :) [14:10] what your hitting is that the OMAP and Beagle emulation requires a special boot loader installed on the image [14:11] But qemu refuses to start without a kernel image [14:11] (in ARM mode) [14:12] Oh, hang on [14:12] Ah, I should be using -M beagle or something? [14:12] yes [14:12] then the images need to have u-boot or some sort of firmware installed [14:13] OK I follow [14:13] is there a spec for what the bootloader built into -M beagle is doing? [14:14] http://elinux.org/BeagleBoard [14:16] utlemming: Oh? I thought qemu-system-arm could exec the kernel directly? [14:23] Daviey: yeah, it can, but the OMAP kernel says, "no boot device found" [14:25] hello, am I seeing this right? Ubuntu 10LTS will install solr version 1.4? [14:25] robos: solr | 1.4.0+ds1-1ubuntu1 | lucid/universe | source [14:25] robos: Yes. [14:25] Seems like it. [14:25] yep. [14:26] yikes. Even 11.04 is 1.4.1. The latest version out is 3.1 [14:28] utlemming: I'm pretty sure that message is from the qemu's beagle bootloader emulator rather than the kernel [14:29] rbasak: interesting === himcesjf1 is now known as himcesjf [14:30] utlemming: going the beagle emulation route seems a bit obtuse for what we need [14:31] utlemming: http://code.google.com/p/qemu-omap3/wiki/UserManual [14:31] (have to generate a separate nand image maybe, etc) [14:32] utlemming: what do you think of using a different machine type (and thus kernel) to save having to create an additional intermediate image? [14:32] rbasak: "bit obtuse" would be overly generous [14:32] utlemming: (given that qemu already supports not having to go down the bootloader route for arm in versatibleab) [14:33] I mean we're not trying to emulate beagle here, qemu is our target [14:33] rbasak: I'm absolutely fine with that -- but AFAIK, we don't have a kernel with the meta package that will work on versatibleab [14:33] OK, I don't really know what would be involved in adding a different kernel [14:34] stgraber: ping for pushing libcgroup 0.37.1-1ubuntu6 for bug 838380 ? [14:34] Launchpad bug 838380 in libcgroup "non-root should not be able to create files in /sys/fs/cgroup" [High,In progress] https://launchpad.net/bugs/838380 [14:34] hallyn: archive is still frozen for beta1 [14:35] robos: solr is well behind at the moment - however work is going on to package lucene3 + solr3 (now part of the same source package) in Debian which should land next release (12.04) [14:35] stgraber: oh, i thought i heard a rumor that it was unfreezing today :) [14:35] drat [14:35] stgraber: are you on ubuntu-server m-l? [14:36] zul: do i understand right that there is nothing you need from the candidate libvirt 0.9.3-5 merge? you just need an arm patch that can just as well go into the current package? [14:37] jamespage, when do you think the next LTS version will be out? [14:37] hallyn: it can go in the current package i just have to rediff it [14:37] Well, you probably can't answer that. :-) [14:37] But... I wonder if compiling solr and making it an ubuntu package will do the trick [14:38] zul: ok, we have enough crap to fix without the merge, i'm thinking of marking the FFE bug invalid [14:38] hallyn: ok cool with me [14:38] for that matter, Daviey was right to ask if there was anything in qemu 0.15 that we needed. I'm not convinced. [14:38] ill get on it after what im done here [14:38] great, thx. i'll roll it in with the other two fixes i'm waiting to push [14:39] which fixes? [14:39] Daviey: pls do put out the call for testing for qemu 0.15 like you were going to, and we'll see if anyone says "i need this" [14:39] zul: for bugs 828211 and 832123 [14:39] Launchpad bug 828211 in libvirt "libvirt lxc should not rely on 'ip' and 'ifconfig' inside container" [Medium,In progress] https://launchpad.net/bugs/828211 [14:39] Launchpad bug 832123 in lxc "when booting oneiric cloud-images under libvirt, no pty is available" [High,In progress] https://launchpad.net/bugs/832123 [14:39] hallyn: cool beans [14:40] hallyn: yeah, it should unfreeze once beta1 is out a bit later today. I'm not subscribed to ubuntu-server [14:41] rbasak: hang on -- it looks like Daviey might have you put on a wild goose chase [14:42] rbasak: I'm now seeing a linux-image-versatile (https://launchpad.net/ubuntu/+source/linux-meta) [14:42] stgraber: ok, i wanted to have a conversation about fixing the race with libcgroup startup. i'll cc: you? [14:43] robos: should be april 2012 [14:43] utlemming: OK, so can we build the image using that? [14:43] In fact to test I can just grab the binary image [14:43] hallyn: yep, just CC me [14:43] great, thanks [14:43] april 2012 for LTS? sweet! [14:44] robos: the packages should land in Debian in the next month or so - its just not good timing for oneiric as that would be very later in the cycle [14:44] LOL! of course there *is* a simpler solution to the libcgroup startup. Make '/' the default cgroup. [14:45] utlemming: OK, so can we build the image using that? [14:45] In fact to test I can just grab the binary image [14:46] It doesn't seem to be in the latest build :-/ [14:47] robos: oh - and solr skipped version 2.x altogther so its not that out-of-date :-) [14:49] hey guys i am having issues after following the postfix and dovecot wiki's to setting up the servers. now my issue it seems like i am having issues authenticating with the imap server [14:50] rbasak: humm....where are you looking for the kernel? [14:50] can anyone help me with debugging the issue as I am getting increasingly frustrated [14:50] utlemming: same page, https://launchpad.net/ubuntu/+source/linux-meta [14:53] rbasak: your right [14:53] any ideas as to my issue ? [14:57] any dovecot imap authentication experts in here cuz i am at a total lost [15:10] anyone know if Daviey's around? [15:15] skaet: he was... [15:16] hallyn, thanks. [15:17] hey guys question [15:17] what should my permissions be set to for squirrelmail [15:17] both dovecot's and postfix's websites explain how to do it correctly [15:17] permissions? doesn't matter at all [15:17] patdk-lap: im following ubuntu squirrelmail documentation [15:17] well im having issues with it loading right [15:18] oh, the website won't even load? [15:18] ERROR: Could not complete request. Query: SELECT "INBOX" Reason Given: [SERVERBUG] Internal error occurred. Refer to server log for more information. [2011-09-01 17:16:39] [15:18] and [15:18] ERROR: Could not complete request. Query: LSUB "" "*" Reason Given: [NOPERM] No permission to read subscriptions [15:19] That sounds like an IMAP error, not a squirrelmail error [15:19] cloakable: O_o [15:19] how so [15:19] what? [15:20] how is that an imap error [15:20] dovecor isn't giving squirrelmail permission to check folders [15:20] *dovecot [15:20] and how do i fix that [15:20] Have a look at the permissions on your mailfolders. [15:21] the permission is fine [15:21] Check your dovecot configuration to see if it's looking in the right place. Check the error log [15:21] try logging in using telnet [15:22] logging in via telnet works [15:23] hmmm. [15:31] hallyn: archive unfrozen. Looking at that upload now. [15:31] stgraber: great, thx. [15:33] hallyn: uploaded [15:37] cloakable: [15:37] problem solved [15:38] need to update the dovecot wiki [15:38] eagles0513875: Go ahead and update it, wiki are editable :) [15:38] ya i will [15:38] Ok, good. [15:39] already updated the dovecot wiki with an issue i discovered with encrypted home directories [15:40] now the problem im having it doesnt seem like my emails are coming through to my server [15:50] hello, i have 2 HDs (sda and sdb) and sda is not listed on blkid [15:50] how can i format it to ext4 and mount? [15:53] smoser, Hi, why did you set bug 1 for test case http://iso.qa.ubuntu.com/qatracker/result/6432/519 ? Is there a real bug or an infrastructure failure or ? [15:53] Launchpad bug 1 in ubuntu "Microsoft has a majority market share" [Critical,In progress] https://launchpad.net/bugs/1 [15:58] jibel, i updtaed it. i there were 2 boot failures seen there, i dont have a bug for the second, but i changed it to the first. [15:59] smoser, cool. thanks [16:01] but i do find it annoying that i can't log a failure without a bug [16:13] RoAkSoAx: Have you been able to document ensemble for bare metal? [16:13] Daviey: i've notes here and there but nothing definitive just yet [16:14] Daviey: wanted to actally have things working on the test righ though I'm just waiting for a fix from cloud-init to get that working again [16:14] Daviey: ut i will start writing a wikipage soon [16:14] Daviey: this weekend/next week [16:27] SpamapS, around ? [16:28] New bug: #825127 in python-novaclient (main) "[MIR] python-novaclient" [Undecided,Incomplete] https://launchpad.net/bugs/825127 [16:31] RoAkSoAx: please tell me when you have it -- so that I can build a test env Foolowing The Standard [16:31] hggdh: coolio, will do [16:32] hggdh: I promise it will be ready for next week ;) [16:32] RoAkSoAx: heh. No need to promise ;-) [16:32] hggdh: hehe, better to promise that not do so :P [16:33] Say, my tests over gigabit ethernet have given me only 40 megabytes per second on file transfers... is that normal? [16:33] RoAkSoAx: Okay, thanks [16:33] One end: Ubuntu 11.04, on HP Microserver with 5GB RAM and two Samsung F4 drives with ZFS mirroring. [16:34] The other end: HP EliteBook 8530w -- Intel 82567 ethernet, ICH9M-E, and an OCZ Vertex. [16:36] New bug: #838931 in backuppc (main) "package backuppc 3.1.0-9ubuntu1 failed to install/upgrade: le sous-processus script post-installation installé a retourné une erreur de sortie d'état 1" [Undecided,New] https://launchpad.net/bugs/838931 [16:45] danag, and what is the test? [16:53] ah, the test case can be either simple file copy (samba), or Windows 7 backup (also samba). [16:53] those are horrible tests [16:53] 40MB/sec sounds right then [16:53] ah. [16:53] That may just be the answer, then. [16:53] using like ftp, around 60MB/s [16:54] What's most likely the bottleneck: network bandwidth, samba, or drive? [16:54] who knows wit hthe laptop [16:54] I've benchmarked file copy from SSD to SSD, through same switch, at like 80 megabytes per second. [16:54] but tuning tcp windows makes a huge difference at gigabit speeds [16:55] I've benchmarked 110MB/sec over gigabit on my systems [16:56] By TCP window do you mean jumbo frames? [16:59] The Broadcom 5723 in the microserver doesn't do jumbo frames. :( [17:03] DanaG: tcp window has nothing to do with jumbo frames [17:03] ah, recieve window size. As in sliding window algorithm? [17:04] I've taken networks class, but forgot the precise name of the algorithm. [17:04] yeah, for tcp. doesn't matter much for low-latency transfers though [17:04] I'd test pure network performance and see if it is there, using iperf [17:09] DanaG: how much ram is in the laptop. as well what kinda hdd is in said laptop [17:10] for low latency I find larger windows hurt performance [17:10] and jumbo frames will only cause more issues [17:10] The 80 test was from EliteBook 8530w to a Fujitsu Pentium M system. [17:10] why would jumbo frames cause more issues? [17:10] jumbo frames works extreemly well, and no issues [17:10] DanaG: i know what your talkin bout i just took a compression class and studied that this past year [17:10] patdk-lap: wont you have massive amounts of fragmentation [17:11] why? [17:11] 1500 -> 9000? why would increasing it cause fragmentation? [17:11] EliteBook had 4 gigs of RAM and an Intel 82567, with OCZ Vertex 2. [17:11] why do you have fragmentation at all i nthe first place? [17:11] patdk-lap: woudl depend on the route taken not all routes to the destination support jumbo frames [17:11] eagles0513875, like we said *DIRECT* [17:11] Target was a Pentium M + ICH6 AHCI + OCZ Vertex 1, with Broadcom NIC of some sort. [17:12] Now the Vertex 1 is back in the EliteBook, since the Vertex 2 died. Twice. [17:12] and you still wouldn't have fragmentation due to pmtu over a long multihop path [17:12] DanaG: i think the 2 biggest issues would be the speed of the nic's as well as the rpm's of the disk drives [17:12] if they are SSD then the bottle neck woudl be the NIC [17:13] not all ssd's can substain >100mbit [17:13] and rpm's don't matter for the disk [17:13] doing sequential read of the drive will give you it's max [17:13] normally between 80-120MB [17:14] Direct-attached benchmark of the Samsung was at least 80, I believe. [17:14] I'll bet ZFS is part of the bottleneck. [17:14] so you will never go >80MB/s [17:14] zfs will cause some latency, slowing things down [17:14] also depends on how everything is attached [17:14] does the nic and drives have their own pcie lanes? [17:14] do they share a southbridge? .... [17:15] if they are all connected over pci, then max would be 50MB/sec [17:15] half for drives, half for nic [17:15] same goes for your laptop [17:16] you could always try doing a transfer from /dev/zero to /dev/null [17:16] if you get gigabit [17:16] then it's a harddrive + nic sharing pci most likely [17:17] It's on PCIe. Not even one PCI legacy slot in the thing. [17:18] hardly anything has pci any more [17:18] utlemming: I've got your image booting with a different kernel image now, but it seems to require an initrd? [17:18] which kernel image are you using? [17:18] utlemming: fedora's :-) [17:18] (for the moment) [17:18] The southbridge is an SB850, I think. [17:18] lol [17:18] Server kernel. [17:19] http://paste.flashrom.org/view.php?id=578 [17:19] lspci. somebody else's paste. [17:19] the initrd is going to be awkward if we're going to boot it like this [17:19] oh that's interesting, qemu does support an initrd [17:20] rbasak: there is a linario vexpress kernel that works, but networking is flaky [17:21] * utlemming looks for URL for linario kernel [17:21] utlemming: what are you doing with the initrd? I tried the one in your image but of course that's mismatched with the kernel now and it doesn't seem to work [17:22] rbasak: nothing, I would fully expect the in-image initrd to completely fail [17:22] it was generated in qemu-arm-static (i.e. chroot) [17:26] New bug: #838959 in rabbitmq-server (main) "rabbitmq-server startup_log should include a date/timestamp" [Undecided,New] https://launchpad.net/bugs/838959 [17:27] rbasak: I just found something that might help -- from linaro https://wiki.linaro.org/Platform/DevPlatform/Ubuntu/ImageInstallation [17:27] rbasak: the problem with this is we _can't_ ship any of this in the default cloud images. We're can only ship stuff from main or restricted [17:27] Once btrfs gets a fixing fsck, I'll probably switch. [17:29] utlemming: their linaro-media-create looks like it'll be able to help, then we could use our omap image [17:30] rbasak: that is the road that I just started barking down [17:30] utlemming: I was reluctant because it's ugly, more work and harder to debug, but if they've already done it...! [17:31] rbasak: agreed, its the reason why I didn't do it as well. I'm more likely to try and implement a toolchain that has sane instructions and good documentation [17:32] utlemming: I need to go. Do you want to see how far you get or would you like me to look at it further tomorrow morning? [17:33] rbasak: I'll beat my head against this for a bit. I generally am on around 14:00 UTC, we can sync up then if that works for you [17:33] utlemming: OK, speak to you tomorrow [17:52] any mail experts in here [17:52] need some urgent help [17:58] mount.nfs keeps the uid/gid as it is shared by nfs server.. but mount.nfs4 adds some odd uid/gid.. like below [17:58] http://pastebin.com/AZ4XGhqi [17:58] line 14 (nfs) and line 29 (nfs4) [17:58] any idea why? [17:59] mount.nfs works.. but mount.nfs4 b0rkes it [18:02] axisys: see https://help.ubuntu.com/community/SettingUpNFSHowTo, and the section on mapping ID names with /etc/idmapd.conf [18:02] anybody know what version of ubuntu started defaulting NFS to version 4 instead of version 3? [18:04] RoAkSoAx, adam_g bug 838968 is what is stopping cloud-init and apt from working sanely on mabolo [18:04] Launchpad bug 838968 in ifupdown "static-network-up event does not wait for interfaces to have an address" [Undecided,New] https://launchpad.net/bugs/838968 [18:06] jj995: looks like what you are showing is for nfs server.. [18:06] jj995: I am trying to do nfs client [18:07] as you can see on line 14.. nfs keeps the uid/gid sane.. unlike in line 29 it does not with nfs4 [18:08] axisys: from the link I gave, if the client is mounting nfs4, "In order for the ID names to be automatically mapped, both the client and server require the /etc/idmapd.conf file to have the same contents with the correct domain names." [18:09] jj995: oh ok [18:09] How do I format a drive into XFS? [18:19] cjwatson: Ping. Do you remember the name of the IPv6 test suite you mentioned at UDS? === lullabud is now known as warzauwynn [18:47] jj995: thanks, that was it! [18:54] Hello . . . I have been trying for a few days to connect with a guest operating system ( Ubuntu Server Edition ) from the host operating system ( Ubuntu Desktop Edition ) via SSH. I have tried port forwarding, changing the guest network device, and taking suggestions from people on the channels #VBox, and #OpenSSH. Nothing I have done was successful. How do I establish this connection? [18:56] internal or external network? [18:56] have you edited the sshd files on the server at all? [18:57] Did you installl openssh-server on the server? [18:57] and of course, you've installed openssh-server on the guest, since it's not installed by default [18:58] https://help.ubuntu.com/10.04/serverguide/C/openssh-server.html [19:02] well, ugly as all sin, but https://gist.github.com/1186970 would help us wait for IPV4 to comeup (bug 838968) [19:02] Launchpad bug 838968 in ifupdown "static-network-up event does not wait for interfaces to have an address" [Undecided,New] https://launchpad.net/bugs/838968 [19:03] SpamapS, ^ [19:04] lickalott, ScottK, ppetraki: I have tried Host-only adapter, Internal Network, Bridged Adapter, and have it currently set to NAT. I have not edited the sshd files, openssh-server is installed. [19:09] can you ping the box? [19:10] How do I do that? [19:10] Ping the ip address or ping the name of the server? [19:11] smoser: interesting [19:12] smoser: I wonder if there isn't some netlink socket that we can listen to for these things, a-la network manager... IIRC, n-m doesn't poll.. [19:14] stgraber: sorry to bother you again, but can't think of anyone else to ping on this - if you get a chance could you look at the proposed debdiff for bug 838410 ? [19:14] Launchpad bug 838410 in lxc "sshd template is broken" [Medium,Triaged] https://launchpad.net/bugs/838410 [19:14] yeah, there has to be some way. we're interested in network interface changes. [19:14] zul: will you be able to get that patch for libvirt to me today? [19:14] hallyn: ill do my best [19:15] do you have a bzr tree [19:15] either should suffice [19:15] zul: i don't, but debdiff is at people.canonical.com/~serge/libvirt-0.9.2-4ubuntu12.debdiff [19:15] (i'd toss up a bzr tree, but since it's so out of sync i don't have the bandwidth right now, don't ask) [19:15] if you're Nat'd correctly you should "hop" all the way through to the server. if you can't ping it successfully then i would look at your routes and/or NAT statements then re-engage [19:16] hallyn: looking [19:16] stgraber: thx. noone really cares (obviously) about sshd template, but it's something ppl often try for fun... [19:19] hallyn: looks good, uploaded [19:20] oh, one note actually :) [19:20] hallyn: -- Ubuntu Thu, 01 Sep 2011 16:01:31 +0000 [19:20] doh [19:20] is that in quilt file? [19:20] nope, debian/changelog [19:21] replaced with your @ubuntu.com [19:22] stgraber: great, thanks [19:22] i'll send the template patch upstream of course [19:22] good thing lintian detected it ;) I didn't see it by reading through the diff [19:22] at some point... [19:23] hm, guess i only built it with fakeroot debian/rules binary [19:23] i need to get my sbuild box up [19:29] smb: xen fix has been uploaded [19:31] hallyn: people.canonical.com/~chucks/libvirt-0.9.2-4ubuntu12.debdiff [19:31] zul: thx, i'll test a bit and push [19:32] hallyn: ok ask if you need someone to upload [19:32] its already signed locally ;) [19:43] Im trying to run mysql server using a different datadirectory [19:43] and its not liking it [19:43] when I try to start the service it just hangs [19:45] fuckin apparmor isnt it [19:46] zul: gah, bad build. a few more minutes, and i'll ask you to push your built version :) [19:46] (bad build = my fault, not the package) [19:51] zul: success - pls push :) [19:52] now on to this libcgroup crap. [20:05] hallyn: uploaded [20:07] zul: rockin'! thx. === aliverius is now known as anthropos === anthropos is now known as aliverius [20:37] how to downgrade a minor version of python2.7? [20:45] nm - downgraded the 3 files from apt-cache [20:47] ChmEarl i would downgrade the file from the apt-cache dir [20:47] :p [20:48] lickalott, first time I ever did this [20:51] anyone know about an httpd update for https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2011-3192 ? [20:51] astrostl: The byterange filter in the Apache HTTP Server 1.3.x, 2.0.x through 2.0.64, and 2.2.x through 2.2.19 allows remote attackers to cause a denial of service (memory and CPU consumption) via a Range header that expresses multiple overlapping ranges, as exploited in the wild in August 2011, a different vulnerability than CVE-2007-0086. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2011-3192) [20:51] red hat has issued one [20:52] astrostl: it was just published [20:52] for lucid? [20:52] for hardy-natty [20:52] so, not for the *LONG-TERM SUPPORT SERVER DISTRIBUTION...* [20:53] hardy through natty [20:53] i seriously don't get ubuntu sometimes [20:53] hardy, lucid, maverick, natty [20:53] astrostl: read again [20:53] ooops :) [20:53] not showing on an apt-get update here [20:53] i'm constantly screwed by bugfixes that are only applied to new releases, pardon the outburst! [20:53] it will be available in approximately 12 minutes directly from LP, about 45 from security.ubuntu.com [20:54] tyvm [22:13] Hi everybody and first thanks for your time. I have a VPS 512 MB RAM running Ubuntu 11.04, MySQL and NGINX standard conf, but after 24 hours it is slow as hell, almost unreachable which makes it worthless. Any help on this issues would be most appreciated, here is the output from top: http://pastebin.com/eziWYrK8 === alexy_ is now known as alexy === alexy is now known as alexy_ === alexy_ is now known as alexy [22:25] Who wants to have a go at testing ubuntu-server beta1 candidate, last minute respin? :) [22:32] Daviey: sure, if it runs in a VM [22:37] done [22:51] I have an upstart process stuck in stop/waiting [22:51] When I try to run start it tells me that. How can I reset this stop/waiting [22:54] TheEvilPhoenix: sounds good to me! [22:55] TheEvilPhoenix: automated test cases have been achieved, but hand testing is helpful - http://cdimage.ubuntu.com/ubuntu-server/daily/current/oneiric-server-i386.iso [23:13] and http://cdimage.ubuntu.com/ubuntu-server/daily/current/oneiric-server-amd64.iso [23:14] plus http://iso.qa.ubuntu.com/qatracker/build/ubuntuserver/notcompleted with what is missing