[00:10] gents has anyone seen leguin.acc.umu.se within a netstat on their rigs? I did netstat -a | grep 127.0.0.1 and that came up. [00:13] lickalott, why would we? [00:13] does a user need to be logged in on the server, for samba to be running? (i want to set samba to share /media/drive1, /media/drive2) [00:13] I'm suprised that worked at all [00:14] cause that means your rdns is broken that grep matched 127.0.0.1 on any line [00:14] km0201, nope [00:14] hmm [00:15] cuz freenas has irritated me for the last time, and i'm seriously considering putting ubuntu server on a flash drive, booting it and just using it as my file server. [00:16] i'm pretty familiar w/ Ubuntu, and i've successfully done what I want to do, in Vbox, but I done it w/ a GUI... not sure I know how to do it w/ a CLI. [00:16] i guess i'll practice doing it w/ ubuntu server in vbox now.. :) [00:17] patdk-lap, the other question i have though... i have two drives on my NAS/File Server... (one is a mirror of the other).. i'm pretty sure i can figure out how to mirror them, but i don't need to be logged in, for those drives to auto mount (i'll add them to fstab) right? [00:18] ya [00:18] ya... i don't need to log in, or ya.. i do... ? [00:19] heh? [00:19] for what? [00:19] do i need to log in, for the drives to automount [00:19] the only reason to ever log in, is to perform mantaince [00:19] well, i can ssh in and do maintenance right? (this will be headless) [00:19] ya [00:20] ok. [00:20] add drives to fstab [00:20] ok... [00:20] configure samba normally using smb.conf [00:20] its smb.conf i'm not sure how to configure. [00:21] not hard, mainly depends on what kind of security your going for [00:21] just basic stuff.. user login w/ a password.... i'm used to the easy way with Nautilus (right click, sharing options, ).. [00:21] never done it by CLI [00:21] with or without an AD server? [00:22] AD server? [00:22] guess without [00:22] i guess i'll say w/o, cuz i don't know what that is. [00:22] this is just a home file server, no outside access, etc. [00:22] storage, etc. [00:22] then your going have to build a user database for samba, and users on the system [00:22] that is the most annoying mode I find [00:23] what do you mean?.. using it as a file server is annoying? [00:23] using samba in user login mode [00:23] security=user [00:23] oh ok... yeah, i'd like to have a user log in though, why.. i dunno.. but i would likee that [00:24] i'll cross that bridge when i come to it. [00:24] the only point of it is if you want people to have different rights to different folders on a share [00:24] and i don't. [00:24] you're actually right, there's no real point to it, now that i think about it. [00:24] you want everyone to just have access to everything? [00:24] well "everyone" is really only 2-3 people. [00:25] doesn't matter how many everyone is [00:25] 2 adults, one teen, and there's nothing on there that is compromising... home movies, music, family pics, etc. [00:25] i think thats what i'll do, is just allow anyone to access it. [00:26] normally that is easy then [00:26] use security=share [00:26] and just use guest access on the shares [00:26] well, i'll cross that bridge when i come to it. [00:26] and everyone logs in as guest :) [00:26] hmm [00:26] ok. [00:26] well, i'll cross that bridge when i come to it. [00:27] right now, i'm perfecting my testing in Vbox (5gig install drive, 2x5gig virtual drives, basically my setup on a much smaller scale)... [00:28] call me anal, i hate rolling stuff out w/ testing, i made that mistake going from freenas 6-8 [01:16] I have been using ispcp, just done a format and installed 10.10. I done apt-cache search control | grep panel and found out that dtc is included in the apt-cache, how ever I can not find any help docs on this. could someone please point me in the right direction [01:17] lifeless: good point [01:18] lifeless: `~/.ensemble/$envname is probably a sane default.. [01:19] SpamapS: apt is bust in oneiric atm for debootstrapping :( [01:19] SpamapS: also, that mail I sent you with friction - did you convert it to bugs, or should I, or are they all known ? [01:20] lifeless: I believe serge and I fixed that last Tuesday [01:20] lifeless: I fixed getall to be more sane... [01:20] lifeless: the rest I don't recall.. let me look it back up [01:20] bug 816606 [01:20] Launchpad bug 816606 in apt "apt postinst failure if ubuntu-keyring not installed" [Undecided,Confirmed] https://launchpad.net/bugs/816606 [01:20] is the apt thing [01:21] lifeless: yeah we worked around it by adding ubuntu-keyring as an explicit package to debootstrap [01:21] SpamapS: except ensemble-lxc doesn't seem to be nabbing it [01:22] tho I think I'll mark our bug, 817233, as a dupe of that one [01:22] lifeless: you have to rm -rf /var/cache/lxc/oneiric [01:22] sry patdk-lap got pulled away for a sec. [01:22] i have no idea what that host is. I saw "acc" and got a little worried. [01:24] SpamapS: trying [01:24] SpamapS: also, probably need to handle lxc containers already existing - e.g. ensemble-0 existing [01:25] bbiab, gotta run cat to vey [01:26] lifeless: right.. I have thought about changing to use a UUID. [01:27] SpamapS: or make it in /tmp ? :P [01:27] lifeless: hm.. it does just have to store a tiny state yaml and the formula zips [01:31] New bug: #817233 in lxc (main) "oneiric containers need ubuntu-keyring package (dup-of: 816606)" [High,Fix released] https://launchpad.net/bugs/817233 [01:33] patdk-lap, u still around? [01:33] anybody know a halfway decent samba tutorial? [01:40] !serverguide [01:40] The Ubuntu server guide may be found at http://help.ubuntu.com/10.04/serverguide/C/ [01:40] KM0201: ^^ [01:41] SpamapS, thanks... [01:42] i actually found that via google.. but it's telling me to find "security=user" in smb.conf in the global section, and i don't see it. [01:42] nevermind [01:42] its not under global, its under authentication [01:43] If the documentation has a bug, you should report it [01:43] doc bugs are arguably the worst kind [01:44] yeah, they are... thats why i'm fed up w/ freenas [01:44] they are viral.. copying themselves into peoples' memories [01:44] my needs are minimal, and i don't know why i never considered just using Ubuntu server (i'm pretty famiiliar w/ the Ubuntu OS, but.. never ran a completley CLI system, so im testing it in Vbox right now... with a GUI..lol) [01:45] lifeless: btw, the last time I was messing with that branch.. I had problems with the oneiric bootup.. something was borking hard [01:46] KM0201: you can run a GUI on a server, but we don't recommend it -- and that applies to ANY server, not just Ubuntu Server. [01:47] twb, yeah, i know.. but.. i know my way around the various GUI's... so i'm setting it up w/ an LXDE front in Vbox now.. (total of 3 virtual drives, to simulate my NAS on a smaller scale)... once i've taken good notes on what i done in the GUI, etc.. I'll test it w/ the server ISO. [01:47] honestly, as long as i've been using Linux, i should be a lot more comfortable in CLI than I am.. i've been spoiled by GUI tools. [01:49] what is nmbd, by the way?.. never heard of that one [01:50] nmbd is part of samba [01:51] Probably stands for something like "netbios message bus daemon" [01:51] oh ok. [01:51] See, Windows' filesharing was devised before TCP/IP and DNS had won [01:51] So Windows has its own alternative to DNS: netbios [01:52] windows has its own alternative to everything [01:52] And until XP, if not later, you need to have this complete replacement for netbios around, and what nmbd does is translate between NetBIOS and DNS [01:52] Is that still true with Win 2008 ? [01:52] SMB2 is allegedly MUCH better, on par with NFSv4, but I haven't looked at it. [01:53] Delerium_: I believe DNS is the default as at Vista or 7, but I don't know if netbios needs to still be there in the background [01:53] Delerium_: ##windows probably knows [01:53] "nmbd, which provides the NetBIOS-to-IP-address name service. NetBIOS over TCP/IP requires some method for mapping NetBIOS computer names to the IP addresses of a TCP/IP network." [01:53] https://secure.wikimedia.org/wikipedia/en/wiki/nmbd [01:54] twb, Thanks, I was just curious ;) [01:54] https://secure.wikimedia.org/wikipedia/en/wiki/Server_Message_Block#SMB2 [01:54] twb, didn't work with as a SysAdm on Windows for about 10 years... [01:54] Delerium_: I haven't even *used* Windows for about eight years [01:55] twb, Lucky you... I still support WAS on a couple of Windows boxes, but I don't support windows itself [01:56] twb, and I'm force to use a WinXP Laptop for "Windows Only" software... puke [01:56] I'm trying to setup a static ip on my server.. I'm curious on what a couple of the settings should be.. address is the IP address that I want it to set, subnet is the subnet of the network, what do I need to put into network, broadcast and gateway [01:57] On a trusted network, I would use DHCP for everything except the DHCP server itself. [01:57] (And instruct the DHCP server to assign fixed addresses to other servers.) [01:58] twb: I want to be able to ssh into the server not having to figure out what the ip address is [01:58] cjohnston: 11:57 (And instruct the DHCP server to assign fixed addresses to other servers.) [01:59] I'm not sure on my router how to do that [01:59] Well, if the router isn't running Ubuntu, you'd turn it off and have your server handle DHCP [01:59] (And DNS) [01:59] But if you want to go static, we can talk about that instead [02:00] twb: I have some stupid router provided by the provider.. its a modem/router in one that takes care of the TVs and everything.. I think I may have just figured it out tho.. 1 sec [02:03] what's a good tool to mirror a drive automatically. [02:03] dd [02:03] twb: I think I figured it out.. thanks [02:12] is there a SWAT alternative? (since it doesn't seem to be supported anymore) [02:12] An alternative to what? [02:13] Sorry, misread. IMO you should learn to manage samba from its config files. [02:14] ok, i figured that would be your position, cuz thats what the FAQ said..lol [02:14] i think i've about got it anyway, but it's all set up w/ a GUI front, so i'm gonna reboot, and set it up w/ server edition by my notes. [02:16] only thing i need to figure out is mirroring 1 drive to another. [02:18] back ina few.. [02:18] KM0201, One time deal? Or live mirror on 2 drives? [02:18] Delerium_, live mirror [02:18] i think i read about rsync [02:18] but thats the next thing i want to look ati. [02:19] KM0201, or use Software RAID 1 [02:19] can you hoold that htought?.. i'll be back in about 2min [02:19] check this out: http://advosys.ca/viewpoints/2007/04/setting-up-software-raid-in-ubuntu-server/ [02:19] oh, awesome. [02:19] Not sure how complete it is .. but looks pretty good at first sigh [02:19] sight... [02:19] well, i guarantee i will find errors..lol [02:20] KM0201, yeah ... Just found that this article date from 2007.. might be a bit old and maybe it has changes since then [02:20] If you want a live mirror you should be using either linux md RAID1, or a $500 raid card. [02:20] well, i'm definitely not using a $500 raid card [02:20] well, maybe "live raid".. isn't the exact word i was looking for. [02:21] maybe something that sync's once a day? [02:21] KM0201: if you want an archive of what the files were as at , you want something more like rsnapshot [02:21] hmm, ok.. well, hold that thought, i'll be back in 2min. [02:21] Which is the hard disk equivalent of rotating incremental tape archives. [02:28] so... rsnapshot? [02:31] It depends whether your goal is disaster recovery or archival [02:32] or live redundancy [02:32] Delerium_: I was folding that into DRP [02:33] twb, k ;) We often have our own wording depending of the infrastructes [02:33] er.. infrastructures [02:33] What I would do would be something like md RAID1 a pair of disks, then on it have /boot md0, LVM PV md1, LV root filesystem and LV rsnapshot filesystem. rsnapshot performs a nightly cp -al and rsync from the former two to the latter, you end up with incremental daily/weekly/monthly/yearly archives, and up to one disk can die. [02:34] To guard against *real* disasters (as oppose to SPOF of a HDD); I would probably also/instead have rsnapshot backing up to an offsite medium. [02:35] offsite or online backup is always good, incase of lightening strikes and other power issues [02:36] I was thinking more "building burns down" [02:36] twb: this is a simple file server man... lol, i just want a mirror of the drive in case of drive failure, i back up to a large USB drive once a week also, so.. [02:37] KM0201: well, that sounds more like RAID1 [02:37] twb: ok, how would i set that up? [02:37] ya, but power issues taking out all drives is more common than buildings, I would think [02:37] KM0201: you need to do it at install time [02:37] oh really? [02:37] hmmm, [02:37] ok. [02:37] (Technically you can probably do it post-facto, but it's a huge pain in the arse) [02:37] it's easiest at install time :) [02:37] Mainly because grub is such a bastard [02:38] gotcha, so when its asking all that install crazyness, choose a "raid" [02:38] i remember seeing that during the install process. [02:38] KM0201: when it asks you about HDD partitioning, you need to choose "manual" and set up the disks appropriately. [02:38] right. [02:38] i've done manual set up a lot. [02:38] Unfortunately there is no "Automatic - use RAID1 entire hard disks" choice [02:39] understood. [02:39] I should probably complain about that to the d-i maintainers -- surely 2×RAID1 and 3×RAID5 are common enough occurrences to automate [02:40] Speaking of HDD management, I didn't check lately, does ZFS makes its way to Linux!? [02:41] ZFS has the cuddle of death [02:41] Wait for BTRFS [02:42] hummm BTRFS is from Oracle, right? [02:42] Technically both zfs and btrfs are from oracle. [02:42] well, other than rsnapshot, i think i got all this figured out [02:42] KM0201: rsnapshot is not very complex; it basically just does a nightly cp -al, then rsync [02:43] twb: and i can schedule that? [02:43] i guss my final question is (and i'm pretty sure i know the answer..).. [02:44] should i install Ubuntu Server, onto a Thumb drive, the way I did freenas... or knock off about 10-15gigs of one of my drives, and put the OS there [02:46] rsnapshot is triggered by cron, so sure [02:46] i know squat about cron [02:46] Ubuntu will run off a USB key, though not very fast for writes [02:46] It makes e.g. "apt-get update" a pain in the arse [02:47] well, honestly, i dont plan on updating this, etc.. [02:47] i might even go through my router, and block the server from accessing the internet [02:47] I would probably allocate 256MB /boot, 2 to 8GiB / and (to begin with) 128GiB /srv for your data [02:47] ooh i'm not gonna do all that craziness [02:47] Assuming you're using LVM, you can expand the latter two as needed. [02:48] twb, LVM can't expend /boot ?! [02:48] (just wondering...) [02:48] Delerium_: not if it's not on LVM [02:49] twb, make sense [02:49] And IMO it is more dangerous than helpful to put /boot on LVM, particularly since it should have negligible growth -- 256MB is enough for (say) a dozen kernels. [02:49] twb, but can Linux boot on a LV ? [02:49] grub2 can boot from a RAID1 LVM /boot, at least, but it's flaky and fugly. [02:50] twb, thanks ... I was a SysAdmin in the past, but I now mostly support Middleware product... I'm just trying to catch up with all those lost years.. [02:53] ok, downloading the server ISO [02:53] A mini.iso will be faster, because it'll only download what you need [02:53] hmm, good point [02:54] twb: that doesnt seem to be working. i think i need to property setup static address.. i have the hdd right now in a machine that has a head, however the HDD will go to a headless machine when it starts working [02:54] http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/current/images/netboot/mini.iso [02:54] cjohnston: OK [02:54] cjohnston: so what do you need help with? [02:54] i'm gnna use 11.04 [02:54] KM0201: I recommend you deploy LTS in production [02:54] twb: naa. [02:55] Shrug. On your own head be it. [02:55] i'll deploy LTS when 12.04 comes out. [02:55] yeah, i won't hold you accountable.. :) [02:55] i've been testing all this in vbox on 11.04 [02:55] twb: I have auto etho \ iface eth0 inet static \ address 192.168.1.100 \ netmask 255.255.255.0 [02:55] you know what though [02:55] i think i'm gonna do a server install, cuz i'm gonna try to set up a raid i think [02:55] twb: I guess there are network broadcast and gateway tha im not sure about [02:56] cjohnston: one moment [02:56] cjohnston: I think you only need address, netmask and gateway [02:56] (gateway only if you want to be able to access other networks.) [02:56] what would I put in gateway? the address of the router? [02:57] yes [02:58] ok... let me switch the hdd to the other computer and see if itll work [03:00] cjohnston: you should check /etc/udev/rules.d/70-persistent-net.rules first [03:00] cjohnston: make sure the MAC for (say) eth0 is right [03:00] o [03:01] Or just delete that file entriely, which shoudl result in eth0 being the iface on next boot === ejat- is now known as ejat [03:07] twb: if i install server in a Raid1, is there a way to verify it is working properly, and is being properly mirrored to the "second" drive? [03:08] KM0201: cat /proc/mdstat [03:09] http://paste.debian.net/124729/ shows a synced three-disk RAID1 array, md0 /boot, md1 LVM [03:09] When the drives are out of sync, there will be a progress bar instead and an ETA. [03:10] hmm, ok [03:10] thanks [03:11] Note that immediately after install they *will* be syncing; this is because it has to sync even the unused blocks, because the md layer can't tell which blocks are important [03:11] ok... [03:11] (an ATA TRIM-like syscall to md, or btrfs' built in RAID layer, would fix that.) [03:12] hmm, ok. [03:12] Also, write-intent bitmaps are off, so if you reboot the box before the sync finishes, it'll start over again from scratch. [03:12] well, i'm gonna try it in vbox first... i hope it works. [03:12] *bitmaps are off _by default* [03:12] ok, thats good info, thanks [03:13] i think i'm gonna like this a lot better than FreeNas. [03:13] should've done it a while ago, but Freenas 8 finally pushed me over the edge [03:14] twb: and when i set up samba, i only need to set setup samba to see Drive 1, right? (since the server will handle syncing the two drives "behind the scenes") [03:15] KM0201: md RAID will mean your system sees three drives -- the sda and sdb low-level drives, and the md0 RAID pseudo drive. [03:15] Only the md drives will be mounted. [03:15] Samba itself doesn't care about that, because it just deals with the VFS layer -- i.e. files and directories, not drives. [03:16] ok.. === ejat- is now known as ejat [03:36] twb: got time for a quick question? [03:36] !ask [03:36] Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) [03:39] i know, i know, i didn't know if you were still here.. [03:39] ok [03:39] i've got scsi3 /swap, / and /media/Disk [03:40] scsi4, is currently unformatted. [03:40] i want scsi4, to be th emirror. [03:40] so i go up to "configure software raid" [03:41] then create MD device [03:41] raid 1, which is mirror. [03:42] now, this is where i get confused, its asking for "active devices in the raid1 array"... [03:42] it's set to 2.. (which is minimum) [03:42] i guess if i choose 3, then it will mirror the entire drive.. right? [03:44] KM0201: what scsi drives do you have ? [03:44] i'm testing this in vbox. [03:45] KM0201: uh, the two nodes should be your two hard disks [03:45] Or two partitions, one on each disk [03:45] You will need to move any data on those disks elsewhere before you do the install [03:45] ok. [03:45] right, i understand that (the virtual disks are blank) [03:49] hmm === cerber0s is now known as cerberos [04:32] twb: i don't ned to set up swap space to be mirroed do i?.. i just set my ext 4 / to mirror, and that was it. (but both drives have a swap partition) [04:36] twb: it worked! (at least in vbox) [04:36] now i just gotta do it again, and take detailed notes [05:46] hey guys. i want a root shell. vps i think. where i can do anything with root acces. install apps. do anything. do i need vps hosting . thats what i see always. i thing its just for hosting websites with some additional acces..? so what do i need. and i want a cheap one. 15 dollars / mo ? [05:47] jane1: take a look at http://prgmr.com [05:49] greppy is it yours? [05:52] hi, does anyone know if I need udev (udevd) on a VPS? [05:52] janel: If you need root shell access a VPS will work... There are lots of providers Linode.com, MediaTemple.com, Prgmr.com, VPS.net the list is long :) [05:53] greppy is it yours? [05:53] defunct: I suggest you keep it... udevd handles events and such [05:53] I personally love Linode. [05:53] greppy ahem [05:53] rww: I second that although MediaTemple gave me thousands of dollars in free hosting at OSCON :P [05:54] Free VPS even :P [05:54] janel: Probably not his and he is likely idling [05:54] rww: Can I PM you? [05:54] bkerensa: hrmm. I was hoping I could free up some memory by removing it [05:54] bkerensa: about? [05:55] rww: #ubuntu-server contributions :) [05:55] I've only been in here for four hours, but sure, why not. [05:55] jane1: no, but I am a customer of theirs :) [05:55] defunct: I would not suggest it... It does use a bit of memory but likely you can make other optimizations to reduce memory usage [05:56] janel: Linode has the best benchmarks and likely the most economical pricing model (Google) [05:56] bkerensa: alright, thank you :) [05:56] greppy if ineed 512 or ram. it will be? [05:57] defunct: Is your VPS for web stuff? [05:57] jane1: they offer that as an option, yes. [05:57] greppy 4 dollars + 512/64? [05:57] k [05:58] defunct: I just have lighhttpd running on it, but mostly it's for irc and and a ts3 server [05:59] defunct: but still, only 100mb ram free atm (from 512) [05:59] whoops meant bkerensa :P [06:00] defunct: No MySQL? [06:00] bkerensa: not yet, no php or mysql [06:00] bkerensa: hrmm I can probably shut off getty [06:01] defunct: Well just make sure your swap is good so you dont thrash [06:02] defunct: 1gb swap should do... [06:06] I'll just leave it all for now, until swamp is actually getting used, so far it hasn't [06:07] syslog 439 0.0 0.2 34452 1476 ? Sl Jul30 0:00 rsyslogd -c4 [06:07] doesn't syslog use like less than half of that normally [06:49] hallyn: around ? [06:57] lifeless: still banging away on it? [06:58] SpamapS: put ensemble on hold till apt is fixed (because ensemble wants oneiric) [06:58] SpamapS: but I plan to nab mvo soon to talk about the apt-key postinst script call [06:58] SpamapS: right now I'm running into [06:58] lxc-start: Device or resource busy - failed to remove previous cgroup [06:58] on start [06:58] after a clean vm boot [06:59] using cgroup-bin? [06:59] it is present, yes. [07:00] I have yet to be successful w/ that installed [07:00] worked fine on natty. Will try removing it. [07:00] it kindof sucks though, to have a nice helper and have it be incompatible. [07:01] so the thing I was trying next was back to 'why does echo /dev/null | sudo lxc-start -n ,,, -d ' reset my sudo password ticket [07:01] which I'm hoping oneiric fixes. [07:01] :-P === smb` is now known as smb [07:37] morning [07:52] Hi all, I have an udev rule: ACTION=="add", SYSFS{model}=="RD1000", NAME="RD1000%n" Now, when I insert the first RD1000 (internal), it becomes /dev/RD10001, I then insert the second RD1000 (usb), and it becomes... /dev/RD10001, while I was presuming the %n would make it RD10002. What am I grossly overlooking? === jamespage1 is now known as jamespage [09:36] <_ruben> hmm, what takes care of displaying the motd on login via ssh? PrintMotd in /etc/ssh/sshd_config is set to "no", yet it is displayed on each *first* login, tho i'd like to show it on *all* logins [10:21] SpamapS: yeah, removed cgroup-bin, works. >< >< >< >< >< >< >< [10:23] <_johnny> perhabs your keyboard is broken [10:51] why centos is considered a better server than ubuntu? [10:53] <_ruben> ask the centos ppl [10:54] whats your choice in ranking [10:55] <_ruben> 1. ubuntu [10:55] <_ruben> [10:55] <_ruben> never used centos [11:02] hm [11:02] fedora ? [11:02] _ruben [11:02] debian [11:02] <_ruben> in the past i've used slackware/debian/suse (roughly in that order) [11:03] The question "why centos is considered a better server than ubuntu?" is essentially meaningless, since it fails to specify who is doing the considering [11:03] I doubt most of us would be on this channel if we upheld that opinion [11:04] <_ruben> exactly [11:12] any suggestions for a domain name that has music, education, videos, students , chating , bloging... video chats. text chats.. etc. multi purpose communication . ? [11:14] _ruben: I suspect many people really appreciate the lack of predictable release schedule, and a not having much confidence in the future stability. [11:14] === Ursinha-afk is now known as Ursinha [11:15] <_ruben> Daviey: agreed :) [11:15] buenos dias [11:17] Ursinha: Salem! [11:18] any suggestions for a domain name that has music, education, videos, students , chating , bloging... video chats. text chats.. etc. multi purpose communication . ? [11:18] jane1: no [11:19] (you don't need to keep asking, people will respond if they do... ) [11:19] 'no' is nice but it makes a negative sence [11:19] oh ok. [11:19] Although, it is pretty Off Topic for here. [11:23] ya.. only the website is in ubuntu hosting [11:33] apt-get update [11:35] * _ruben redirects Shaggy2's terminal [11:36] <_ruben> jane1: music-education-videos-students-chating-bloging-videochats-textchats-etc-mulit-purpose-communication.{org,net,com,...} === bazhang__ is now known as bazhang [11:54] i have a script which runs /etc/init.d/nfs-kernel-server. when running the script from roots crontab it fails to start. [12:02] <_ruben> trapmax_: quite likely an environment varaible thing, like lack of /sbin in the path [12:06] k, thanks. i'll look into that [12:07] trapmax_: You may want to make sure that the cron line is running at all, check /var/log/syslog for an entry. [12:10] Pici: i used "0 * * * * /path/to/my/script > /path/to/script/log", and got nfs output as log, so i know it ran, but errors were sent via email concerning the modprobe and start-stop-daemons. it would make sense that env vars are not set properly [12:12] trapmax_: Put 2>&1 at the end of that. That will ensure that anything that was sent to stderr will be redirected to stdout, which you've already directed to the logfile. [12:19] Pici [12:19] any suggestions for a domain name that has music, education, videos, students , chating , bloging... video chats. text chats.. etc. multi purpose communication . ? something that have hub or communications... gatway... [12:20] jane1: no. And I don't understand what that question has to do with Ubuntu either. [12:20] its offtopic [12:20] and thx [12:21] jane1: Try #ubuntu-offtopic then [12:21] no response [12:22] <_ruben> probably noone cares there either :) [12:22] ya [12:51] hi [12:51] Can someone help with krb5 and nfs4? [12:51] when i touch a file on the mount, it is owned by nobody, even when the user has his kerberos ticket and a ticket for the nfs-server principal [12:53] what is the ram and proccesor requirments for ubuntu server? [12:54] jane1, the base system will boot and run in 256 [12:54] after that, it really depends on what you're hoping to do with it. [12:55] gui . gnome or kde will run in 256 too ? [12:56] smoser ? [12:56] jane1, i dont really know. gnome or kde are not really ubuntu server. [12:56] i only have 256 of ram at a vps. need to run apache, mysql, and an ircd. will do ? [12:56] jane1, most certainly people do that. [12:57] jit wil run [12:57] ? [12:57] those people don't run cnn.com or slashdot.org [12:57] but it will run. [12:58] jane1: yes, it will run [12:58] if you expect high load, get a better server [12:59] jane1, i would suggest not running an X stack there though [12:59] that is just going to waste precious memory [13:00] smoser x stack? [13:00] X11 [13:00] (you mentioned gnome and kde) [13:00] ya. but whats x11? [13:13] lifeless: i'm around now. what's up? [13:41] hey is there a way to limit /home/user1 and /home/user2 each to a certain limit of size [13:42] ex: each to 50GB in an OpenVZ vps? [13:56] sw0rdfish: if you want it independent of the owners of the files [13:56] xfs projectquota can do that [13:56] Otherwise, normal kernel-quota stuff [13:57] lynxman, cloud-init test ? [13:57] smoser: looks okay here [13:57] you tested ? [13:57] trunk for your changes [13:57] or two separate filesystems [13:58] sw0rdfish: that was for you as another suggestion [13:58] (context was perfectly fine inside my head) [13:59] hmmmmmm [13:59] sw0rdfish, openvz supports user quotas inside a container, but it's global not path specific [14:00] smoser: yes sir I did [14:00] I see. [14:00] smoser: maybe the permissions could be an issue but tbh not too concerned [14:03] the permissions in trunk on the private key are 600 or 400 [14:03] no way should a file with credentials have 644 perms [14:04] SpamapS, ping when you're in [14:09] smoser: they should be 400 I do agree [14:37] New bug: #815504 in bind9 (main) "glibc double free when using postgres dlz" [Undecided,New] https://launchpad.net/bugs/815504 [14:42] New bug: #819319 in lxc (main) "more lxc scripts are broken with multiple cgroup mounts" [High,In progress] https://launchpad.net/bugs/819319 [14:47] New bug: #817220 in bacula (main) "package bacula-director-common 5.0.3-0ubuntu2 failed to install/upgrade: problèmes de dépendances - laissé non configuré" [Undecided,New] https://launchpad.net/bugs/817220 [14:47] New bug: #817779 in mysql-dfsg-5.1 (main) "package libmysqlclient16 (not installed) failed to install/upgrade: trying to overwrite '/usr/lib/libmysqlclient.so.16.0.0', which is also in package mysql-cluster-client-5.1 0:7.0.9-1ubuntu7" [Undecided,New] https://launchpad.net/bugs/817779 [14:47] New bug: #818105 in qemu-kvm (main) "package qemu-kvm 0.14.0 noroms-0ubuntu4.4 failed to install/upgrade: erro ao escrever para '': Arquivo ou diretório não encontrado" [Undecided,New] https://launchpad.net/bugs/818105 [15:06] robbiew: http://ubuntuqa.wordpress.com/ [15:08] patrickmw: sorry, but what am I supposed to notice exactly :/ [15:09] robbiew: this is the qa automation status report. you had asked about the package testing. its on the backlog at the moment [15:09] I will be adding dates soon [15:11] patrickmw: ah...so basically a wiki to tell me the same info...heh [15:12] robbiew: yup. but when we start on the package stuff, this will be the link that will show what packages have been "checked off". I will keep you posted [15:14] ah...gotcha [15:51] I'm trying to find the users using the most disk space on a server -- is there a quick command to do this? e.g. like "du /*/* -s --block-size=1G | sort -g" except for sorting by user instead of file [15:53] smoser: pong, I'm in [15:54] SpamapS, updates to https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/810044 [15:54] Launchpad bug 810044 in cloud-init "cloud-init will have race conditions for cloud-config with multiple network adapters" [Medium,Confirmed] [15:56] that is one big block [15:57] Hey guys, I'm getting increasingly more angry with the dhcp3-server. [15:57] smoser: so.. I've been thinking hard about how that works.. and I'm worried that the event will be too rigid for my eventual intended use (as a part of rc-sysinit) [15:57] It's not reading any configuration files. [15:57] Anywhere. [15:57] SpamapS, why to rigid? [15:58] smoser: because one can have an interface in /etc/network/interfaces that comes and goes.. [15:58] There's no /etc/default/dhcp3 , first off, and all the guides I've seen tell me there should be... [15:58] smoser: for instance, one could list a USB network interface as "auto" .. that just means "when its there, bring it up" [15:58] smoser: its a bit of a corner case, but I think we can handle it gracefully [15:59] well.. [16:00] i think the implementation i have there is corredt for that [16:00] your case is handled (i think) best by the user not using 'auto' [16:00] but 'auto-hotplug' [16:00] in which case, the code i have there right now works correctly [16:00] smoser: I was thinking after networking (ifup -a), if static-networking-up hasn't been emitted, we need to make some kind of call about the remaining interfaces.. not sure what though. [16:00] as it would be ignored as "auto" but still brought up on hotplug [16:01] smoser: is auto-hotplug already a thing? [16:01] i am going off what i read in interfaces(5) only [16:01] * SpamapS should read it again. ;) [16:01] sorry [16:01] allow-hotplug [16:02] Right, ok [16:02] that makes sense. [16:02] yes, but whether or not it actually happens, i dont know [16:02] :) [16:02] it makes for good doc [16:02] How about a timeout that drops you to single user mode if the "auto" interfaces haven't all come up? [16:03] the other thing is that your event will possibly fire multiple times in the lifetime of a sytem [16:03] ie, if i : ifdown eth0 ; ifup eth0 [16:03] it will fire again [16:03] exec ifup --allow auto $INTERFACE [16:03] That is what happens *now* on net-device-added [16:03] which is basically "hotplug" [16:04] well that sucks. [16:04] smoser: I was thinking about using a job, 'static-networking' instead of the event. [16:04] smoser: well we could do if ! ifup --allow auto $INTERFACE ; then ifup --allow hotplug $INTERFACE [16:05] smoser: ifquery may help here tho [16:05] ifquery is broken [16:05] for mapping devices [16:05] so i dont trust it [16:06] smoser: are you sure the mapping interfaces is a problem? Don't all mapped interfaces show up as real=mapped in ifquery --list ? [16:07] smoser: ahh, reading the comment now [16:07] SpamapS, did you read what i wrote there? [16:07] yeah [16:07] ok. [16:08] i'll be back in 10 minues === zz_ng_ is now known as ng_ === oCean_ is now known as oCean [16:17] SpamapS, so i suspect that in ubuntu 'allow-hotplug' would be difficult to implement. [16:17] as that is basically how *all* interfaces are brought up [16:17] on udev events at startup [16:19] smoser: well ifup -a will try to bring up all that exist, in case they don't emit a udev event [16:20] SpamapS, probably not [16:20] smoser: since that starts on stopped udevtrigger .. that should be sufficient for any servers. [16:20] or it could be considered a bug if it did [16:21] smoser: it does. the point of /etc/init/networking.conf is a "last ditch" to bring up any critical networking pieces [16:21] should 'ifup -a' try to bring up a device that was listed in interfaces with 'allow-hotplug' [16:21] no [16:21] auto [16:21] right. you said "all that exist", which is what i was confused by. [16:22] so anyway, i think what i have there in my branch should work for all cases other than the usb case. [16:23] and ofr my use, that is not a big deal. [16:23] For the USB case I'm ok w/ release notes saying "use conman for usb interfaces that you want to bring up automatically" [16:24] smoser: appreciate you completing my work.. I've been fretting over this for a week trying to think of all the error conditions [16:25] connman or network-manager ? [16:25] nI guess network-manager --no-install-recommends won't bring in anything evil [16:26] SpamapS, i pinged slangasek in -devel hoping to ask about the intent of 'ifquery --list' [16:26] it generally seems broken to me [16:26] smoser: ifupdown is *evil* [16:26] smoser: did you try to read the code? [16:26] main.c ? [16:26] smoser: that is not the code [16:26] :-/ [16:26] yeah, and i thought WTH [16:26] smoser: that is generated [16:26] well, obviously. [16:26] but it wasn't clear from where. [16:27] ifupdown.nw [16:27] Which is written in some weird "literate programming" language [16:27] Some crazy idea where you write code and a postscript doc together [16:28] what processes .nw -> main.c ? [16:28] 'nowebm' [16:29] wow. [16:29] ok [16:29] so where do you think we are ? [16:29] oh.. [16:29] th eothe rissue i had to talk to you about SpamapS in this regrad [16:30] smoser: well I have oft wondered if we wouldn't be better off abandoning the ifupdown code base and just re-implementing it with C++ or python or something. [16:31] i *think* that my need for blocking is handled fine. [16:31] a diff to cloud-init like: [16:31] http://paste.ubuntu.com/656489/ [16:31] SpamapS, it can't be python [16:31] it has to be /bin or /sbin [16:31] (which my removal of 'cut' was good for that too) [16:32] smoser: yeah I actually have a branch that cleaned that up too w/ shell only [16:32] SpamapS, in theory, would you tink my cloud-init-nonet would function the same with this change? [16:33] smoser: but I was definitely relying on ifquery --list [16:33] ie, it shoudl block for me [16:33] smoser: and don't feel bad that we've been working in parallel. The amount of knowledge we've gained is enormously valuable. :) [16:34] smoser: I'm not sure, what were you going to change the start on to? [16:34] smoser: or wait, its the stop on that you want to change right? [16:34] (i just started looking at this this morning, hoping i could shove something in before alpha3) [16:34] see pastebin above. [16:34] right. [16:34] on a single instance system that should be the same moment [16:34] err [16:34] interface [16:35] right. [16:35] but i need to handle multiple interfaces [16:35] And if you have 3 interfaces and a bridge, then it would be a more correct version of the same moment. :) [16:35] smoser: it would make your comment wrong.. you'd be blocking until all auto interfaces are up, not just a non lo interface [16:36] i'm concerned about blockign system boot [16:36] thats what i *want* [16:37] i think i get that because of cloud-init.conf is 'start on mounted MOUNTPOINT=/ and stopped cloud-init-nonet [16:37] ' [16:37] smoser: I'd like to see this accompanied by a test case in the build... I'm worried about weird /etc/network/interfaces files that work w/ ifupdown but not with your ghetto parser. ;) [16:37] smoser: for your purposes though.. yes I think this solves your case. [16:38] i dont really know how the parser could go wrong [16:38] smoser: for the purpose of delaying rc-sysinit's switch to runlevel 2.. I'm not entirely sure about it. [16:38] white space delimited first field is "auto" [16:38] (or allow-auto) [16:38] anything after a '#' is stripped [16:38] smoser: " A line may be extended across multiple lines by making the last character a backslash. [16:38] take all the other tokens [16:39] actually, readline works then. [16:39] interestingly [16:39] OH cool [16:39] i'm pretty sure. let me test that though. [16:39] i think i've beeen pissed off by that before [16:39] :) [16:39] ok reading the man page, I think you're covered if read line works [16:40] So, I'm willing to upload this for A3 [16:40] let me test it out on a VM tho. :) [16:40] SpamapS, http://paste.ubuntu.com/656500/ [16:41] just run that, you get the 3 lines concat'd [16:41] yeah that makes sense [16:42] SpamapS, you build this, i'll build a cloud-init and give you a ec2 instance to play with [16:42] build for amd64 [16:43] ok, I'm also going to configure a VM with 3 interfaces and a bridge.. ;) [16:44] hrm.. why haven't we dropped the size of the server iso yet? [16:45] RoAkSoAx: hey [16:48] hallyn: hey [16:48] hallyn: what's up [16:48] smoser: built, seems to emit on my laptop which has only lo.. :) [16:48] RoAkSoAx: hey, cani get some sponsorin luv? :) [16:48] RoAkSoAx: http://people.canonical.com/~serge/lxc.debdiff [16:48] hallyn: sure [16:49] RoAkSoAx: thanks [16:49] smoser: were you thinking that we should touch a file in /var/run or something so we don't emit this twice? [16:49] SpamapS, i dont know. it seems like it may be ok to emit twice [16:50] right? [16:50] smoser: since its not waited on, it won't affect ifup/ifdown commands [16:50] smoser: but it could cause issues for things that 'start on x and static-networking-up' [16:52] SpamapS, i can defer to you on that. [16:52] what will happen for my case? [16:52] smoser: for yours nothing, because yours only runs at start [16:52] smoser: I think the right thing to do is to view this as a singular event, much like mountall does.. and never emit it again [16:53] smoser: Otherwise the system gets super reactive to any admin activity [16:54] well, not so much, SpamapS [16:54] on 'start on x and static-network-up' [16:54] those would not really fire more than once, right? [16:54] smoser: ifdown eth0, ifup eth0 .. would emit it again on a system with auto eth0 [16:55] smoser: so if thats a task, its now waiting for 'x' to happen again [16:55] ok. [16:55] I can't see a time where this needs to happene very time the static network comes up [16:55] so then, we can touch a file that said its been done [16:55] This is an event to boot things.. the "state" we're looking for is runlevel 2 [16:56] smoser: I'll actually touch a file before emitting.. as thats more atomically correct. :) [16:56] It would be interesting to look at how ChromeOS is doing their upstart boot now.. [16:57] SpamapS, http://paste.ubuntu.com/656507/ [16:57] don't touch [16:57] use : > [16:57] or something else [16:57] Keybuk was saying they just have two jobs .. boot-services and services .. and everything follows those two jobs' starting and stopping events [16:57] touch is /usr/bin/touch [16:57] damnit! ;) [16:58] no, no no no not return 1 [16:58] hallyn: done! [16:58] smoser: this is a racey lock.. [16:59] true [16:59] you want me to un-racify it ? [16:59] (i think return 1 is correct, htough) [17:00] its not an error [17:00] smoser: we're just choosing not to emit [17:00] : > /var/run/network/static-network-up-emitted || exit 0 [17:01] something like that [17:01] returning 1 is not an error. [17:01] oh, you can't do it there. ;) [17:01] i'm confused. [17:01] thats the race [17:01] yes i can fix a race there. [17:01] you check the lock in all_interfaces_up, but create it blindly later [17:02] i or you are missing something [17:02] hmm.. : > doesn't error if the file exists already.. :-/ [17:02] there *is* a race condition there, you are correct [17:02] doesn't matter. [17:02] you dont get there unless you had the lock [17:02] but we dont have a lock [17:02] for a lock, the easiest thing to do is 'mkdir' [17:03] ahh good point it is atomic [17:03] mkdir /var/run/network/static-network-up-emitted || exit 0 [17:03] RoAkSoAx: thanks! [17:04] http://paste.ubuntu.com/656510/ [17:04] http://paste.ubuntu.com/656513/ [17:05] yeah. [17:05] 6 [17:05] 1/2 dozen [17:05] hmmmm which is better? [17:05] I like the && [17:05] one exit is simpler to read [17:06] the only issue is that we're taking the lock before we've emitted the event [17:06] so at very least we should rmdir on failure of the event to emit [17:06] but... [17:06] we're kinda screwed in that scenario [17:06] yep [17:06] all kinds of stuff breaks if initctl doesn't work right .. all over the place [17:07] we've stipulated that it works [17:07] and its authors are careful to make it as reliable as possible [17:07] smoser: push that, I'll merge it [17:09] when I upgraded from 9.10 to 10.04, there was a bug with dovecot [17:09] what is a good way to reinstall dovecot on 10.04 [17:09] SpamapS, pushed. [17:09] I want to completely whipe out the configs for it etc... [17:10] and start fresh on 10.04 [17:11] SpamapS, lets test it a bit, and then go with it [17:11] orudie: apt-get purge name_of_package_to_remove [17:11] orudie: note that that will also remove all the *data* that belongs to said package [17:11] SpamapS, thanks it worked [17:11] orudie: so, just make a backup first :) [17:12] SpamapS, :) [17:14] SpamapS: I don't have write access to lxc tree yet. Any chance I could convince you to do 'pull-lp-source lxc natty; bzr branch lp:ubuntu/natty/lxc; cd lxc; bzr import-dsc ../lxc*.dsc; bzr push' and do that for lucid, maveirck, and natty? (when you get a chance) [17:16] hallyn: importer not working? [17:19] Daviey: apparently [17:20] at least for natty it is one changelog entry behind [17:20] hallyn: are you sure? [17:21] hallyn: ahh, using the wrong pocket.. [17:21] natty-updates [17:22] hallyn: lp:ubuntu/natty-updates/lxc [17:23] Daviey: damn. i was thinking bzr trees just tracked updates [17:23] Daviey: thx [17:25] hallyn: it would make sense, would it not? :) [17:26] Daviey: how it's done makes sense :) [17:53] hello: when doing a large file copy across a NFS mount I see around 80% of load (according to top) is in iowait. Does this sound right to you? [17:54] only 80%, wonder why it's not 100% [18:05] i've got two disks mounted under /media (Disk_1, Disk_2).. both are ext4, i can read/write them from the server, but not via samba, i always get permission denied. [18:09] i've got two disks mounted under /media (Disk_1, Disk_2).. both are ext4, i can read/write them from the server, but not via samba, i always get permission denied. [18:10] Daviey: there? [18:28] hggdh: always [18:33] * negronjl is away: out to lunch [18:36] smoser: any ideias? http://paste.ubuntu.com/656580/ [18:36] i've got two disks mounted under /media (Disk_1, Disk_2).. both are ext4, i can read/write them from the server, but not via samba, i always get permission denied. [18:37] i'm assuming this is something i've done in smb.conf , but i don't know what. [18:38] RoAkSoAx, looking [18:38] KM0201: I assume you have verified the Samba share permissions? [18:38] where do you see that ? [18:38] what is the metadata, RoAkSoAx [18:39] KillMeNow: how would i check that? (no i don't think i have, i just followed a basic tutorial) [18:39] obviously its a permission error, just.. not sure how to fix it. [18:40] RoAkSoAx, i suspect that you have a dict in the yaml that you gave it for cloud-config rather than a list [18:41] i can do this no prob w/ NTFS.. [18:41] so it's obviously something i'm doing wrong [18:42] smoser give me a sec and ill send u the late command using that [18:42] RoAkSoAx, am i right in guessing it has 'ssh_authorized_keys' [18:42] ? [18:43] smoser give me a.sec as im reinstalling should be done in ~4 [18:51] smoser: so this would be the error: not an array right? [18:51] right. i suspect that your meta-data or user-data is incorrectly formed [18:52] smoser: yeah I guess that came from changes made to ensemble [18:52] KM0201: IIRC it should be ls -la /path/to/share [18:52] Daviey: are you still around? [18:53] RoAkSoAx, note, that ssh keys is strange in metadata [18:53] KillMeNow: what does that do? [18:53] KM0201: you should see the typical file permissions but instead of a drwx it should show srwx or something like that [18:53] been a while since i dinked with samba [18:53] oh ok. [18:53] then you can set up permissions like any other directory using chmod [18:53] RoAkSoAx, http://paste.ubuntu.com/656588/ [18:54] and chown if necessary [18:54] hello all [18:54] if you're feeding it meta-data and putting ssh keys in it, you'll have to do like that [18:54] KillMeNow: i have the directory right.. (like i said, i can create files in it from the server, and browse from a network machine).. wha ti can't figure out, is how to write from a network machine [18:54] i've tried several chown/chmod commands [18:54] smoser: ok cool. I think it's just a change on the way of doing things cause it was working just fine [18:55] are you doing it in user-data or meta-data ? [18:55] I was asked by #httpd to show the output of 'apache2 -S' but I get the error "apache2: bad user name ${APACHE_RUN_USER}" [18:55] smoser: fwereade was working on some changes to convert stuff to twisted so he might have changed that [18:55] smoser: user-data [18:55] How can I get that to work? [18:57] RoAkSoAx: do you have a few more minutes? (and are you sufficiently familiar with UDD?) [18:58] RoAkSoAx: if you'd be so inclined, there are two bzr trees linked to bug 753308. they need to be pushed to -proposed. [18:58] Launchpad bug 753308 in lxc "Restart containers on upgrade" [High,Fix released] https://launchpad.net/bugs/753308 [18:58] hallyn: i am indeed [18:59] hallyn: i'm actually on my way out but can take a look at them when I get back unless someone beats me to it [19:03] RoAkSoAx: then i'll ask Daviey, thanks :) [19:03] Daviey: can I get 3 minutes of your time? there are two bzr trees linked to bug 753308. they need to be pushed to -proposed. [19:03] Launchpad bug 753308 in lxc "Restart containers on upgrade" [High,Fix released] https://launchpad.net/bugs/753308 [19:06] Daviey: and before complain, I'm working on https://wiki.ubuntu.com/SergeHallyn/ServerDeveloperApplication :-) [19:06] \o/ [19:08] hallyn: not to be a wet blanket, but wouldn't this be suitable to add to a future SRU? [19:08] as in, it has no change except not restarting containers on package upgrade? [19:08] the impact of the current packages is 0, it's a -security or -updates future issue? [19:10] Daviey: i dont' grok your last sentence? [19:11] as for 'future SRU', not to sound naive :) but I don't really expect any, especially for maverick (note it is as present the -unstable package) [19:11] hallyn: this bug doesn't cause upset with packages *currently* in -updates, right? [19:11] lxc is not in -updates yet, is that what you mean? [19:11] it's if the packages get upgraded in the future, the containers restart? [19:12] ah [19:12] yes, or on a do-release-upgrade i guess [19:12] your point makes sense; only if osmeone upgrades to natty from maverick would they see this [19:13] you sort of expect to reboot following a do-release-upgrade [19:13] infact the user is prompted to [19:13] natty might have more I suppose. [19:13] Daviey: do you have a good idea for how to make sure thi schange would make it into the next SRU? [19:14] hallyn: funny you say that... [19:14] if we could be sure UDD was going to be used, we could just commit it to bzr but not push a package [19:14] technically, staging it in the udd branch *should* be enough [19:14] Daviey: i'm a comedian, you just didn't know it [19:14] but nobody checks the udd branch before they upload their latest kick [19:14] eh, i do :) [19:14] we can ask stgraber and zul to. tha tshould cover it [19:14] KM0201: here is a article for quick samba permissions: http://www.cyberciti.biz/tips/how-do-i-set-permissions-to-samba-shares.html [19:14] -security do maintain a list of things they need to do when touching some packages. [19:14] where? [19:17] hallyn: good question.. trying to find it [19:18] Daviey: of course i still don't have write access to that tree. I don't mind if you just remove the release tag and push to {maverick,natty}-proposed [19:18] KillMeNow: thanms [19:18] *thanks [19:20] hallyn: if you still have the tree on your local machine, fancy changing the pocket to UNRELEASED and submitting a merge proposal? :) [19:21] Daviey: will do [19:21] (in about 10 mins) shoudl i assign to you? [19:21] assign the review to you, that is [19:21] the default should be good [19:22] ok, thanks. ttyl [19:24] i just upgrade the to the latest linux kernel on lucid and now I am getting this error [19:25] [ 402.925779] bonding: bond0: Warning: the permanent HWaddr of eth0 - 00:23:8b:11:22:10 - is still in use by bond0. Set the HWaddr of eth0 to a different address to avoid conflicts. [19:25] do I need to revert to the prev linux kernel ? [19:25] I did not see this alert before.. I have this bonding setup for a month now [19:26] Daviey: done [19:44] hallyn: -security have added an alert to include this fix if they touch the package [19:45] Subject: [ubuntu/oneiric] jenkins 1.409.1-0ubuntu1 (Accepted) [19:45] !!!!! [19:45] w00t [19:46] SpamapS: would be better if it wasin't in dep-wait :P [19:47] something get rejected or just many things in NEW? [19:47] NEW queue :( [19:47] Daviey: cool, thanks [19:58] hallyn: just more lxc shenanigans [19:58] hallyn: e.g. on oneiric cgroups-bin does something to make lxc startup fail [19:58] hallyn: https://code.launchpad.net/~serge-hallyn/ubuntu/natty/lxc/fix-restart/+merge/70064 .. is showing two commits, stgrabbers aswell? [19:58] stgraber rather [20:00] hallyn: Ahhh! [20:00] your target is the release pocket, not -proposed [20:00] Daviey, https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/791607, any chance we'll have a fix for A3? [20:00] Launchpad bug 791607 in eucalyptus "Oneiric Eucalyptus fails to start up" [High,Confirmed] [20:01] skaet: with no doubt, no. [20:01] skaet: That package should be in universe by then. [20:01] Daviey, have the mandatory tests been updated? [20:02] it's not something that will be tested as part of the A3 QA, skaet [20:02] skaet: yus [20:02] :) [20:02] :) [20:02] skaet: the A3 release will be an easy ride for us.. you read it here first! [20:03] If we come across difficult bugs, we just rip it off the cd.. that is how we roll. [20:03] Daviey, fingers crossed. [20:03] lol [20:04] skaet: bug #767225, is still of *major* concern to me tho. [20:04] Launchpad bug 767225 in fortune-mod "Slightly incorrect attribution in a fortune" [Wishlist,Incomplete] https://launchpad.net/bugs/767225 [20:04] LOL [20:05] :) [20:06] Daviey: I've done one of those before: https://bugs.launchpad.net/ubuntu/+source/fortune-mod/+bug/184824 [20:06] Launchpad bug 184824 in fortune-mod "Oxford Union rules out of date" [Undecided,Fix released] [20:07] There's a couple of others that are still open reported with A2, any of them likely to get fixed? [20:07] jcastro: Looks like it was resolved promptly, but this Oneiric release critical issue is still unresolved. :( [20:13] * skaet notes that Daviey considers a wishlist bug as release critical - hmm... could make for a rather long list for next agenda [20:21] lifeless: are you sure you are fully uptodate? (bc this weekend you reported a bug that i'd already worked around in the lxc package in archive) [20:24] Daviey: you want me to try an dtarget the merge request to -proposed? (not sure it will let me as that doesn't exist right now) [20:26] hallyn: yes, I'm sure. [20:26] hallyn: the workaround you have doesn't work. [20:27] oh? It worked on my system... [20:27] but that's not what you're asking me about right now - what's gong on with cgroups-bin? [20:27] your adding of ubuntu-keyring to the install line? doesn't work if apt decides to do two runs, because its a recommends not a depends [20:27] so it may work. [20:27] It may not. [20:27] hm [20:28] the right fix (just discussed in ubuntu-devel now) is changing apt. [20:28] ok. (regardless it still would have needed to have a *real* fix, for non-lxc users) [20:28] well, I don't know whats up with cgroups-bin, on natty it didn't /stop/ lxc working, now it does. I haven't got a cause yet. [20:29] can you list any variables from when it worked to now? [20:29] it shows up as 'failure to remove previous cgroup' which appears to be a generic failure mode when something is messing with new namespaces [20:29] with what commands? [20:29] lxc-start -n foo [20:29] lemme try to reproduce on ec2 [20:30] I have a clean oneiric vm; installed lxc, cgroups-bin, changed the conf file for it to be all ../cpu (carrying over the workaround from natty) [20:30] and yeah, ENOTWORKY [20:31] what does 'all ../cpu' mean? [20:32] (vaio overheated. grr) [20:40] hallyn: https://dev.launchpad.net/Running/LXC step 3 [20:42] lifeless: step 3 shouldn't be needed in oneiric... [20:42] lifeless: if you add freezer to that list, does it work? [20:43] will try later, have to run, sorry. [20:43] I'll try skipping step 3 first in fact. [20:46] New bug: #819507 in cloud-init (main) "consume_userdata is only called once per instance" [Undecided,New] https://launchpad.net/bugs/819507 [20:55] lifeless: yeah, i can reproduce with your step 3, not without. [20:55] lifeless: ttyl :) [20:55] hallyn: awesome! thanks. [21:04] is there somewhere i can look to see if slapd is compiled with tcp wrappers support or not? [21:05] for 10.04 [21:05] other than getting the source and digging thru it [21:09] r4__ digging through debian/rules [21:12] thanks [21:12] looking === ng_ is now known as zz_ng_ [21:21] does anybody know a ncurses app to manage users and groups? [21:52] New bug: #819535 in mysql-5.1 (main) "mysqldumpslow looking for slow.log file on an incorrect directory" [Undecided,New] https://launchpad.net/bugs/819535 [21:59] how can i mount a folder from the host into an lxc container? seems like simply adding a bind mount to the container's fstab is not possible. [22:02] lifeless, ever tried that? ^^ [22:23] New bug: #810731 in mysql-5.1 (main) "MySQL Server installation crashed, now unable to remove MySQL server and Client Binaries" [Undecided,New] https://launchpad.net/bugs/810731 [22:30] can apt-get -o pass CFLAG options? [22:30] it seems so === vDubG__Gone is now known as vDubG === davros- is now known as davros [23:50] I wonder why neither cPanel or Plesk support Ubuntu Server :P [23:53] bkerensa: why do you want them? [23:53] JanC: Clients want panel :P you know how it goes ;) [23:55] well, I have seens hosting providers offering cpanel & plesk on Ubuntu [23:56] but I guess the issue is that in general, Debian/Ubuntu have more strict rules about how apps can change config files etc., and those panels often don't obey those rules, which might cause problems [23:59] and where Ubuntu Server is used in hosting environments, plesk and cpanel are not capable enough [23:59] Hello all...