[00:08] <axisys> _KaszpiR_: it is wired link on a server
[00:09] <axisys> _KaszpiR_: where should I look to troubleshoot? planning to enable the tac_plus debug
[00:09] <_KaszpiR_> first, check your cabling
[00:10] <_KaszpiR_> like another network card, different switch port
[00:10] <_KaszpiR_> replace cable
[00:11] <axisys> _KaszpiR_: can you please tell me why ? (learning)
[00:13] <_KaszpiR_> what is the length of the network cable between switch and the server?
[00:13] <_KaszpiR_> is if a full duplex or half duplex switch, or maybe hub (ugh...)
[00:14] <axisys> its a cat3528 switch in short distance
[00:14] <_KaszpiR_> define short distance
[00:14] <axisys> Duplex: Half
[00:14] <axisys> yikes!
[00:15] <Patrickdk> not quarter?
[00:15] <axisys> _KaszpiR_: you are right on money.. thanks a lot
[00:15] <axisys> _KaszpiR_: I will start monitoring this
[00:15] <_KaszpiR_> replace cable, the first guess
[00:16] <axisys> _KaszpiR_: thanks
[00:16] <_KaszpiR_> or check the plugs, sometimes they are getting loosened up or the connection may be dirty etc
[00:16] <_KaszpiR_> worse scenario when one of the cables is wrong or using poor quality cable
[00:17] <_KaszpiR_> get some simple network cable tester for that
[00:18] <axisys> ethtool eth1 shows half duplex.. dmesg | grep eth1 also shows it.. is there any other logfile that should have that info?
[00:18] <axisys> _KaszpiR_: ^
[00:18] <_KaszpiR_> i doubt it
[00:18] <axisys> [   12.482613] igb: eth1 NIC Link is Up 100 Mbps Half Duplex, Flow Control: None
[00:18] <axisys> that was from dmesg
[00:18] <axisys> i wish I can tell when it switched to half
[00:19] <axisys> i guess this will work
[00:19] <axisys> # grep eth1 /var/log/dmesg
[00:19] <axisys> that shows it too
[00:19] <axisys> but I need to a timestamp
[00:19] <axisys> or depend on the monitoring tool to catch it
[00:20] <_KaszpiR_> you may try to change settings with the ethtool
[00:20] <_KaszpiR_> but usually forcing certian settings is just dangerous or performance degrading
[00:26] <axisys> _KaszpiR_: yeah, I am planning to do it.. may be when traffic is low :-)
[00:26] <axisys> we have a scheduled reboot coming up on Sunday
[00:27] <axisys> how do people usually handle ubuntu server security and kernel updates?
[00:27] <axisys> other distro seems to be way behind and not much reboots..
[00:27] <axisys> wish ksplice were available for all and not bought by oracle
[00:28] <axisys> we reboot the ubuntu server more often than others to stay on top those kernel and security upgrade
[00:58] <smoser> hallyn, just to be sure, the patch doesn't actually do anything
[00:58] <smoser> right?
[01:02] <_KaszpiR_> axisys ubuntu is imo the most often changed distro
[01:02] <_KaszpiR_> sometimes it is not a good thing
[01:02] <_KaszpiR_> on the other thing noone forces you to make reboots ;)
[01:03] <_KaszpiR_> another story - some time ago I did kernel update (with distro update) and after reboot I was surprised that network card was no longer working
[01:03] <_KaszpiR_> they moved drivers to nonfree
[01:03] <_KaszpiR_> guess my enjoyment
[01:13] <mdeslaur_> axisys: if you're comfortable with other distros updating the kernel less frequently, you can simply reboot your Ubuntu less frequently...
[01:24] <Tohuw> Is there some better solution to the permissions conundrum WebDAV creates than setting the Apache umask to 000?
[01:41] <ignarps> resolvconf -u doesn't seem to be updating my resolv.conf information from the settings in /etc/network/interfaces.
[01:42] <ignarps> Does anyone know how to get resolvconf to reread the entries and generate a new resolv.conf file ?
[01:52] <stgraber> hallyn: patched kernel + new ifupdown => no more network-interface-* entries left after the container is killed!
[02:05] <ignarps> I have verified the /etc/resolv.conf is the proper symlink to /run/resolvconf/resolv.conf and I have checked that updates-enabled is set
[02:05] <ignarps> I have tried to restart resolvconf via /etc/init.d/resolvconf restart;  tried the resolvconf -u that the manual shows
[02:06] <ignarps> but my /etc/resolv.conf file won't update the dns list from /etc/network/interfaces
[02:10] <hallyn> smoser: right
[02:10] <hallyn> just installs a sample sysctl.conf file to show *how* to do it
[02:10] <hallyn> stgraber: \o/
[02:12] <hallyn> heh, s/blow/below/
[02:14] <axisys> security updates may sometime need kernel upgrade or libc or openssl upgrade .. what's the workaround ?
[02:14] <axisys> in other words how to keep the system security uptodate w/o reboot? looking for suggestions
[02:15] <axisys> I guess looking for best practice
[03:07] <pmatulis> axisys: you need to reboot to get the new kernel obviously
[03:58] <axisys> pmatulis: right.. I do not want to upgrade new kernel unless it fixes security issue
[04:00] <axisys> so on unattended-upgrade just uncomment the security line is all I needed. and if that means kernel upgrade sometimes then be it..
[04:00] <axisys> I am on right track ?
[04:00] <axisys> am I ?
[04:28] <oneseventeen> any tips on making SSH stay open longer?  one site said change the ClientAliveInterval but I don't see that in the sshd_config file
[04:29] <escott> oneseventeen,  ControlMaster auto  ControlPath /tmp/ssh_%r@%h:%p  ControlPersist 4h
[04:29] <escott> oneseventeen, but on the server you have to make some modifications
[04:31] <bradm> oneseventeen: you could possibly try TCPKeepAlive and ServerAliveInterval as well
[04:31] <oneseventeen> TCPKeepAlive is already yes, but I don't see a ServerAliveInterval... should I just add that to sshd_config
[04:32] <oneseventeen> escott: thanks, I'll do some searching on ControlMaster/Path/Persist
[04:32] <bradm> oneseventeen: its a client side thing
[04:33] <oneseventeen> ahh... I'm working with some 3rd party clients, I'll see what they have in the way of options.
[04:35] <bradm> oneseventeen: ClientAliveInterval is definately the server side version of it, and I see it in the sshd_config man page
[04:35] <bradm> oneseventeen: you can tweak it a bit more with ClientAliveCountMax as well
[06:14] <Lartza> VPS with 3GB of RAM, 64-bit or 32-bit? I ran it 32-bit before but not sure if 64 would have any benefits
[06:14] <Lartza> OpenVZ
[06:15] <demona> Rule of thumb was always, 32-bit unless you have more than 4Gb RAM
[06:16] <demona> whether virtual or physical server
[06:16] <Lartza> Yeah I've thought to too, just never really thought about it :)
[06:16] <Lartza> *so too
[06:16] <Lartza> Bad english... not a native speaker :P
[06:16] <demona> English is bad all over
[06:16] <Lartza> That is too true ;)
[07:32] <AnAnt> Hello, I am trying to install Ubuntu precise on a Dell PowerEdge T410 server, the installer is asking me for disk driver, can anyone help ?
[07:39] <RoyK> AnAnt: what sort of disk controller does that thing have?
[07:39] <RoyK> AnAnt: and which ubuntu version is this?
[07:39] <AnAnt> Ubuntu Precise
[07:40] <AnAnt> RAID controller
[07:40] <AnAnt> PERC600 I thnk
[07:40] <RoyK> what usually happens, is something new arrives and re-uses the old chipset, but with a new PCI ID, so that the driver doesn't find it
[07:41] <RoyK> and btw, Dell normally supports RHEL and SuSE and that's about it... The custom drivers they might ship will probably *not* load properly on ubuntu
[07:43] <Kartagis> anyone can help me with postfix?
[07:43] <RoyK> !ask
[07:44] <AnAnt> RoyK: so why is PowerEdge T410 a certified machine on Ubuntu website ?
[07:44] <Kartagis> http://daf2e3092688b173.paste.se/
[07:44] <Kartagis> I get these errors
[07:44] <AnAnt> http://www.ubuntu.com/certification/hardware/201011-6780/
[07:44] <RoyK> AnAnt: erm - it is?
[07:44] <Kartagis> my mysql server is running, and the credentials are correct (I've checked)
[07:45] <RoyK> Kartagis: then I guess there's a new version of the hardware, with a new PCI ID - that happens - it shouldn't, but it may
[07:46] <Kartagis> erm. you mean AnAnt, right?
[07:46] <AnAnt> or the machine I got has a different storage
[07:46] <Kartagis> new version of the hardware? /me is confused
[07:46] <RoyK> anant: contact ubuntu/canonical support. if you have paid support for it, it'll be fixed quickly - otherwise it may take some time. Also, you might want to try to boot on Quantal
[07:46] <AnAnt> the website says Samsaung
[07:47] <RoyK> AnAnt: right - that doesn't list the PERC
[07:47] <RoyK> try Quantal, please
[07:47] <RoyK> probably new chipset or just pci id
[07:47] <AnAnt> hmm
[07:47] <RoyK> IIRC the PERC60 is just an LSI SAS2 controller
[07:47] <RoyK> but different PCI ID
[07:47] <RoyK> as usual
[07:55] <RoyK> AnAnt: if it works with quantal, you should be able to hack up a custom precise boot cd or pendrive with the quantal kernel (so to avoid upgrading the whole thing to something non-LTS)
[07:55] <RoyK> AnAnt: btw, is this a new T410?
[07:56] <AnAnt> RoyK: erm, we just bought it, yes
[07:56] <RoyK> please try with Quantal - probably new pci id with old, working chips
[07:57] <RoyK> if you have console access (with networking), try lspci -vvv
[07:57] <Kartagis> RoyK: can you help me with that?
[07:58] <RoyK> Kartagis: looks like your pfuser mysql user has an issue connecting to its database - fix that
[07:58] <RoyK> AnAnt: and lspci -vvvn
[07:59] <RoyK> AnAnt: that'll list the PCI IDs associated with each PCI (express) unit in the system (including those soldered to the motherboard)
[07:59] <Kartagis> RoyK: I wouldn't know where to start, since credentials are correct (I've checked)
[07:59] <AnAnt> RoyK: I can run lspci -vvnn , since I am running the text installer
[07:59] <RoyK> AnAnt: ctrl+alt+f2
[07:59] <AnAnt> yes
[07:59] <RoyK> there should be a console there
[08:00] <RoyK> well, try quantal first
[08:05] <AnAnt> RoyK: ok, how could I know which PCI unit is the disk controller ?
[08:12] <RoyK> AnAnt: http://paste.ubuntu.com/1427051/
[08:14] <RoyK> http://www.pcidatabase.com/
[08:14] <RoyK> so with my 1849:8168, 1849 is vendor id, 8168 is device id
[08:21] <AnAnt> RoyK: ok, how could I know which PCI unit is the disk controller ?
[08:21] <AnAnt> RoyK: funny thing is that megaraid_sas driver is indeed loaded !
[08:21] <AnAnt> yet it still asks for a disk driver
[08:22] <RoyK> Quantal does?
[08:23] <RoyK> AnAnt: have you tested Quantal (12.10)?
[08:23] <AnAnt> not yet
[08:24] <RoyK> https://help.ubuntu.com/community/InstallCDCustomization
[08:25] <AnAnt> I thought of running lspci before rebooting
[08:25] <RoyK> did you do so?
[08:25] <AnAnt> yup
[08:25] <RoyK> if you can find the PCI ID of that thing...
[08:25] <AnAnt> but I didn't find out which is the disk controller !
[08:26] <RoyK> run lspci -v first
[08:26] <RoyK> without -n
[08:26] <AnAnt> I just ran lspci
[08:26] <RoyK> that should show you its position - in my example, it's 04:00.0
[08:27] <RoyK> does that show something like RAID controller?
[08:27] <RoyK> or SAS
[08:28] <RoyK> well, forget it - try quantal
[08:28] <RoyK> if it works, it works
[08:28] <RoyK> then you can use a quantal kernel in a custom setup somehow
[08:28] <AnAnt> wellI can only see SATA controllers
[08:29] <AnAnt> isn't it little bit dangerous to mix stuff from different releases, especially the kernel
[08:30] <RoyK> the kernel is usually the least dangerous - the worst is all the userspace stuff
[08:30] <AnAnt> ah, there is the SAS controller
[08:30] <RoyK> then lspci -n should show its pci id
[08:32] <RoyK> perhaps try a kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.6.3-quantal/
[08:32] <RoyK> I can check in 3.6.3 if that pci id is listed there
[08:33] <RoyK> or - well - try quantal - if you can live with a non-lts build for another 18 months...
[08:33] <RoyK> or 16
[08:33] <AnAnt> I am trying to avoid non-LTS
[08:33] <RoyK> so am I
[08:34] <RoyK> just saying it's a wee bit tricky to install on something not supported by the kernel
[08:34] <AnAnt> ok, so the PCI ID is 0104: 1000:0079
[08:35] <AnAnt> where shall I go from there ?
[08:35] <RoyK> try quantal
[08:35] <RoyK> to see if it's in that kernel
[08:35] <RoyK> then we can see where to go from there
[08:35] <AnAnt> funny thing is that lspci says that this controller is indeed using megaraid_sas module
[08:35] <RoyK> lspci doesn't say anything about kernel modules
[08:35] <RoyK> but 1000 is indeed LSI
[08:36] <AnAnt> lspci -v -s 05:00.0 did give this output at the end:
[08:36] <RoyK> standard SAS2 controller
[08:36] <AnAnt> Kernel driver in use: megaraid_sas
[08:36] <AnAnt> Kernel modules: megaraid_sas
[08:37] <RoyK> please
[08:37] <RoyK> the PCI IDs are hardcoded in the driver
[08:37] <RoyK> when a new PCI ID is introduced, like here, 0097 is not listed in the driver
[08:37] <RoyK> meaning megaraid_sas won't load
[08:37] <RoyK> or perhaps it'll load, but it won't find the card
[08:37] <RoyK> s/card/controller/
[08:38] <AnAnt> well, the install log did list some stuff from megaraid
[08:39] <AnAnt> last line is: megasas: INIT adapter done
[08:39]  * RoyK gives up and goes to do something better
[08:48] <AnAnt> find, I'm trying quantal now
[08:55] <RoyK> AnAnt: I guess no luck?
[08:57] <AnAnt> RoyK: same problem with quantal
[08:57] <RoyK> yeah, I checked the source up to linux 3.7, and that PCI ID is not there
[08:57] <RoyK> however
[08:57] <RoyK> http://www.mjmwired.net/kernel/Documentation/PCI/pci.txt
[08:58] <RoyK> it seems it's possible to hot-add PCI IDs
[08:58] <RoyK> I didn't know that
[08:58] <RoyK> New PCI IDs may be added to a device driver pci_ids table at runtime
[08:58] <RoyK> echo "vendor device subvendor subdevice class class_mask driver_data"> /sys/bus/pci/drivers/{driver}/new_id
[08:59] <RoyK> though YMMV
[08:59] <RoyK> I would consider using rhel/centos with the drivers shipped with the server (on a CD/DVD/something)
[09:00] <RoyK> swearing in church, sure, but it's easier that way
[09:07] <AnAnt> there is no /sys/bus/drivers during install
[09:07] <AnAnt> oops
[09:08] <RoyK> you'll probably want to load that driver first too
[09:09] <RoyK> although I don't know how to fix this so it happens automatically after a reboot - it'll have to happen during kernel bootup or somewhere in the initrd
[09:09] <RoyK> since nothing can be mounted before the driver is loaded
[09:10] <AnAnt> funny thing is that the driver is already loaded !
[09:10] <RoyK> well, that's good
[09:10] <AnAnt> class_mask ?
[09:12] <AnAnt> it's optional anyways
[09:13] <RoyK> never tried that before, so no idea...
[09:13] <AnAnt> still asking me for disk driver !
[09:13] <RoyK> cat /proc/partitions
[09:13] <RoyK> anything shows up?
[09:13] <AnAnt> sr0
[09:22] <RoyK> AnAnt: then no idea - sorry
[09:22] <RoyK> AnAnt: what was the line you gave the driver?
[09:22] <AnAnt> the echo line ?
[09:22] <RoyK> yeah
[09:24] <RoyK> TBH, what I'd do if I were you in this, was testing centos and shipped drivers from dell
[09:24] <RoyK> centos is a PITA, but it works
[09:24] <AnAnt> echo "1000 0079 1028 1f16 0104" > /sys/bus/pci/drivers/megaraid_sas/new_id
[09:25] <RoyK> AnAnt: any output in dmesg after that?
[09:28] <AnAnt> no
[09:29] <RoyK> AnAnt: sorry - really no idea
[09:29] <AnAnt> thanks
[09:34] <RoyK> AnAnt: you may want to try to ask in #kernelnewbies @ irc.oftc.net
[09:37] <RoyK> AnAnt: but really, just try centos
[09:44] <AnAnt> RoyK: I'd rather give the server back !
[09:53] <RoyK> AnAnt: well, if you can... There should be other servers around, or perhaps if you just got a PCI-ex controller, or perhaps even better, don't use "hardware" raid, you probably have sufficient hardware to run that part in software ;)
[09:53] <RoyK> "hardware" raid is just software after all, it's just hidden
[09:54] <RoyK> better let linux do that job - it's usually faster and just as safe
[10:27] <AnAnt> RoyK: sorry for troubling you, I found out that I had to configure the RAID before installing the OS !
[10:27] <AnAnt> I mean adding logical volumes
[10:28] <RoyK> ROTFL
[10:28] <AnAnt> it's first time I got a machine with RAID
[10:28] <AnAnt> we didn't need the RAID actually, it just came with the machine
[10:58] <jamespage> ivoks, around? I have a corosync question
[11:16] <ivoks> jamespage: yep
[11:26] <Daviey> jamespage: DO you happen to know what wordpress plugin we use for 'rebloggig' ?
[11:26] <jamespage> Daviey, sorry - no
[11:27] <jamespage> ivoks, I'm having trouble restarting corosync on quantal - but I can't figure out why - the process is just not responding to QUIT
[11:27] <jamespage> ivoks, how can I diagnose further
[11:27] <jamespage> ?
[11:27] <ivoks> jamespage: are you sure it just doesn't take very long time?
[11:27] <ivoks> jamespage: corosync spawns lots of other processes
[11:28] <jamespage> ivoks, hmm - might be
[11:28] <jamespage> let me check
[11:28] <ivoks> jamespage: like crmd, lrmd, stonithd, etc... and then, depending on cluster configuration, it can take unlimited amount of time to stop it
[11:28] <jamespage> maybe I was just being impatient!
[11:28] <ivoks> jamespage: ideally, if nothing is set up, it takes a minute to get it to full start and then 10-15 seconds to stop it
[11:28] <ivoks> meaning, after start it can't quit for next 60+ seconds
[11:29] <jamespage> ivoks, ah - I see
[11:29] <jamespage> ivoks, I see pacemaker running lots of those things
[11:29] <ivoks> tbh, i haven't looked at pacemaker in quantal yet
[11:29] <ivoks> i'm sure, once again, everything was redesigned :D
[11:30] <jamespage> ivoks, ah - I see why "Dec 12 11:28:44 juju-openstack-lcy01-instance-10 corosync[987]:   [pcmk  ] notice: pcmk_shutdown: Preventing Corosync shutdown.  Please ensure Pacemaker is stopped first.
[11:30] <jamespage> "
[11:30] <ivoks> in the old days, corosync would start pacemaker
[11:30] <ivoks> right
[11:30] <ivoks> i think now pacemaker has its own init script
[11:30] <ivoks> and one needs to stop it before corosync
[11:30] <jamespage> ivoks, it does but its not enabled by default
[11:30] <jamespage> ...
[11:31] <ivoks> look at /etc/corosync/corosync.conf
[11:31] <ivoks> if it has pcmk service configured, corosync starts pacemaker
[11:31] <ivoks> grep -sr pcmk /etc/corosync
[11:32] <jamespage> ivoks, nothing - I think roaksoax did that on purpose
[11:32] <ivoks> :)
[11:33] <ivoks> if that machine is available, i could check it out
[13:03]  * cloud_away is now away - Reason : taking the dog out
[13:13] <pmatulis> axisys: so don't upgrade the kernel then.  or upgrade it but don't reboot
[13:26] <Daviey> zul: btw, CA for grizzly is open
[13:34] <zul> Daviey:  yay!
[13:40] <Daviey> adam_g: don't suppose you are around yet? :)
[13:50] <zul> Daviey: its still like 5:50 am for him
[13:51] <Daviey> zul: I know, he's slack. :)
[13:51] <zul> totally
[13:51] <Daviey> We were up jogging at this time,.
[13:52] <koolhead17> Daviey: zul i thought adam_g is in Japan these days :P
[13:53] <zul> Daviey: some of us who presumably had bears chasing us because thats the only reason to jog
[13:54] <allballs> Good morning. I've been running an Ubuntu Server 10.04 instance on EC2 for about 8 months now. I would like to upgrade it to 12.04-LTS. Is there any recommended "EC2" way of doing this, or, should I just do the standard ol'fashioned do-release-upgrade?
[13:54] <MrTorque> i got degraded raid-status just after creating the raid 1 and adding a partition. Is this because I did simultaniously create 3 raid 1 (each consisting of 2 hdds)?
[13:55] <xnox> MrTorque: did they fully sync after creating?
[13:55] <xnox> or is that what is happening now?
[13:57] <simplexio> MrTorque: i think it works like it should be . i mean there is no mirrored data on new partition
[13:58] <simplexio> MrTorque: so it has to rebuild
[14:04] <smoser> hallyn, ping
[14:05] <MrTorque> xnox: they are right now mirroring
[14:06] <MrTorque> and while the first raid was about 1% complete i did create the next raid and another one. then i got 3 system mails that the raid is degraded...
[14:06] <MrTorque> simplexio: so every raid is degraded after creation?
[14:07] <simplexio> i think so . i can be wrong too
[14:11] <Teduardo> Howdy friends, I'm trying to install 12.10 on a Supermicro machine that has the Intel Rapid Storage Enterprise SCU, the installer installs fine on the raid but then when the server reboots it doesnt seem to be finding grub.
[14:12] <smoser> hallyn, ok. so when you see this, 2 things.
[14:14] <smoser> a.) am i right that if i do 'lxc-start -n foo' (without '-d') then i can never detach from the container
[14:14] <smoser> b.) could you add the escape/detach sequence (ctrl-a q) to 'man lxc-console'
[14:23] <hallyn> smoser: a. right
[14:23] <hallyn> smoser: b. of course.  i'll open a bug for it
[14:27] <YamakasY> Hi guys, is there a package where I can manage applications like dns, apache, vpns, etc per node? I have look at all and webmin seems to come closest to this but it can not manage per node/ per service..
[14:28] <smoser> hallyn, thanks. if you do 'b', you'll reduce your pingcount from me by ~ 1/month.
[14:28] <smoser> maybe document in lxc-start man page also that there is no way to detach ?
[14:31] <hallyn> smoser: i'll add that to the same bug
[14:33] <MrTorque> simplexio: hm, i will wait the 5h it takes to rebuild and see if the status changed.
[14:33] <MrTorque> thank you
[14:36] <simplexio> MrTorque: send msg if my answer wasnt correct, it is good to learn
[14:37] <MrTorque> I hope I won't forget :)
[14:38] <Teduardo> hmm
[14:38] <Teduardo> wonder why this thing wont boot =)
[14:38] <Teduardo> I hate supermicro =D
[14:40] <Teduardo> is there any easy way to boot with a live cd or something and check grub?
[14:57] <Daviey> yolanda: Your karama is going to go through the rood :)
[14:58] <yolanda> yes, that quantum SRU generated more than 100 tasks!
[15:29] <Pici> "|
[15:37] <Daviey> thanks for the update uvirtbot
[15:40] <yolanda> !
[15:47] <Pici> soren: uvirtbot is going a bit nuts, ideas?
[16:33] <halvors> I get an error from dovecot, and can't figure out how to fix it: http://paste.ubuntu.com/1427860/
[16:34] <halvors> May anyone update the roundcibe package to the latest version?  0.8.4 http://www.roundcube.net/
[16:41] <halvors> Nobody knows how to fix my issue?
[16:48] <rbasak> soren: uvirtbot spam ^^  - any chance you could shut it up, please?
[17:06] <adam_g> Daviey: sup
[17:27] <Darkstar1> Hello everyone.
[17:27] <RoyK> evening
[17:30] <Darkstar1> Just a qq. The php max upload filesize is set at 64M I can't attach anything bigger than 8Mb to an email on the when trying to send
[17:30] <Darkstar1> Anyone know what else could be the limiting factor?
[17:37] <halvors> I get an error from dovecot, and can't figure out how to fix it: http://paste.ubuntu.com/1427860/
[17:37] <halvors> Is the way to go to just make dir /var/www/Maildir and chown and chgrp to www-data?
[17:39]  * RoyK guesses apparmor may be in the way
[17:41] <rbasak> Why use /var/www/Maildir in the first place? I think a good answer depends on the answer to this question.
[17:42] <rbasak> Or is this a dovecot thing?
[17:54] <zul> adam_g: did you push the quantal sru into the CA
[17:57] <adam_g> zul: no, i wasn't planning on putting any of them there until they get into quantal-proposed
[17:57] <adam_g> back in 30
[17:58] <zul> adam_g: k
[19:12] <adam_g> zul: did you get a bug # from jay's keystoneclient issue yesterday?
[19:13] <zul> adam_g: yeah i asked him to open up a bug against python-keystone but i wasnt able to reproduce this myself
[19:14] <adam_g> zul: did one get opened? curious about it but cant find one
[19:14] <zul> hld on
[19:15] <zul> https://bugs.launchpad.net/ubuntu/+source/python-keystoneclient/+bug/1089194
[19:20] <bieb> I installed 12.04 server, I need gui for one of the people that will be using server.. I did apt-get install xinit, when I startx I get a small white box in the upper left with a command prompt.. any ideas?
[19:22] <sarnold> you might also want to install a window manager
[19:29] <RoyK> bieb: apt-get install ubuntu-desktop
[19:29] <RoyK> just make sure network-manager* isn't installed - it'll mess up networking - binding it to the gui
[19:30] <RoyK> you can easily uninstall it later, though I don't think ubuntu-desktop includes that
[19:32] <bieb> RoyK: does that give teh same desktop as a regular 12.04 install?
[19:34] <bieb> RoyK: unity.. I couldnt remember the name of it
[19:38] <RoyK> bieb: yes
[19:38] <RoyK> it does
[19:38] <RoyK> of course, you can change that later
[19:38] <bieb> ehhh.. not a huge fan of unity
[19:39] <RoyK> but it installs whatever needed for a good desktop
[19:39] <bieb> can you install the ubuntu fallback?
[19:39] <RoyK> I don't use X on servers
[19:39] <RoyK> someone else might know
[19:39] <bieb> ohh.. that would be a good reason to use the ubuntu desktop.. to give a good desktop
[19:39] <sarnold> bieb: was that 'gnome-panel'?
[19:40] <bieb> sarnold: not sure what the old ubuntu gui was.. I thought you could apt get ubuntu fallback.. I will have to look that up
[19:42] <sarnold> heh, apt-cache search ubuntu fallback --> 0 results
[19:44] <bieb> sarnold: gnome-session-fallback
[19:45] <sarnold> bieb: nice. :D
[19:56] <bladernr_> can someone give me a little basic info about the Cloud live image?
[19:57] <bladernr_> Following the instructions here: https://help.ubuntu.com/community/UbuntuCloudInfrastructure#Use_Your_Cloud I am told to go to http://node-address/horizon and log in.  that returns a 404 error
[19:57] <bladernr_> if I go to http://node-address (presuming that's actually the correct way), I DO get a login prompt
[19:57] <bladernr_> however, the suggested UID/PW of admin/openstack does not work.
[20:00] <Daviey> adam_g: bug 1029430 .. I can't quite grok what kernel/userspace combo this impacts. Is it clear to you?
[20:00]  * Daviey wants a truth table :)
[20:03] <adam_g> Daviey: yes. kernel 3.3. made vhost_net enabled by default. when its enabled, virtualized enviornments where a dhcp server is running on the host and serving to guests on the same machine (nova's multi_host, libvirt's default network) wont work without some extra IPtables rules. libvirt takes care of setting this up for its default network (and i believe other networks it manages), but other virt users (nova, for instance) that do not make us
[20:04] <sarnold> adam_g: but off at 'make us'
[20:04] <sarnold> s/but/cut/
[20:04] <Neal_> Can I resize my partitions without rebooting the server?
[20:05] <adam_g> sarnold: huh?
[20:05] <sarnold> adam_g: you wrote too much in one line, you got cut off at "that do not make us"
[20:06] <adam_g> oh
[20:06] <adam_g>  other virt users (nova, for instance) that do not make use of libvirt for that  need to do so manually
[20:09] <Daviey> adam_g: Ah, that is the issue you were working on!
[20:09] <Daviey> makes sense now
[20:09] <escott> Neal_, if you can umount them
[20:09] <Neal_> escott: it's the main partition
[20:10] <adam_g> Daviey: i imagine there are build options to avoid the change, tho
[20:10] <escott> Neal_, then no. not with most normal filesystems
[20:10] <sarnold> Neal_: resize2fs manpage says you can expand ext3 and ext4 without having to unmount the filesystem first
[20:11] <sarnold> Neal_: whether you trust it is another matter :D
[20:11] <Neal_> escott: I have 1.8TB on /home and only 10GB on / and I want / to have 100GB instead, but I can't take the server offline.
[20:11] <Daviey> adam_g: but it's indicative of potential unnoticed regressions
[20:11] <Neal_> sarnold: will look into it, thanks
[20:14] <chilicuil> hi, good morning, I'm trying to setup bonding in my ubuntu machine, I use wicd as network manager, I've following wiki to get an bond0 device, https://help.ubuntu.com/community/UbuntuBonding , it works to set it up, however when I use wicd-curses to stablish a conection with 1 of the slave devices it pushes wlan0 out of the slaves interfaces, should I stop using wicd bonding?
[20:15] <RoyK> chilicuil: what are you trying to accomplish?
[20:16] <chilicuil> RoyK: to use 2 wireless connections as 1 in order to obtain as much assurance in the conection as possible
[20:16] <RoyK> same access point?
[20:17] <chilicuil> RoyK: no
[20:17] <RoyK> same L3 network?
[20:17] <chilicuil> RoyK: no
[20:17] <RoyK> then you'll need a routing protocol of sorts
[20:17] <chilicuil> they're 2 different isp
[20:17] <sarnold> isn't bonding intended more for "doubling speed" than "doubling reliability"? sort of like a striped RAID rather than mirrored RAID?
[20:18] <RoyK> sarnold: bonding is a lot of different things
[20:18] <sarnold> RoyK: hehe, figures :)
[20:18] <RoyK> sarnold: with LACP, you can have redundancy and higher bandwidth, with master/slave setup, you only get redundancy etc
[20:18] <RoyK> read up about it
[20:18] <sarnold> RoyK: thanks! :D
[20:19] <chilicuil> RoyK: ok, I'll look at it, thanks for your time
[20:54] <hallyn> stgraber: so lxc-attach works (at leaset without userns) with the userns kernel...  but apparmor profile doesn't get changed of course.  which is sort of dangerous.  just something to heed
[20:55] <stgraber> hallyn: hmm, indeed. I guess we'll want lxc to change to the container's apparmor profile too then
[20:56] <hallyn> stgraber: probably safest with a lxc-attach-specific profile which allows bypassing lxc-start on the way to the container profile
[20:56] <hallyn> sigh
[20:56] <hallyn> somehow didn't see thta one coming
[20:57] <hallyn> yeah and attach to userns does nto yet work
[20:57] <stgraber> see, we still have lots of things to do for alpha2 ;)
[20:57] <stgraber> (upstream lxc alpha2 that's)
[20:57] <hallyn> well thats a kernel bug
[20:58] <hallyn>  / missing feature
[20:58] <hallyn> u guess ill file a bug about the profile switch
[20:58] <hallyn> sigh
[20:58] <hallyn> s/^u/i/
[21:01] <jcastro_> smoser: new stable-apt fast release, feels more transparent than bolted on now.
[21:01] <jcastro_> smoser: it's pretty insanely fast with aws mirrors
[21:01] <smoser> ?
[21:01] <smoser> link?
[21:02] <jcastro_> http://www.iloveubuntu.net/apt-fast-181-released-aria2c-default-and-under-hood-improvements
[21:05] <jcastro_> smoser: I put like three regions in the mirror= section and cranked up the connections to 35 and it was real fast.
[21:05] <jcastro_> smoser: I have no science for you though, just anecdotes and possibly misleading localized performance.
[21:05] <jcastro_> :)
[21:25] <smoser> jcastro_, can i 'apt-get update' somehow ?
[21:26] <jcastro_> apt-fast update
[23:21] <adam_g> Daviey: still about?
[23:51] <Daviey> adam_g: yah
[23:52] <adam_g> Daviey: http://people.canonical.com/~agandelman/nova_2012.2-0ubuntu5.2~cloud0/
[23:53] <adam_g> Daviey: just hit quantal-security. thats for ca. also, theres the 2012.2.1 update thats stewing in queue for quantal-proposed, that doesn't include that security update. should i just trump that with a new stable/folsom snapshot that includes this cve?
[23:56] <Daviey> adam_g: yeah, the copy from quantal-proposed to updates could happen now
[23:56] <Daviey> So base on quantal-proposed, and we'll fast track it
[23:58] <keithzg> Hmm, is there any hope for qemu-kvm 1.2 to be backported to 12.04? Being unable to blockpull really puts a kink in external snapshotting . . .
[23:58] <adam_g> Daviey: not sure i follow that