[02:05] <trimeta> I just put together a box and am trying to boot Ubuntu onto it. I get to the Ubuntu boot menu ("Run Ubuntu," "Install Ubuntu," "Run first HDD," "Test Memory," etc.), but when I select Run Ubuntu, the screen goes dark and nothing happens.
[02:05] <trimeta> I decided to test the memory, but something weird happens: it shows the right amount of memory in the line right after the L3 Cache line, but the Settings: line is nearly empty, and it doesn't seem to be running any tests.
[02:06] <trimeta> And there's no options along the bottom for escaping or configuration.
[02:06] <trimeta> Esc and c do nothing.
[02:08] <trimeta> And now it's not even booting off the USB...
[02:11] <trimeta> Oh, wait, this is a Sandy Bridge chip. Could that be it?
[02:27] <Roasted> hey fellas
[02:28] <Roasted> I'm tr ying to figure out who is limiting me in this scenario. I'm transferring a 5.9GB file to my personal owncloud server on the LAN and it tanks @ 4.3GB on the nose each time. Who's limiting it? Apache? Webdav? PHP?
[02:29] <patdk-lap> who knows
[02:29] <patdk-lap> how are you transfering it?
[02:29] <Roasted> throguh webdav in nautilus
[02:29] <patdk-lap> not php then
[02:30] <patdk-lap> apache/webdav or nautilus
[02:30] <Roasted> kind of doubtful it's nautilus. I heard about a 4.3GB limit on google but some people were saying that was IIS related.
[02:30] <Roasted> Since I'm not running IIS I wasn't sure if that would imply apache or what.
[02:31] <Roasted> It doesn't help that I get an error after asying my system doesn't have enough memory to process the error to send to devs. It maxes uot all 8GB only to leave me to cancel it with no idea what happened.
[02:34] <patdk-lap> heh?
[02:34] <patdk-lap> apache doesn't do that
[02:34] <patdk-lap> so that would have to be on your computer, nautilus
[02:34] <Roasted> hm
[02:35] <Roasted> I'll try it again and watch my memory. I wonder if it scales up during the transfer or if it just spikes @ the end
[02:35] <Roasted> up up and away
[02:35] <Roasted> 50% memory
[02:35] <Roasted> and failed
[02:36] <Roasted> Error in stream protocol: End of stream
[03:57] <trimeta> It seems my grub won't continue booting until I hit "enter," despite GRUB_TIMEOUT being set in /etc/default/grub. What could be happening?
[04:03] <trimeta> Even on a clean reboot, it seems it's getting the failure thing that makes it hang at Grub.
[06:09] <branant> I am getting very slow performance running Drupal websites from my new Ubuntu Server 12.04 installation.
[06:10] <branant> It takes about 1-2 minutes to load up the website and after that everything works fine
[06:10] <branant> HostnameLookups is set to Off in the apache2.conf
[06:11] <branant> Any other ideas on what else I can check please?
[06:39] <taipres> anyone use ngnix in here?
[06:49] <ironm> taipres, it looks like there no deb packages for ngnix (in official repositories at least)
[06:49] <ironm> +are
[06:50] <ironm> taipres, it looks like there are no deb packages for ngnix (in official repositories at least)
[06:57] <taipres> someone needs to fix that
[06:57] <taipres> ngnix is awesome
[06:58] <taipres> almost have everything setup right, ngnix is easy mundin isn't
[07:14] <cwillu_at_work> !info nginx
[07:15] <cwillu_at_work> ironm, did I miss something?
[07:15] <cwillu_at_work> taipres, ^
[07:24] <ironm> thanks a lot cwillu_at_work. I have found deb packages in repositories on ngnix.org .. even von preicse :)
[07:24]  * cwillu_at_work points out that precise _has_ nginx debs in the repositories already
[07:28] <ironm> cwillu_at_work, well .. I have checked it on nginx .org  ... *sorry* .. there are also packages for debian's sid (I misspelled it using apt-cache search ngnix ... instead of nginx)
[07:39] <Guest87207> hi all
[07:40] <Guest87207> something strange has happened to one of our ubuntu servers.. it is no longer possible to su - root
[07:40] <Guest87207> but in the vmware console at the login prompt if you enter "root" as username you are immediately dumped to a prompt as if you had entered a password
[07:40] <Guest87207> but no password is requestest
[07:40] <Guest87207> any ideas?
[09:01] <elst> hi
[09:01] <elst> what is the link to download ubuntu server for network installation?
[09:01] <elst> for normal ubuntu lftp -c "open http://it.archive.ubuntu.com/ubuntu/dists/precise/main/installer-i386/current/images/; mirror netboot/"
[09:01] <elst> but for ubuntu server?
[09:07] <taipres> is posix the easiest to setup for mail?
[09:09] <taipres> postfix rather
[09:14] <yeats> taipres: postfix is very straightforward, yes
[09:45] <ironm> hello. Do I really need "ifenslave" for interface bonding (LinksAggregation) on ubuntu-server 12.04?
[09:55] <Daviey> SpamapS: Hey, is there any reason not to sync txaws?
[10:00] <Daviey> jamespage: Hola, how much effort would it be to generate a chart of the openstack ci utilisation?  Ie, be able to spot peak usage and idleness ?
[10:04] <ironm> does anyone of you run interface bonding on ubuntu 12.04?
[10:07] <jamespage> Daviey, there is a history view for each executor configured - gives you a rough idea
[10:07] <jamespage> but as its commit driven....
[10:07] <jamespage> varies massively
[10:25] <halvors> Hi! I'm trying to use the "net.ipv6.conf.all.forwarding=1" in sysctl.conf to give connectivity to my lan computers using the wan interface of my Ubuntu Router Box. But when i'm trying to ping from the inside, nothing happends...
[10:49] <Daviey> jamespage: I really want to try and determine if we can disable CI testing over the weekends.
[10:49] <Daviey> jamespage: if so, it would be nice to be able to hammer time longevity every weekend.
[10:55] <halvors> Hi! I'm trying to use the "net.ipv6.conf.all.forwarding=1" in sysctl.conf to give connectivity to my lan computers using the wan interface of my Ubuntu Router Box. But when i'm trying to ping from the inside, nothing happends...
[11:24] <soren> Say I have an interface that isn't configured for VLAN's at all.. What happens if it receives VLAN tagged ethernet frames? Does it ignore the VLAN tagging or does it reject the packet?
[11:25] <soren> Some kind of reference for this would be awesome. I only have virtual hardware to test with and that's not good enough (since I'm suspecting a bug in QEmu's handling in this particular area).
[11:48] <patdk-lap> soren, it's ignored
[11:49] <patdk-lap> basically think of it like this, you have many switchs each named vlan1 to 4096, but you only plug your machine into the ones you configured to listen to that vlan id
[11:49] <patdk-lap> tcpdump will show you the packets though, cause it shows everything received
[11:59] <SpamapS> Daviey: I haven't looked at txaws yet..but it definitely should be syncable
[12:12] <soren> patdk-lap: You appear to be correct. I'm suprised, though. My thinking was something like: The machine doesn't know about vlans at all. It doesn't even have the vlan driver loaded, so why does it go and look at the vlan tag in the ethernet frames?
[12:12] <soren> patdk-lap: What does it even compare it to?
[12:15] <soren> patdk-lap: Oh, hang on. No, I'm not sure you're right. Do you have any references to support it?
[12:15] <soren> patdk-lap: I got confused by my test.
[12:25] <patdk-wk> soren, the reason is cause, it totally kill the ethertype header field in the packet, if the machine doesn't understand vlan tagging
[12:25] <soren> patdk-wk: Ah, yes.
[12:25] <patdk-wk> since the ethertype would not match, it would get dropped as an invalid packet
[12:25] <soren> Right.
[12:25] <soren> Hm.
[12:25] <patdk-wk> unless you have some braindead implementation :)
[12:27]  * patdk-wk totally forgot to comb his hair today :(
[12:51] <Daviey> SpamapS: you are the DM, no? :)
[12:52] <SpamapS> Daviey: I am, just haven't looked to make sure its ok to sync
[12:52] <SpamapS> 99% sure it is, looking at it now actually
[12:53] <agc93> hi all, server newbie here. I'm trying to install 12.04 x64 on my server, onto a SCSI drive, but the drive isn't being recognised. It's connected to an Adaptec PCI card which according to the great interwebs works fine with the adaptec driver included in the kernel. Should I just drop to a shell and modprobe the module in, then try again?
[12:54] <SpamapS> Daviey: it still has an Ubuntu only patch (aws-status-add-appindicator.patch)
[12:55] <SpamapS> Daviey: I poked the desktop team to get appindicator into Debian a few weeks ago
[13:08] <Daviey> SpamapS: ah cool.
[13:18] <agc93> trying to install 12.04 on my SCSI disk, but although lspci correctly reports my SCSI card and everything is connected, my drives aren't appearing in /proc/scsi/scsi. Any ideas?
[13:29] <Daviey> jamespage: i see you are currently triaging the incoming bugs.. Which end are you working from?
[13:29] <jamespage> Daviey, bottom
[13:29]  * jamespage is getting bored of mysql-5.5
[13:30] <jamespage> lynxman is doing some as well
[13:30] <lynxman> Daviey: I'm going from the top
[13:31] <Daviey> jamespage: yeah, i just had a look and noticed a hump of mysql ones.
[13:31] <Daviey> lynxman: awesome!!
[13:31] <jamespage> Daviey, only just got subbed to ubuntu-server
[13:34] <Daviey> jamespage: ahh, that makes sense then
[13:36]  * Daviey considers applying for, https://tbe.taleo.net/NA3/ats/careers/requisition.jsp?org=CANONICAL&cws=1&rid=493
[13:37] <Daviey> (I just posted that btw, and spamming it around)
[13:39] <jamespage> lol
[13:40] <lynxman> Daviey: can I apply? :o)
[13:42] <Daviey> lynxman: it's open to everyone :)
[13:51] <jamespage> lynxman, I've see 4+ like bug 1015408 today
[13:51] <jamespage> might be a maintainer script not doing the right thing with permissions issue
[13:51] <lynxman> jamespage: yeah, all of them for different reasons though, which is a pita
[13:51] <jamespage> might not
[13:52] <lynxman> jamespage: some lack information, this one clearly is someone deleting the user.frm file
[13:52] <jamespage> lynxman, no - specifically with the errno 13 - which means permission denied I think
[13:53] <lynxman> jamespage: hm...
[13:54] <lynxman> jamespage: I'll dig this up further then, hrm
[13:55] <agc93> Anyone know why a SCSI card would be perfectly recognised by both installer and Live CD, but the connected drives wouldn't be, even though there is drive activity?
[14:01] <rowin> agc93: are the devices showing in /dev?
[14:12] <claude2> anyone here good with pxe boot?
[14:13] <agc93> rowin: it doesnt appear so. The SATA drives I have connected appear, and then sg0 through 3 but thats it
[14:14] <stgraber> hallyn: just noticed you landed create(), is it ready to get mapped in the python module?
[14:14] <rowin> agc93: to confirm, when you loaded up the installer/live boot were the drives mountable?
[14:15] <agc93> rowin: no. At the moment, im in the Desktop Live CD and it shows the SCSI controller (the card) but not the drives themselves.
[14:15] <agc93> but I know they're working as they were running in a different PC this morning
[14:15] <jamespage> zul: bug 1013782
[14:16] <rowin> ah, ok, what has changed since this morning? were they moved from one system to another all together or was something else changed?
[14:17] <SpamapS> Daviey: oh! looks like appindicator is actually in Debian now so I can patch the Debian aws-status and sync them. woot.
[14:17] <zul> jamespage:  yeah i know about that one, its fixed in foslom but its fixed done in a rewrite/refactor
[14:17] <jamespage> zul: meh - never mind - answered my own question
[14:17] <jamespage> zul, anyway sounds important
[14:17] <hallyn> stgraber: I think so.  can you check teh src/tests/createtest.c and tell me if you're happy with that api?
[14:17] <jamespage> 'High'
[14:17] <zul> jamespage:  ill update the bug with my findings so far
[14:18] <jamespage> zul, is that released in folsom for what we have in the archive already?
[14:18] <agc93> rowin: moved into a different system. Basically picked up the PCI card with connected drives and dropped it in a new system
[14:18] <zul> jamespage: i think so, otherwise it will be fixed in the archive for quantal tomorrow
[14:21] <rowin> agc93: good, not too many things to work with then, no extra variables. next thing, can you pastebin the result of this sudo lshw -C disk
[14:22] <stgraber> hallyn: a bit suprised you went with a single char* instead of char** like for start() but that's not a problem for me
[14:23] <agc93> rowin: that gives no output :/
[14:24] <hallyn> stgraber: ?
[14:24] <rowin> try it without the -C disk, so just lshw as root (# or sudo)
[14:24] <agc93> rowin: I have since taken out all the SATA HDD's so the only remaining drives are an IDE CD Drive and the SCSI's
[14:25] <hallyn> oh
[14:25] <hallyn> stgraber: not a single char*
[14:25] <hallyn> stgraber: varargs.  like startl
[14:25] <hallyn> stgraber: did you prefer the char** (like ->start vs. ->startl)?
[14:25] <agc93> rowin: lshw on its own shows the SCSI controller, but no drives. Hang on, ill try and get a full output
[14:25] <rowin> agc93: it should show us the scsi controller and any other hw attached
[14:26] <rowin> agc93: cool
[14:27] <hallyn> stgraber: i can do the same ->create vs ->createl.  lemme know
[14:27] <hallyn> stgraber: bug 919619 - an arkose competitor?
[14:28] <stgraber> hallyn: I was expecting create() to be like start(), so basically create(char* template, char** argv) from your example, it looks like it's create(char* template, char* arg) where arg can contain multiple arguments ("-r lucid" in your example)
[14:30] <hallyn> stgraber: d'oh, i used it wrong in my test :)  though it worked, hm.
[14:30] <hallyn> stgraber: it was supposed to be create(c, "ubuntu", "-r", "lucid", NULL)
[14:30] <stgraber> ok, can you split it into createl() and create() like for start?
[14:31] <hallyn> stgraber: yup
[14:31] <hallyn> stgraber: but you use startl() from python, right?
[14:31] <stgraber> I believe createl() will be easier for people directly using the C api and maybe for some other language bindings, but create() is easier for me
[14:31] <stgraber> nope, I'm using start()
[14:31] <hallyn> ah.  ok.  lemme do that right now
[14:32] <stgraber> doing the variable args stuff when I already need to parse all the arguments before was tricky. It was easier to build a char** and pass that to start() instead
[14:32] <Kloeji> Hi, if I plan on using Postfix to handle multiple domains, what domain should I use during the setup, when it asks for a FQDN hostname?
[14:33] <agc93> rowin: heres the whole thing http://paste.ubuntu.com/1052723/
[14:34] <agc93> my scsi controller (An adaptec 19160 based on the 7892b chipset) is listed uner pci:5 under scsi, but then no drives...
[14:34] <hallyn> stgraber: oh that's right, the reason i did that was that it occurred to me that i might have a memory leak with the way i'm doing startl
[14:36] <Kloeji> Hi, if I plan on using Postfix to handle multiple domains, what domain should I use during the setup, when it asks for a FQDN hostname?
[14:37] <rowin> acg93: hmm, I think maybe you may not have scsi support enabled. I have only worked with it once or twice but i'm sure there is a kernel module require
[14:38] <agc93> rowin: there is an aic7xxx module that is supposedly the SCSI one and thats loaded (at least according to lsmod)
[14:38] <agc93> rown: there is also an aic79xx module that I can modprobe in, but that doesnt change anything
[14:38] <Daviey> SpamapS: wooooooooot
[14:40] <SpamapS> Something about getting packages back in sync w/ Debian.. its like calling your mom to see how she's doing just as the doorbell rings with the postman carrying a tin full of cookies from mom..
[14:42] <rowin> acg93: any chance you could dump ls /dev?
[14:44] <agc93> hang on, ill just boot it back up.
[14:45] <agc93> but last time I tried that (when there were SATA drives in there as well), there was just sda,sda1,sdc,sdc1 (the two SATA's), then sg0,sg1,sg2,sg3 and sr0
[14:46] <rowin> acg93: on the other system were the drives partitioned? if so, any raid or other possible complications?
[14:47] <agc93> rowin: theres two drives connected, so kind of. On the first system, both drives were just separate Ext4 partitions and mounted and worked fine
[14:48] <agc93> I then formatted both drives, and one drive is just one massive LVM partition. The other one is two partitions, one LVM, the other Ext4. My intention is to install onto the Ext4 partition and use the combined lvm partitions as a data area
[14:48] <rowin> can you run cat /proc/partitions
[14:50] <jamespage> SpamapS, the NEW bugs queue for mysql-5.5 is a bit shorter now
[14:51] <agc93> rowin: I get two entries returned, one for sr0 (my dvd drive) and the other for loop0 (the live squashfs)
[14:51] <agc93> rowin: wait a minute. I want to try something.
[14:59] <agc93> nevermind, that didnt work either.
[15:00] <trimeta> For some reason, when I reboot my server grub always waits for my input (i.e., doesn't automatically load the default choice after N seconds), even if this is was after a clean reboot. How can I fix it (even if this means never waiting for input)?
[15:01] <patdk-wk> normally? /etc/default/grub
[15:01] <trimeta> GRUB_TIMEOUT=2 is already there.
[15:02] <patdk-wk> what ubuntu version?
[15:02] <trimeta> 12.04.
[15:03] <patdk-wk> upgrade or new install?
[15:03] <trimeta> I'm slightly concerned it may have something to do with my hardware, because when I do "sudo halt," the system doesn't actually shut down after it prints "The system is now halted."
[15:03] <trimeta> New install.
[15:03] <patdk-wk> my netboot does that, but boot is fine
[15:03] <trimeta> Ah.
[15:03] <patdk-wk> well, I seriously don't know the answer :)
[15:04] <trimeta> Fair enough.
[15:07] <SpamapS> jamespage: awesome. Any serious problems coming to light?
[15:08] <SpamapS> jamespage: when I perused them, I just saw the same old problems with starting broken mysqld
[15:12] <hallyn> stgraber: thanks for renaming the tests :)  kept meaning to do that
[15:13] <hallyn> all right, new patch seems to be working.  will push in a bit
[15:13] <hallyn> maybe i'll double check ->start first
[15:15] <stgraber> hallyn: hehe, yeah, I'm usually just adding "mv" calls to debian/rules to put all your tests under lxc-test-. I guess whenever we have the API fully implemented, these should be made optional in the build process as most distros won't really need them
[15:18] <halvors> Hi!
[15:18] <halvors> I'm trying to setup ISC DHCP Server to serve both IPv4 and IPv6, is there som way to start it in IPv6 mode?
[15:20] <halvors> There is no documentation for this on Ubuntu...
[15:20] <stgraber> halvors: just create a /etc/dhcp/dhcpd6.conf file and start isc-dhcp-server6
[15:21] <halvors> Is isc-dhcp-server a service under /etc/init.d?
[15:22] <stgraber> it's an upstart job. You have /etc/init/isc-dhcp-server.conf for ipv4 and /etc/init/isc-dhcp-server6.conf for ipv6
[15:22] <stgraber> so you just need to create /etc/dhcp/dhcpd6.conf then do "sudo start isc-dhcp-server6" and dhcpd6 will start
[15:23] <stgraber> (it'll also start automatically on boot as long as /etc/dhcp/dhcpd6.conf exists and is valid)
[15:23] <halvors> Is there a default dhcpd6.conf file i can use?
[15:23] <halvors> Eventually download?
[15:23] <halvors> Just to get started.
[15:26] <hallyn> there, reallocs cleaned up.
[15:27] <stgraber> halvors: http://www.stgraber.org/download/ipv6/dhcpd.conf.ipv6
[15:27] <halvors> Thanks :D
[15:32] <smoser> jamespage, ping
[15:32] <smoser> triaging...
[15:32] <smoser> it seems like you've touched a bunch of the mysql stuff?
[15:32] <smoser> but its still showing ther.e
[15:33] <hallyn> stgraber: ->create updates pushed
[15:33] <stgraber> hallyn: cool, thanks
[15:33] <hallyn> Question for everyone:  can someone lay out advantages (if any) of vde2 over openvswitch (as pertains to libvirt/qemu)?  Is it mainly the ability to use them unprivileged?
[15:34] <hallyn> stgraber: AIUI I'm done for now with the api.  If you were waiting on something more, please poke.
[15:34] <hallyn> (after an SRU I'll start on the user namespaces patch for lxc)
[15:35] <stgraber> that's the main advantage I can see of vde2, a user with just access to /dev/kvm can use an interface using a user network card (no privileges required) and another using vde2 to talk to another VM (private network)
[15:35] <stgraber> though I haven't used that in years
[15:35] <hallyn> stgraber: I'm trying to gauge its value wrt MIR
[15:35] <hallyn> we dont' watn duplicate functionality in multiple packages in main, and ovs is in main, so...
[15:36] <hallyn> all right, I guess I'll go ahead and request it.  There's just (AFAIK) no alternative for unprivileged (or underprivileged) access.
[15:36] <hallyn> well, there's userspace networking, which frankly works quite well for me :)
[15:37] <halvors2> stgraber: Now i get an IPv6 address on the clients, but seems like they don't know the gateway, is there any way to specify that+
[15:37] <halvors2> ?*
[15:38] <stgraber> halvors2: you're supposed to run radvd for that
[15:38] <halvors2> Ok.
[15:38] <halvors2> Then i think i'll need some help for that...
[15:38] <stgraber> halvors2: http://www.stgraber.org/download/ipv6/radvd.conf
[15:39] <stgraber> halvors2: if you want stateful, use found-dhcpv6, if you want stateless, use found-radvd-dhcpv6
[15:40] <stgraber> halvors2: the source of these examples being an old blog post of mine: http://www.stgraber.org/2011/07/26/state-of-ipv6-in-ubuntu-oneiric/
[15:40] <stgraber> things changed a bit in 12.04, but most of the config still applies
[15:45] <halvors3> stgraber: Got it working :D But how do i set radvd to start on boot?
[15:46] <stgraber> halvors3: it usually starts on boot just fine for me
[15:46] <stgraber> as long as it has a config, it's starting (at least here)
[15:46] <halvors3> Tried to do "start radvd
[15:48] <stgraber> halvors: it's an old style init job, so you'd need "/etc/init.d/radvd start"
[15:48] <stgraber> halvors: or use the service command that works with both systels "service isc-dhcp-server6 start" and "service radvd start"
[15:48] <halvors> Well, i'll have to do a reboot then, just to test :) Thanks you very much for help, now i finally got my lan router running :)
[15:49] <halvors> stgraber: Is it possible not specify a range in the ISC-DHCP-Server but instead provide a prefix to automaticly generate ip adresses for?
[15:50] <halvors> Like you do on Cisco IOS.
[15:52] <stgraber> halvors: range6 allows a CIDR noted subnet so yeah, you should be able to just give it a subnet.
[15:52] <halvors> Just like "range6 2001:xxxx:xxxx:xxxx::/64" ?
[15:53] <stgraber> halvors: yeah, the manpage says it should work
[16:00] <halvors2> stgraber: I still get the ip that ends with ::9999, i want all the groups to get random data...
[16:00] <stgraber> halvors2: then you probably want to use stateless dhcpv6 + privacy extensions
[16:01] <halvors2> stgraber: Isn't is possible with stateful ipv6?
[16:02] <smoser> hallyn, since you said 'autofs' recently, i assume you're an expert. could you take a quick read of https://bugs.launchpad.net/ubuntu/+source/autofs5/+bug/1016068
[16:03] <stgraber> halvors2: it might be but I'm not aware of an easy way to get dhcpd to give you non-sequential addresses, if you want random addresses it's much easier to let the clients build their initial address from the RA using EUI64, then turn on the privacy extensions on the client so they generate a fully random IPv6 on top of that
[16:03] <hallyn> smoser: bwahahahahaha
[16:03] <hallyn> (reading)
[16:04] <halvors2> Is that in your example?
[16:04] <stgraber> halvors2: yep, that's found-radvd-dhcpv6
[16:05] <stgraber> halvors2: the privacy extensions are a client side option that's turned on by default in Ubuntu and I believe on Windows as well
[16:05] <halvors2> I can't just turn "AdvOtherConfigFlag on;"?
[16:08] <hallyn> smoser: funky.  no idea.  his auto.cifs entry looks funky to me (
[16:08] <hallyn> box -fstype=cifs,rw,credentials=/etc/fritz.cred ://192.168.6.1/FRITZ.NAS)
[16:08] <stgraber> halvors2: yep, that's basically the only difference, well, that and removing the range6 definition (as it's no longer the dhcp server's role to give ip addresses)
[16:08] <hallyn> but i'm not up on all the latest autofs config stuff
[16:08] <hallyn> smoser: his nsswitch.conf is the stock one, so i'm guessing autofs is really crashing on teh rest of the config
[16:10] <smoser> yeah, that is what i saw too.
[16:17] <jamespage> smoser, pong - no there was just alot of it
[16:17] <jamespage> not done all ofthem
[16:27] <halvors2> Nice :D
[16:36] <claude2> anyone have any idea what this might mean? I'm getting these logs on my NFS server as I try to nfs mount the root on a diskless install
[16:36] <claude2> Jun 21 12:32:50 seahorse rpc.mountd[3547]: authenticated mount request from 10.13.9.205:790 for /netboot/stax00 (/netboot/stax00)
[16:37] <claude2> Jun 21 12:32:50 seahorse rpc.mountd[3547]: refused unmount request from 10.13.9.205 for /root (/): no export entry
[16:37] <claude2> that repeats over and over
[16:54] <jkyle> in hosts created from a uec image of ubuntu 11.10 I'm seeing this in hte /etc/hosts. Instead of 127.0.0.1 localhost, it says 127.0.0.1 ubuntu.
[16:55] <jkyle> was that a vendor choice at some point?
[16:56] <jkyle> ah, here we go: https://bugs.launchpad.net/ubuntu/+bug/906687
[17:18] <adam_g> zul: ping
[17:21] <zul> adam_g: whats up?
[17:26] <adam_g> zul: so, we need a good versioning scheme for the cloud archive stuff. have you given this any thought yet?
[17:26] <adam_g> it just started causing problems for jenkins jobs
[17:26] <zul> adam_g: not yet...it would probebaly be <version>~precise or something
[17:28] <adam_g> zul: so, something like: quantum-plugin-ryu-agent_2012.2+git201206211313~precise-0ubuntu1_all.deb ?
[17:29] <zul> yeah
[17:29] <adam_g> zul: the issue now is commits to trunk trigger to parallel builds for each $release that end up with the same timestamp/version, and only one makes it to the repository
[17:29] <zul> adam_g: wait they should be going to different repositorys should it not?
[17:31] <adam_g> zul: its a single repository with different pockets. pool/main is shared between them all
[17:33] <zul> ok i thought there would be different pool/main for each version but whatever
[17:34] <zul> so yeah package-version~precise-0ubuntu1
[17:49] <jamespage> utlemming, ping re waagent
[17:49] <utlemming> jamespage: pong...I thought you'd be gone by now
[17:49] <utlemming> jamespage: I had meetings in the morning, so I was thinking I would have to ping you tomorrow
[17:49] <jamespage> utlemming, nah - I had a few hours off this afternoon so working this evening
[17:49] <utlemming> :)
[17:50] <utlemming> jamepage: I was just asking if you had a chance to look at the package after I applied the feedback
[17:50] <jamespage> utlemming, yep - I did - and wanted to grab you to discuss
[17:51] <utlemming> G+ then?
[17:51] <jamespage> utlemming, sure - lemme just grab a coffee
[17:52] <utlemming> jamespage: k, invite posted when you get back with caffinated goodness
[17:57] <jamespage> utlemming, no invite so invited you :-)
[17:57] <utlemming> jamespage: odd....joinging yours
[17:58] <utlemming> jamespage: your room is empty
[17:59] <zul> adam_g: im going to upload a new snapshot tomorrow
[17:59] <jsnapp> can anyone here help me with a cloud-init question?
[18:00] <jsnapp> i'm using nocloud-net with an http datasource ... everything works well except manual_cache_clean: True never seems to get picked up by cloud-init
[18:01] <adam_g> zul: k, everything should be building fine now
[18:01] <adam_g> for trunk at least
[18:09] <jsnapp> can anyone here help me with a cloud-init question?
[18:13] <jamespage> !ask jsnapp
[18:13] <jsnapp> i'm using nocloud-net with an http datasource ... everything works well except manual_cache_clean: True never seems to get picked up by cloud-init
[18:13] <jamespage> bah - wrong subcomman
[18:15] <jsnapp> awfully quiet in here
[18:18] <jamespage> smoser, ^^ never used that option myself - any ideas?
[18:19] <smoser> jsnapp, it "should work"
[18:19] <smoser> oh.
[18:19] <smoser> jsnapp, no. it wont. i'm sorry.
[18:20] <smoser> that option has to be in local config as its consulted before datasources are considered.
[18:20] <smoser> 2 things:
[18:20] <smoser> a.) file a bug, as we *could* make it work (by trying to load existing cached datasource, and then checking the value)
[18:21] <smoser> b.) you can have your userdata/cloud-config write a new file in /etc/cloud/cloud.cfg.d with 'manual_cache_clean: True'
[18:21] <smoser> then reboot will have it.
[18:22] <jsnapp> ok, i was starting to suspect it wasn't designed to work with nocloud
[18:22] <jsnapp> i'm not sure i even need it though
[18:23] <jsnapp> i am setting kernel parameter ds='nocloud-net;s=http://192.168.25.77:8000/' for the first boot of a vmware ubuntu vm
[18:24] <jsnapp> then i add ds='nocloud' to the grub.cfg so any subsequent boots will also be nocloud thereby skipping the search for EC2 datasource
[18:31] <jsnapp> smoser, does my strategy seem to make sense?
[18:32] <jsnapp> since instance-id will default to "nocloud" on both the first nocloud-net boot as well as all subsequent nocloud boots i should be able to count on cloud-init not actually changing anything on subsequent boots?
[18:33] <smoser> jsnapp, what version of ubuntu? 12.04 ?
[18:33] <jsnapp> i'll experiment with your option of having cloud-config write it to a file
[18:33] <jsnapp> yes 12.04
[18:34] <smoser> how are you manipulating the kernel parameter on first boot?
[18:34] <smoser> you know you dont need to manipulate kernel parameters with 12.04, it is perhaps easier for you to attach a second disk.
[18:34] <jsnapp> vmware vsphere client has a console of the vm as it boots ... i just modify grub menu
[18:35] <smoser> oh. wow. so you do that manually. yuck. (sorry).
[18:35] <jsnapp> yeah, i saw that option just today ... you mean create an iso right?
[18:35] <smoser> would you have the ability to attach a cdrom?
[18:36] <jsnapp> yeah, that's possible ... so would i just put the http seed in the iso so i don't have to keep updating the iso if i change my user-data?
[18:37] <smoser> or it can be any disk, really.
[18:37] <smoser> http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/doc/nocloud/README
[18:37] <smoser> jsnapp, you could just have a '#include' in the user-data for your url
[18:38] <jsnapp> oh yeah that makes sense
[18:38] <smoser> and then, also http://smoser.brickies.net/git/?p=tildabin.git;a=blob;f=make-seed-disk;hb=HEAD
[18:38] <smoser> the first is doc on how ot make the disk.
[18:38] <smoser> the second basically does it for you.
[18:38] <jsnapp> cool, thanks
[18:39] <jsnapp> so i don't necessarily want to count on the datasource being available on next boot so is that where i use your option of having cloud-config write manual_cache_clean: true to the filesystem so cloud-init effectively becomes disabled on next boot?
[18:42] <jsnapp> and that way i won't have to keep that disk attached AND I won't have to permanently modify /boot/grub/grub.cfg to include ds='nocloud' ?
[18:48] <jsnapp> smoser, sorry to bug you but did you see my last two questions?
[18:48] <al_the_x> Hey, anybody here running 12.04 and running MySQL replication?
[18:49] <glance> mmm...
[18:49] <smoser> no, jsnapp i did not.
[18:49] <matt_keys> al_the_x : yes.
[18:49] <smoser> jsnapp, you are correct, that is the purpose of manual_cache_clean: true
[18:50] <smoser> the other thing you will want to do is '#include-once' rather than '#include'
[18:50] <al_the_x> matt_keys: Did you edit your /etc/mysql/my.cnf directly or use separate files in conf.d/
[18:50] <jsnapp> smoser, ok i wondered about that
[18:50] <smoser> as '#include' will attempt to include every time... although i have to think harder to be sure of which one wins.
[18:51] <matt_keys> al_the:x : I suppose it depends on what you're trying to do
[18:52] <jsnapp> smoser, what two pieces are competing?  include_once and manual_cache_clean ?
[18:52] <jsnapp> sorry, i meant "#include and manual_cache_clean" ?
[18:53] <smoser> jsnapp, right. thats what i meant.
[18:54] <matt_keys> RoyK : you around?
[18:55] <al_the_x> matt_keys: On 11.04, I could put a symlink in /etc/mysql/conf.d/ and mysql would pick it up.
[18:55] <matt_keys> does the symlinked file end with .cnf?
[18:56] <al_the_x> Now in 12.04 (with MySQL 5.5), symlinks aren't read. HARD links are, but not SOFT links.
[18:56] <al_the_x> matt_keys: Yes.
[18:56] <patdk-wk> symlinks iwht mysql has always been fine
[18:57] <patdk-wk> an easy way to shard database load across disks
[18:57] <al_the_x> Another interesting point: if the file contains a dot other than .cnf, the file is ignored, as well.
[18:57] <jsnapp> smoser, so most of my systems will just have a base user-data to get basics setup and then chef will take over from there (i'm not using chef option in cloud-config) ... but i do have a specific chef-server-user-data file that just one vm will need to use
[18:57] <al_the_x> eg: replication.001.cnf = bad; replication-001.cnf = good
[18:58] <smoser> jsnapp, right. that makes sense.
[18:58] <jsnapp> smoser, do i just have two iso's or disks then?  one that include_once the base user-data and another that include_once the chef-server-user-data?
[18:58] <al_the_x> patdk-wk: Sorry, were you chiming in...?
[18:58] <patdk-wk> kindof :)
[18:59] <al_the_x> Okay. I'm having trouble with symlinked config files, though, not data dirs.
[18:59] <al_the_x> Thanks, though.
[18:59] <jsnapp> smoser, sorry but i think you responded at same time as i sent my last little bit ... for clarity you're saying it makes sense to have two iso's or disks right?
[18:59] <patdk-wk> hmm
[18:59] <al_the_x> matt_keys: Do you mind running an experiment on your box?
[18:59] <patdk-wk> you sure it's the symlink that is the issue? or the appprofile?
[19:00] <al_the_x> appprofile?
[19:00] <smoser> jsnapp, yeah, you'd need two different isos. i dont know how else you'd differenciate behavior.
[19:00] <jsnapp> smoser, i agree
[19:00] <patdk-wk> apparmor
[19:00] <smoser> jsnapp, and i am fairly sure that what you want to do is have #includ-once
[19:01] <smoser> that itself would be sufficient if you left the cdrom plugged in.
[19:01] <al_the_x> patdk-wk: That's a good point. I hadn't thought about it...
[19:01] <matt_keys> al_the_x : I suppose. You don't have a dev environment?
[19:01] <smoser> but if you want to be able to eject the cdrom also, then you'd need to make the manual_cache_clean change to local
[19:02] <al_the_x> matt_keys: I've confirmed in both my environments, looking for a second opinion.
[19:02] <jsnapp> smoser, great it's all making sense now
[19:02] <al_the_x> I think that patdk-wk might be onto something with the apparmor profile, though.
[19:02] <jsnapp> smoser, last question ... if the iso uses include_once to get the user-data and the user-data retrieved has #include's in it those shouldn't have to be #include_once's too do they?
[19:03] <smoser> :).
[19:03] <al_the_x> patdk-wk: That's it, sir...
[19:04] <matt_keys> al_the_x : that makes sense to me too, I can see apparmor profile for mysqld
[19:04] <al_the_x> Indeed. I think that's the ticket. Thanks, fellas!
[19:06] <matt_keys> patdk-wk : interesting I see libvirtd in there as well... do you think that would have anything to do with a mmap allocation problem on a guest? I was wondering if maybe it was ksmd, but have no idea how to troubleshoot it
[19:06] <smoser> jsnapp, it looks to me that if you have '#include-once <URL1>' that does '#include <url2>' that only the contents of 'URL1' are cached.
[19:06] <smoser> meaning that it will then try to '#include url2' on subsequent runs.
[19:13] <jsnapp> smoser, hmmm, if only the contents of 'URL1'  ('#include <url2>') are cached and it's semaphored as include_once then it seems like cloud-init wouldn't try to '#include <url2>' again
[19:13] <jsnapp> smoser, but your the author so i'll take your word for it :)
[19:13] <smoser> jsnapp, well, it caches the contents of the url
[19:14] <smoser> so the next time through, it says "oh, i dont have to get URL1, cause i've got it locally"
[19:14] <smoser> it opens up the local file
[19:14] <smoser> and sees
[19:14] <smoser> #include <url2>
[19:14] <smoser> and says "oh, i need to get that"
[19:15] <jsnapp> yeah, but i thought the point of include_once is that it semaphores that content to not be included the next time through
[19:16] <jsnapp> i think i'm misunderstanding the "_once" part
[19:16] <jsnapp> include_once doesn't mean run_once?
[19:16] <jsnapp> it literally means don't retrieve content from the url again?
[19:17] <jsnapp> therefore it will run "again" any cached content?
[19:17] <jsnapp> which in my case would be '#include <url2>'
[19:17] <jsnapp> i think i follow you
[19:21] <jsnapp> smoser, thanks a lot ... i should have tried irc earlier ... you've been very helpful
[19:22] <smoser> yeah. it basically means "cache local content"
[19:22] <smoser> jsnapp, if you ahve other questions, feel free to ask.
[19:22] <smoser> the other hting...
[19:22] <smoser> id be interested in knowing why you're not using the chef support in cloud-init
[19:23] <smoser> as i'd like to make people like you say "oh, that makes sense, i'll use that"
[19:23] <smoser> (ie, the reason its there is that so people use it)
[19:26] <ironm> does anyone of you run interface bonding on ubuntu 12.04?
[19:27] <ironm> hello. Do I really need "ifenslave" for interface bonding (LinksAggregation) on ubuntu-server 12.04?
[19:27] <ironm> all configurations (without ifenslave) didn't work ...
[19:27] <ironm> all configurations I have tested ...
[19:32] <matt_keys> patdk-wk : Nope, that wasn't it... symlinked the qemu-<uuid> to /etc/apparmor.d/disable but the behavior didn't change on the guest :(
[19:32] <patdk-wk> hmm?
[19:33] <patdk-wk> apparmor profile on a host won't affect a guest
[19:33] <patdk-wk> but now your talking qemu, I have no clue what your doing :)
[19:33] <matt_keys> patdk-wk : was worth a shot. :) having a hard time figuring this one out... keep seeing mmap allocation errors on the guest that i haven't seen before
[19:33] <matt_keys> only change was upgrading from 10.04 to 12.04 lts
[19:34] <matt_keys> well, i take that back... there was a centos 6.2 kernel update recently, but i rolled it back to test if that was it and it wasn't
[19:36] <matt_keys> make that libvirt-<uuid>
[19:37] <jsnapp> smoser, in regards to the chef support I'll start by saying that I'm no expert in regards to deployment but I'm trying to keep things as cookie cutter as possible ... i'd rather not have to have multiple iso's each pointing to a different user-data file (e.g. web-user-data, db-user-data, etc)
[19:37] <smoser> jsnapp, i dont follow that.
[19:38] <jsnapp> probably because i'm making weird decisions? cause i'm not an expert
[19:38] <jsnapp> :)
[19:38] <smoser> well i know very little about chef
[19:39] <jsnapp> i would love to have someone (cough..smoser) tell me the "right" way
[19:40] <smoser> i know very little about chef.
[19:40] <smoser> but doc/examples/cloud-config-chef.txt
[19:40] <jsnapp> right that's what i'm looking at now
[19:40] <smoser> you'd have ot have a chef server (your chef-server-user-data)
[19:40] <smoser> and then one fo rech of the nodes that new where the serer was
[19:40] <jsnapp> if i use cloud-config-chef then at least node_name and run_list would be different for each vm
[19:41] <smoser> but then how you turn a generic node who calls home to the chef server into a specific funciton, i don tknow.
[19:41] <smoser> and that may force you to have separate isos with "node-1" "node-2" somewhere.
[19:42] <jsnapp> right, then i think we're on the same page as long as we use cloud-config-chef to *only* create generice nodes and then use chef to configure them into a specific function
[19:43] <smoser> yeah. i see hte difficulty there.
[19:44] <jsnapp> in which case i could use cloud-config-chef but i'm not sure i see too much benefit over just using chef's knife bootstrap command which uses ssh to send the private validation key, private databag key and doesn't rely on cloud-init so it's platform independent
[19:45] <hallyn> smoser: we need a ubuntu-cloudimg-query bot in here :)
[19:45] <hallyn> @smoserbot ebs quantal
[19:45] <smoser> ami-36b1115f
[19:46] <hallyn> smoserbot: thx :)
[19:46] <smoser> us-east-1 ami-36b1115f canonical    ebs/ubuntu-quantal-alpha1-amd64-server-20120606.1
[19:46] <smoser> but you probably wanted
[19:46] <smoser> @smoserbot ebs quantal daily
[19:46] <smoser> ami-fa9c3e93
[19:47] <hallyn> oh, the other is older?
[19:47] <hallyn> yeah daily is good, thx :)
[20:18] <ironm> hello.  When I use ifenslave on ubuntu 12.04 interface bonding works as expected (with following config files: interfaces - http://paste.debian.net/175728/ ... and bonding.conf - http://paste.debian.net/175729/)
[20:18] <ironm> I have tested some configurations (see config files in brackets) without ifenslave, and *no* one works :( interfaces - http://paste.debian.net/175730/ ... bonding.conf - http://paste.debian.net/175731/ ... modules - http://paste.debian.net/175732/ ). Do you have an idea what do I miss? Thank you in advance for any hints or examples of working config files without using ifenslave.
[20:19] <ironm> it is *not* documented in unbunt administration guide
[20:20] <jsnapp> smoser, given my descriptions before i wanted to add that i'm currently just using an empty meta-data file ... since i'm trying to use cloud-init to create generic nodes it made sense to me not to populate meta-data since that seemed to typically consist of unique info ... does that make sense to you too
[20:21] <smoser> jsnapp, well, you need to have an instnace-id in it.
[20:21] <smoser> but it does not have to be unique.
[20:21] <smoser> that is a limitation i'll probably attempt to remove.
[20:21] <jsnapp> smoser, but that defaults to nocloud which seemed sufficient
[20:21] <smoser> oh. ok.
[20:21] <smoser> sure.
[20:21] <smoser> :)
[20:22] <jsnapp> cool
[20:25] <jsnapp> smoser, as an aside do you see user-data and cloud-init typically getting used to create many unique nodes ... like in AWS EC2 do people actually use user-data that much for complicated unique configurations?
[20:26] <smoser> jsnapp, i'm sur ethat some people do have complex user dta that they feed into a node.
[20:26] <smoser> i certainly have lots of user-data snippets that i create instances with to then do something to.
[20:26] <smoser> and for single systems, its useful
[20:27] <smoser> but i suspect (and if not hte case right now, then i expect it will be sometime soon) that people will primarily use cloud-init as a bootstrap to a configuration management solution
[20:27] <smoser> or something along those lines
[20:27] <smoser> (ie, "get me to juju, puppet, chef, salt"...)
[20:28] <jsnapp> yeah that makes sense
[20:30] <jsnapp> I'm just curious. cloud-init has been ported to RHEL or Fedora right?
[20:33] <jsnapp> smoser, I'm just curious. cloud-init has been ported to RHEL or Fedora right?
[20:34] <smoser> it is in fedora (F16 and beyond i think)
[20:34] <smoser> but it is less functional
[20:34] <jsnapp> ok
[20:34] <smoser> but someone is working on making it more functional there right now.
[20:34] <smoser> so i would suspect that for the next version it will be much closer to parity.
[20:35] <jsnapp> ok, thanks for the info
[21:22] <daff> why would "apt-get install foo" complain that "The following packages cannot be authenticated!" when "aptitude intall foo" works fine? the correspoding signing key for package foo is installed and listed in "apt-key list". any ideas?
[21:24] <ironm> daff, you have to install appropriate gpg-keys
[21:25] <ironm> daff, an example for MariaDB deb packages; http://kb.askmonty.org/en/installing-mariadb-deb-files/
[21:26] <daff> ironm: thanks, but the appropriate key is already installed. "apt-key list" lists it
[21:26] <stgraber> hallyn: "lxc_container: failed to save starting configuration for test" when calling create("ubuntu") on a non-existing container
[21:27] <daff> and aptitude doesn't seem to have a problem, only apt-get. which I find strange.
[21:28] <daff> what's also interesting is that the .gpg file for the repo/key is nowhere to be found in /var/lib/apt/lists
[21:31] <ironm> daff, check with "dpkg -l | grep ..." if you can see it too
[21:31] <ironm> daff, did you "apt-get update"
[21:34] <daff> yes, of course I did apt-get update :)
[21:35] <daff> what should "dpkg -l" show me, other than a list of packages?
[21:54] <adam_g> jamespage: FYI all the jenkins jobs have been renamed internally to ${ubuntu-release}_${openstack-release}-${component}-${target} if you want to make the required changes on the public jenkins side
[21:56] <hallyn> stgraber: nothing more in lxccontainer.log?
[21:58] <stgraber> hallyn: nope, just that single line
[21:59] <hallyn> stgraber: can you send me a test case?  will try after dinner
[22:00] <hallyn> oh does /var/lib/lxc exist?
[22:00] <hallyn> i dont recursively mkdir
[22:01] <hallyn> (yet :)
[22:03] <stgraber> hallyn: yeah /var/lib/lxc exists
[22:04] <stgraber> hallyn: I'll quickly try to reproduce that in C, should be a two lines test
[22:06] <stgraber> hallyn: http://paste.ubuntu.com/1053379/
[22:08] <hallyn> huh
[22:10] <hallyn> oh
[22:11] <hallyn> yeah, I guess we could allow for this, but you have to first create the config (i.e. by doing a c->set_config_item("lxc.network.type", "veth");)
[22:12] <hallyn> stgraber: I'll fix that tonight to create an empty dummy config at save_config() if one doesn't exist.
[22:12] <hallyn> I think I purposely didn't do that at first bc i wasn't sure it would be the right thing.  but i guess it is
[22:13] <stgraber> hallyn: just load /etc/lxc/lxc.conf if no config exists?
[22:13] <stgraber> hallyn: that'd match our current behaviour
[22:13] <hallyn> right.  i'll  have it do that tonight
[22:13] <hallyn> stgraber: for the record http://pastebin.com/aqZY8rSW works
[22:15] <stgraber> hallyn: right, I've changed my test to do a load_config("/etc/lxc/lxc.conf") for now
[22:15] <hallyn> stgraber: mind you, upstream we'll just want to init an empty one :)
[22:15] <hallyn> ok, i'll fix that up tonight, thanks.  bbl
[22:24] <stgraber> hallyn: it also seems like it's always returning true even when failing (calling create on a container with an existing rootfs)
[22:33] <stgraber> hallyn: my bad apparently, it's something going wrong in the python code
[22:52] <hallyn> stgraber: the c code for bool rets is pretty simple so would've been tough to muck up, but if anyone could i could :)
[22:53] <agc931> If I have an Adaptec PCI SCSI card connected and Ubuntu 12.04 is recognising the card properly but not the (confirmed working) drives attached to the card, am I missing a kernel module or something?
[23:01] <freesbie_> some controllers dont present the drives unless you have them configured in a raid or as jbod, could that be the issue ?
[23:04] <slide> Does anyone know if it is possible to change the visible directory of a nfs share? for instance on the server i want to share the folder /mnt/Downloads, but have it be visible as /Downloads
[23:05] <SpamapS> slide: mount --bind /mnt/Downloads /Downloads would be the simplest way
[23:06] <slide> i guess, but that clutters up my server
[23:08] <SpamapS> slide: https://help.ubuntu.com/community/NFSv4Howto
[23:08] <slide> yea i read that, did i miss something?
[23:09] <slide> hrm kinda like a chroot in /export ?
[23:09] <slide> no i still think it would use the /export prefix...
[23:10] <SpamapS> slide: not according to the Howto
[23:10] <SpamapS> Note with remote NFS paths
[23:10] <SpamapS> They don't work the way they did in NFSv3. NFSv4 has a global root directory and all exported directories are children to it. So what would have been nfs-server:/export/users on NFSv3 is nfs-server:/users on NFSv4, because /export is the root directory.
[23:10] <slide> i think thats because /export was shared as the root of the filesystem fsid=0
[23:10] <slide> hrm
[23:10] <slide> let me try to set it up like that and see
[23:10] <SpamapS> slide: yes that is why, but fsid=0 is special
[23:10] <slide> so should my /export/Downloads be a link to /mnt/Downloads or a mount --bind ?
[23:12] <SpamapS> slide: bind mount
[23:12] <SpamapS> slide: or you could share /mnt as your fsid 0 :)
[23:12] <slide> true
[23:19] <slide> hrm
[23:19] <slide> so i set /mnt to fsid=0 but my /mnt/Downloads gets the error "/mnt/Downloads requires fsid= for NFS export"
[23:20] <slide> maybe because /mnt/Downloads is actually a mhffds mount of /mnt/Downloads-01 and /mnt/Downloads-02 ?
[23:20] <SpamapS> slide: no idea.. but yeah maybe its time to use bind mounts :0
[23:25] <slide> man wtf
[23:50] <pdtpatrick> Question - can someone please share a guide/wiki they used to successfully add Ubuntu server to Active Directory