/srv/irclogs.ubuntu.com/2012/09/06/#ubuntu-server.txt

=== gary_poster is now known as gary_poster|away
adam_gDaviey: what do you think of this approach to eliminating horizon's node.js dependency? https://code.launchpad.net/~gandelman-a/ubuntu/quantal/horizon/1024326/+merge/12298200:39
adam_gsmoser: ^00:39
=== arrrghhhAWAY is now known as arrrghhh
ninjixanyone knowledgeable about cloud-init in channel?00:42
=== arrrghhh is now known as arrrghhhAWAY
ninjixtrying to figure out why my bootcmd sed command isn't running00:43
ninjixhere's my user-data http://paste.ubuntu.com/1188054/00:46
=== cpg is now known as cpg|away
ninjixbasically, I'm trying to get the apt-get update to go faster by commenting out the deb-src00:48
mikaladam_g: that looks reasonable to me00:48
adam_gmikal: oh cool, thanks. i'll push up the actual horizon to gerrit for some review tomorrow.00:48
adam_g*actual patch to horizon00:48
mikaladam_g: my reading of lessc is that it supports other compressor modes, but I'm very not an expert00:49
=== troy__ is now known as Troy^
zuladam_g: looks ok to me although i know less than squat about node.js00:53
adam_gmikal: im not either, but what ive gathered is that lessc is one of the supported precompilers that python-compressor can use for doing this kind of stuff. there are others, only one of which is in our main archives, and it requires java. so... i think its just easier to disable by default and make it easy for a user to enable if they want.00:53
adam_gzul: cool00:53
mikaladam_g: works for me00:54
zuladam_g: although lessc should be a suggests if they want to use it00:55
adam_gzul: oh, word.00:55
uvirtbotNew bug: #1046620 in tftp-hpa (main) "tftpd-hpa 5.2-1ubuntu1 failed to purge, exit status 127" [Undecided,New] https://launchpad.net/bugs/104662001:16
=== n0ts_off is now known as n0ts
=== [1]unsecur3d is now known as unsecur3d
hallynstgraber: I'm going to push lp:~serge-hallyn/ubuntu/quantal/lxc/lxc-fix-getitem-segv to ubuntu:lxc ?02:27
stgraberhallyn: sounds good02:32
stgraberhallyn: can you also push to your git? I think the utsname fix is in there but not the clear_config one02:34
hallynhm?  yeah, i'll check and push.  thought i had02:35
hallynoh i was probably waiting for testing02:36
=== n0ts is now known as n0ts_off
crystalcastwhat is ubuntu?02:43
hallynstgraber: done03:02
stgraberhallyn: thanks03:03
hallynstgraber: np.  pls let me know if i messed any of it up!03:04
hallynstgraber: did you walk away from ubuntu-devel feeling we should make / ro in ubuntu containers before starting init?03:05
hallyncould experiment with adding a lxc.conf option to dothat03:06
stgraberif it's not too much trouble, an option might be interesting to better replicate what we have on Ubuntu03:06
stgraberbut we need to make sure apparmor actually allows that and that it'll work for cases where we use a block device for rootfs03:07
hallynthe option is no trouble.  I'm just not sure how it fits in with the rest of the pre-boot mounts that we do03:07
hallynright03:07
stgraberI didn't read the whole backlog, so I'm not sure what problem we're trying to fix exactly03:07
hallyneh i can give it a shot.  I'm so far having no luck with userns, wouldn't mind trying something different03:07
stgraberand as we'll never run fsck at container boot time (well, at least not for these sharing the fs with the host), I'm not sure I see the point03:07
hallynlikewise - but i think it has to do with cloudinit wnating to be able to run when / is mounted, as detected by an upstart event03:08
stgraberwell, adding the option and seeing what blows up might still be an interesting experiment ;)03:08
hallynobviously subject to FF :)  but easily done in a ppa03:09
hallynall right lemme see03:09
stgraberyeah, for now I'm pushing all the new stuff to git for inclusion upstream, then if we see that we really need some of the that in 12.10, we'll just file an FFe03:11
stgraberbtw, I see you fixed a bug in lxc-destroy (symlinks). highvoltage mentioned to me that apparently lxc-destroy fails when using btrfs, something about trying to rm what's essentially a mount point03:11
stgraberI don't have that kind of setup here so I didn't look at fixing it yet, maybe he'll be sending a patch if he gets bored of seeing the failure and having to flush it manually ;)03:12
hallynstgraber: depending on his setup, my fix may fix his03:19
hallynor, his may be an "unsupported" (i.e. nigh upon impossible to guess) setup...03:20
hallyndo you know if he'll be at next uds?03:20
stgraberhallyn: yep, he'll be there03:20
=== n0ts_off is now known as n0ts
hallyncool03:21
ninjixanyone having slow connections to security.ubuntu.com?03:32
uvirtbotNew bug: #1022721 in python-greenlet (main) "Switch from python-support to dh_python2" [High,Fix released] https://launchpad.net/bugs/102272103:37
=== cpg|away is now known as cpg
=== n0ts is now known as n0ts_off
=== n0ts_off is now known as n0ts
tonyyarussoI seem to be having trouble with user preferences for SpamAssassin.  I'm running 3.3.2 on Ubuntu 12.04 with Postfix 2.9.3 and Amavis 2.6.5, with user preferences stored in a PostgreSQL 9.1 database.  The issue appears to be that those preferences aren't being read or aren't being honored.  For instance, I have a preference of "blacklist_from" with a username of "me@ourdomain.org" and a value of ...04:40
tonyyarusso... "another_address_of_mine@differentdomain.net", and I can still send messages through from the supposedly blocked account.  I do have user_scores_dsn, user_scores_sql_username, user_scores_sql_password, and user_scores_custom_query set (I've tried with and without that last one).  What else should I check?04:40
hallynstgraber: perhaps you should respond to Dan Kegel with your plans to reimplement lxc-start-ephemeral?04:44
SpamapStonyyarusso: that sounds really specific to SpamAssassin .. have you tried their official forums/lists/channel?05:56
tonyyarussoSpamapS: I posted a few hours earlier on their IRC, but no sign of life so far.  May have to seek out the other options.06:03
koolhead17i all06:10
koolhead17*hi06:10
uvirtbotNew bug: #1046684 in lxc (universe) "debian-templates missed hostname setting" [Undecided,New] https://launchpad.net/bugs/104668406:51
Davieyadam_g: missing a bzr add?07:29
uvirtbotNew bug: #1046696 in lxc (universe) "lxc-shutdown can not shutdown or reboot debian host" [Undecided,New] https://launchpad.net/bugs/104669607:31
=== n0ts is now known as n0ts_off
=== n0ts_off is now known as n0ts
=== mcclurmc_away is now known as mcclurmc
=== mcclurmc is now known as mcclurmc_away
=== mcclurmc_away is now known as mcclurmc
=== mcclurmc is now known as mcclurmc_away
=== Kiall is now known as zz_Kiall
=== zz_Kiall is now known as Kiall
uvirtbotNew bug: #1046717 in clamav (main) "package clamav-daemon (not installed) failed to install/upgrade: unable to create `/usr/sbin/clamd.dpkg-new' (while processing `./usr/sbin/clamd'): Permission denied" [Undecided,New] https://launchpad.net/bugs/104671709:11
=== cpg is now known as cpg|away
=== mcclurmc_away is now known as mcclurmc
uvirtbotNew bug: #1046771 in qemu-kvm (main) "Time runs too fast" [Undecided,New] https://launchpad.net/bugs/104677110:51
=== n0ts is now known as n0ts_off
=== n0ts_off is now known as n0ts
=== n0ts is now known as n0ts_off
n_npHello, my nic connects to my network, establishes that its 100 full dublex and rx tx flow control is on, then it just goes down again directly11:34
n_nphow can i correct this?11:34
patdk-laplack of dhcp?11:34
=== n0ts_off is now known as n0ts
=== gary_poster|away is now known as gary_poster
briggzdoes anyone here have any experience in disabling specific usb ports?13:05
=== n0ts is now known as n0ts_off
jamespagehallyn, can you take a look at bug 1046771 when you start please13:32
uvirtbotLaunchpad bug 1046771 in qemu-kvm "Time runs too fast" [Undecided,New] https://launchpad.net/bugs/104677113:32
jamespageI know what cjwatson means (feels like that every day to me) :-)13:33
hallynjamespage: hm, ok13:39
ninjixI have a 12.04 machine running puppet client that keeps getting a puppet class assigned to it even though I've now set the specific host to it's own node definition13:39
ninjixanyone experienced this with Precise?13:40
jamespagehallyn, I was able to confirm the symptoms13:40
hallynjamespage: you've not built a package with those commits to test it though right?13:41
jamespagehallyn, nope - only just confirmed it13:42
hallynis beta freeze lifted today?13:42
jamespagehallyn, once beta is released yes13:42
hallynpresumably i can push to -proposed anyway13:42
hallynjamespage: thanks for looking at the suse pkg :)13:45
ninjixfound my puppet problem, was a lurking include :)13:50
stgraberhallyn: yep, will do. Was planning on including his work in the python port anyway.13:50
hallynjamespage: I can' treproduce this though13:53
hallyntimeout=10 sleeps for 10 seconds for me13:53
jamespagetimeout=2000 took about 2 seconds for me13:53
hallynoh, oops.  i'm probably using wrong rtc13:55
hallynhm, no , you're not specifying any13:55
hallynwonder if it's an intel-only bug13:56
jamespagemight be13:56
jamespageI'm using intel13:56
* hallyn reboots his precise box that was setting up for a cirrus vga bug reproduction13:58
smbhallyn the one where it fails to have vts?14:01
smbOr better where they are corrupted14:01
hallynsmb: bug 104584514:02
uvirtbotLaunchpad bug 1045845 in qemu-kvm "12.10 guest crash on login when using 12.04 qemu-kvm with cirrus driver" [Low,Confirmed] https://launchpad.net/bugs/104584514:02
hallynsmb: what are the ones you're talking about?14:02
smbbug 103805514:03
uvirtbotLaunchpad bug 1038055 in linux "graphics fail to initialise correctly, in kvm with cirrus graphics (after LUKS install)" [High,Confirmed] https://launchpad.net/bugs/103805514:03
smbBut they might be relared14:03
smbrelated14:03
hallyncould be.14:03
hallynthanks14:03
smbI think there is two states racing: sometimes the cirrus driver fails to get the VRAM resources. Then us use the cirrus driver for x14:04
smbAnd that crashes on login14:04
hallynwonderiffic14:04
smb(meaning we have that problem with Xen all the time)14:04
smbFor qemu/kvm the cirrus drm driver exists which then causes the modesetting x driver to be used14:05
hallynmeaning it keeps being reintroduced?14:05
hallynoh i see14:05
smbThat one does work .... s l ooooo w l y14:05
* smb and unity 3d on VMs are not good friends14:06
hallynsmb: yeah, always adds 10 secs to pull up the dash to get an xterm up14:09
LinuxAdminhi everyone14:10
uvirtbotNew bug: #1046851 in juju (universe) ""juju-origin: lp:juju" fails to install dependencies" [Undecided,New] https://launchpad.net/bugs/104685114:10
LinuxAdminI've just installed ubuntu server 12.04 and I cannot find some usual files14:10
LinuxAdminI cannot find grub.cfg and /var/log/messages14:11
LinuxAdminis this normal?14:11
LinuxAdminwere those files moved to another place?14:11
smbLinuxAdmin, messages is /var/log/syslog and grub.cfg should be in /boot/grub14:12
LinuxAdminthey are not there14:12
LinuxAdminsorry, syslog is there14:12
LinuxAdminsmb, I was trying to open with my unprivileged user14:13
LinuxAdminwith root it opens14:14
LinuxAdminthanks smb14:14
smbAh yeah, some files permissions are tightened up14:14
alaingCork How do I write that length selector14:33
zulsmoser: ping so whats yoru script suppose to prove?14:38
smoseri have more than one script14:39
smoserone of them is about this computer programmer, who works at night as a maple syrup thief.  I'm shopping that script around, trying to get it into a hollywood movie. but i haven't had any buyers yet.14:40
zulasshat14:40
zulthe ec2 script for the missing dates for ec214:41
Picihah14:41
zulsmoser: this bug https://bugs.launchpad.net/bugs/82756914:42
uvirtbotLaunchpad bug 827569 in nova "ec2metadata service does not include 2011-01-01" [Wishlist,Confirmed]14:42
stgraberhallyn: I'll take care of bug 1046684 and bug 104669614:44
uvirtbotLaunchpad bug 1046684 in lxc "debian-templates missed hostname setting" [Undecided,New] https://launchpad.net/bugs/104668414:44
uvirtbotLaunchpad bug 1046696 in lxc "lxc-shutdown can not shutdown or reboot debian host" [Undecided,New] https://launchpad.net/bugs/104669614:44
smoserzul, its just there for reference, if someone wanted to implement additional EC2 MD.14:44
zulsmoser:  ah...i see14:44
stgraberhallyn: the patches contain typos and need proper splitting, but that's trivial. I'll push to my git branch and to ubuntu:lxc14:45
smoserso you dont think my movie idea will work?14:45
hallynstgraber: thanks14:46
koolhead17hallyn: hi514:48
koolhead17someone just tested LXC/ Openstack on precise14:49
hallynjamespage: yeah, imeout goes by faster on intel. it only waited 3-5 seconds instead of 10. guess it's very hw specific.  now to build proposed fix14:53
hallynkoolhead17: excellent - and it worked great i hope :)14:53
koolhead17hallyn: yes. will share the blog on same soon14:53
hallynkoolhead17: awesome, thx14:56
phaidroshow do I map a domain user's primary group ("domain user" from AD) to a local group (gid=33) ?14:56
hallynsmb: grasping at straws maybe, but the timekeeping bugs just may be responsible for cirrus bugs, it seems like.  (we'll see)15:02
smbhallyn, depends which cirrus bug you mean. :) The one I am looking at right now is because the efi vga frambuffer may not quickly enough release its resources and cirrus giving up quickly when trying to allocate the same mem region15:04
hallynsmb: ah.  so that's likely not fixed upstream?15:05
smbhallyn, not when I looked last (yesterday)15:06
smbhallyn, I am doing a test kernel which you may try15:07
hallynsmb: no hurry, i'm not there yet15:08
stgraberhallyn: pushed to git and ubuntu branch15:09
hallynstgraber: ubuntu branch is getting scary :)15:11
stgraberhallyn: yeah :)15:11
RoyKphaidros: idmapd15:17
phaidrosRoyK: thats as far as I came as well, but I seem to be not understand what exactly to do with idmapd in my case. is there any way to just map a specific group?15:20
RoyKit's in the docs15:21
smoserjibel, would you consider lp:~smoser/+junk/qatracker for merge?15:28
jibelsmoser, will do15:31
T3CHKOMMIEHey guys, I am trying to configure SNMP v2 on my ubuntu server so that my Zenoss device can manage and monitor it. But for some reason i cant get snmp to work... or even do a MIB walk on it. anyone have any ideas or links for setting up snmp 2 on ubuntu server 12.04? thanks.15:45
hallynjamespage: can you test with qemu-kvm from ppa:serge-hallyn/virt for grub timeout?15:52
=== matsubara is now known as matsubara-lunch
hallyn(I can't test on the intel right now, and the amd doesn't reproduce :)15:52
jamespagehallyn, LGTM - 2000 second countdown looking OK15:54
hallynjamespage: \o/15:55
hallynthanks.  lemme check release schedule to see if i should sned it to -proposed or wait for freeze to end15:56
jamespagehallyn, I'd just hold it until freeze lifts15:58
jamespagewon;t be that long now methingks15:58
hallynjamespage: is that today?15:58
jamespageyep15:59
hallynstill waiting for page to load15:59
jamespagebut not yet15:59
hallynok, cool.15:59
hallynthx - ttyl :)15:59
hallynstgraber: alas, setting up read-only root in lxc is problematic.  if we do it before pivot_root, we can't create and remove the put_old dir.  If do it after, we need to at least provide lxc-start with the ability to remount / ro, which we don't want16:21
hallynat least, we don't want that until we can tell apparmor to only allow that after pivot_root16:22
hallynjjohansen: ^ there was some sort of labeling coming down the pipline which should allow that right?16:22
stgraberhallyn: Well, we never want to allow the container to remount / read-only, though remounting it read/write shouldn't be a problem right?16:25
hallynstgraber: right, but this would be lxc-start mounting it readonly16:26
hallynbut still , it's '/'.  i dont' want it being able to muck with that16:27
stgraberyeah...16:27
hallynunless we can label it beofre pivot_root, and allow it by label16:27
hallynso i think i'll finish the code minus apparmor and stash it in a bzr tree for now16:27
hallynuntil we can talk with jjohansen16:27
Davieyroaksoax: Remember B1 doesn't include your latest upload.  Can you work with smoser to get refreshed released notes?16:28
Davieyta16:28
hallynstgraber: stashed in lp:~serge-hallyn/ubuntu/quantal/lxc/lxc-guest-start-roroot for now16:31
jjohansenhallyn: not for quantal, there will be a new version with things you want post quantal16:32
smoserjibel, http://paste.ubuntu.com/1189241/16:32
smoserany ideas on that ?16:32
roaksoaxDaviey: sure16:32
roaksoaxsmoser: ^^16:32
roaksoaxsmoser: what are the release notes so far regarding to maas?16:33
smoserroaksoax, ""16:33
smoserthose are the release notes at https://wiki.ubuntu.com/QuantalQuetzal/TechnicalOverview/Beta1#preview reguarding maas16:33
hallynjjohansen: and that will let us explicitly label a path and refer to it by label after mungling the mounts tree?16:34
roaksoaxsmoser: "New version of MAAS is available. This realease drops the usage of maas-provision, a subset of cobbler. Additionally, it provides its own DNS/DHCP server" or similar16:34
jibelsmoser,  probably because the build was marked 'Ready to release'16:35
jjohansenhallyn: I don't know whether explicit labeling will make it yet but hopefully16:35
jibelthere are nicer ways to handle that than a traceback16:35
smoserjibel, ok. that makes sense i guess.16:35
smoserwell, yeah, but that sfine.16:35
hallynjjohansen: great - thanks16:36
smoseri was just testing some change sto my update script16:36
jjohansenhallyn: however if its a file handle, delegation would be able to handle it16:36
roaksoaxsmoser: known issues "The DHCP config file does not get written. (LP: #1046397)" -= Already fixed, waiting for the upload16:36
hallynstgraber: i want to write a lxc.conf(5) section on the hooks this afternoon.  (so i'll probably wait to push lxc until then)16:36
roaksoaxto be approved16:36
smoserroaksoax, ok. i'll add that.16:37
hallynjjohansen: it's not.  i wonder if there is a way to use one for our purposes16:37
smoserwe should try to put together some 2 or 3 sentance blurb on maas.16:37
roaksoaxsmoser: thanks16:37
smoseras a feature / notable thing16:37
roaksoaxsmoser: better yet "New version of MAAS is available. This realease has dropped the usage of maas-provision, and features its own, MAAS managed, DNS and DHCP server"16:39
hallyn(moving to other laptop, biab)16:40
smoserroaksoax, gracias.16:41
smoseri'll update.16:41
stgraberhallyn: ok16:44
=== matsubara-lunch is now known as matsubara
RoyKanyone that knows a good alternative to zimbra? I'm rather fed up with upgrade paths and whatnot17:07
viezerdRoyK: Zarafa17:10
hallynstgraber: manpage pushed to ubuntu:lxc.  If you'd like to proofread, I'd appreciate it17:10
stgraberhallyn: ok, will do17:11
hallynstgraber: oh no, did some of my commits get lost?  did i not push to ubuntu:lxc with 0210-lxc-destroy-rm-symlink: ?17:12
hallyngotta do lunch, will figure it out after.17:13
stgraberhallyn: I pulled before pushing 0210-fix-debian-templates and it wasn't there, so it looks like you forgot to push that one to ubuntu:lxc17:13
hallynstgraber: ok will push it after lunch :)17:14
uvirtbotNew bug: #1046946 in cloud-init "None in cfgmnts second field gets converted to string" [Undecided,Confirmed] https://launchpad.net/bugs/104694617:21
=== mcclurmc is now known as mcclurmc_away
=== cpg|away is now known as cpg
=== arrrghhhAWAY is now known as arrrghhh
zulsmoser: ping cloud-init has ports.ubuntu.com right?17:38
stgraberhallyn: manpage looks good. I'm assuming you'll remove the "not yet implemented" when pushing to your git branch.17:43
hallynright, and in pkg if/when FFE is approved17:44
stgraberhallyn: doh, seems like you commited right before me ;)17:47
* stgraber pulls again17:47
stgraberhallyn: pushed17:48
stgraberhallyn: I granted you the FFe for pre-mount17:49
=== groupcat is now known as mrmist
stgraberhallyn: isn't bug 1046117 fixed by your last commit? I don't see the bug number in the changelog, so not sure.17:52
uvirtbotLaunchpad bug 1046117 in lxc "btrfs via symlink not working" [Medium,Triaged] https://launchpad.net/bugs/104611717:52
smoserzul, in quantal it should17:52
zulsmoser:  cool...just updating blueprints17:53
=== arrrghhh is now known as arrrghhhAWAY
hallynstgraber: g'ah!  yes i didn't put the bug# in changelog17:56
hallynstgraber: by pushed, you mean to archive?  (not seeing that in email or rmadison yet)17:57
hallyn(if no then it's fixable :)17:58
stgraberhallyn: pushed as in commited to ubuntu:lxc. Before uploading, please apply the pre-mount change now that the FFe has been approved18:00
hallynstgraber: phew, and add bug#1046117 to changelog.  will do18:01
hallynstgraber: would you say it's a half-assed job to NOT have an option in the ubuntu template to automtically create an ecryptfs-backed container?18:24
hallynOr would you say doing so is teetering off the edge of the FFE?18:24
stgraberhallyn: well, what I granted was an FFe for adding the hook to LXC, not for adding the ecryptfs option (which was just a valid use case for the hook)18:26
stgraberhallyn: so if you want to also push the ecryptfs option, I'd recommend another FFe18:26
hallynstgraber: sensible.  thx18:27
=== arrrghhhAWAY is now known as arrrghhh
=== arrrghhh is now known as arrrghhhAWAY
=== arrrghhhAWAY is now known as arrrghhh
lifelesshallyn: ecryptfs is a new feature isn't it ?18:53
lifelesshallyn: nvm, saw the rest of the discussion :)18:54
uvirtbotNew bug: #1040626 in keystone "Update user's default tenant partially succeeds without authz" [Critical,Fix committed] https://launchpad.net/bugs/104062619:01
sbeattiezul, can you or adam_g sponsor my keystone upload for bug 1040626?19:12
uvirtbotLaunchpad bug 1040626 in keystone "Update user's default tenant partially succeeds without authz" [Critical,Fix committed] https://launchpad.net/bugs/104062619:12
zuladam_g: do you wanna take this?19:12
Davieyerm19:13
Davieyzul: remember we are in freeze.19:13
zulDaviey: yeah but its a security update19:13
Davieysbeattie: If this can wait until Friday, it can be part of the normal upload set19:13
Davieyis that acceptable ?19:13
sbeattieDaviey: yeah, totally cool19:13
Davieyzul: quantal doesn't promise security rapid updates, right sbeattie ?19:14
sbeattieno, not really, not while in development.19:14
sbeattiefriday is fine; I just want to make sure it's on your team's radar.19:15
Davieysbeattie: thanks!19:17
Davieysmoser: thanks for the overview updates19:20
smoserlooking at maas test ephemeral now.19:21
adam_gzul: sure19:23
adam_gill at least look at it19:23
smoserhallyn, around ?19:28
smoserutlemming tells me that the 'admin' group is deprecated post 12.04 (or possibly *in* 12.04).  IIRC when kvm is installed it installs members of admin into kvm.19:28
smoserour cloud-images do not have ubuntu in 'admin' any more, so additional action is then required to get them into kvm.  that said, if i *did* recall correctly that kvm looks at members of the admin group to add them to kvm group automatically, then it should probably update that to look at sudo19:29
smoseror maybe not19:29
smoserDaviey, ping19:31
smoserhttps://bugs.launchpad.net/ubuntu/+source/linux/+bug/1031090 . i'm seeing that (i htink) on canonistack19:31
uvirtbotLaunchpad bug 1031090 in linux "kvm_intel not loadable in a quantal guest" [High,Fix released]19:31
smoserhttp://paste.ubuntu.com/1189514/19:32
Davieysmoser:19:35
Davieysmoser: you will, there are both arches... I bisected the kernel to try and find the commit via canonistack19:36
Davieysmoser: i am assuming that canonistack hasn't rebooted into the new kernel today19:36
Daviey(note the released date)19:36
hallynsmoser: checking19:37
hallynsmoser: in q it d19:38
Davieyadam_g: hey.. your horizon branch.. did you forget a bzr add?19:38
hallyndoesn't do that at all19:38
smoserhallyn, so it just doesn't add any users, you *have* to add users to kvm group19:39
smoserthats fine.19:39
hallynin precise too, it looks like.  still checking19:39
adam_gDaviey: whihc file? let me check19:40
adam_gDaviey: dont think so. my local branch is clean19:40
Davieyadam_g: but your diff doesn't seem to include everything?19:41
Davieyadam_g: hmm.. seeing more than i did this morning.. shrug19:43
adam_gDaviey: strange.19:44
adam_gDaviey: so that lets us ship something that works with compression disabled, and user install node-less, enable compression and it just works. im going to push the patch upstream and see if we can enable that ability there19:45
Davieyadam_g: Hmm.. why do we need to have pre-generated css... is that because it's statically served ?19:46
sarthorHi, I searched for the appropriate howto on the Internet, but I am not lucky to find some guide that cover my needs, What I Want! I want to host my real domain and website on my locale computer, While my IP is dynamic, I am having 8 mb DSL line, I want to do this stuff on ubuntu-server, Need help please.19:47
adam_gDaviey: yes, with compression disabled it needs to be statically served. when enabled, it invokes lessc to compile horizon.less into a dynamic .css that gets served19:47
Davieyadam_g: but still runs throug django?  or direct from apache?19:48
adam_gDaviey: its served from the static url via apache. its generated by django -> compressor -> lessc -> /usr/share/openstack_dashboard/static/dashboard/css/$some_hash.css19:48
adam_gthe first request generates it19:49
Davieyahh19:49
Davieyadam_g: it's not possible to serve uncompressed css?19:50
Davieyie.. inefficient css..19:50
Davieyadam_g: I hoped, django.conf.settings.COMPRESS_ENABLED = False.. would mean that it would just work.. without needing a static file shipped19:51
adam_gDaviey: its not necesarilly compressed, horizon.less  is  compiled using lessc from a library of less snippets.19:53
Davieyadam_g: right.. but the unanswered question.. does django-compressor allow serving of the snippets?19:56
Davieyie, by using inheritance imports within the css.. rather than flattening it?19:57
Daviey(without node)19:57
sarthorHi, I searched for the appropriate howto on the Internet, but I am not lucky to find some guide that cover my needs, What I Want! I want to host my real domain and website on my locale computer, While my IP is dynamic, I am having 8 mb DSL line, I want to do this stuff on ubuntu-server, Need help please... my webserver is running and UP,19:59
arrrghhhsarthor, we can help - but honestly if you look at things like overall cost and return on investment a hosted solution is MUCH better20:00
arrrghhhbecause what if your ISP doesn't allow port 80 traffic?  or if your uplink is slow?  perhaps if your internet goes down?  all these things will cause your website to fail to customers.20:00
sarthorarrrghhh: I am a learner, I have proxmox installed, and really I can play arround.20:01
arrrghhhok then what's the issue assuming you realize the limitations?20:01
sarthorany my ISP allow that, because I can brows via my IP.20:01
adam_gDaviey: i dont believe,  when the snippets shipped wiht source are in the .less format and not .css.20:02
arrrghhhsarthor, ok?  so you need a domain name?  have you purchased one or are you going to use something like dyndns?20:02
sarthorarrrghhh: I have already.20:02
sarthorarrrghhh: I have a dyndns paid account too. and i have a purchased domain too on sitegrouend registrar20:03
Davieyadam_g: yes.. but my hope was that django-compressor would DTRT, and render uncompressed... I guess i was too hopeful?20:04
arrrghhhsarthor, so... what's the issue?  please ask a question.20:04
adam_gDaviey: that is true of javascript, but not css20:05
Davieyah, ok.20:06
Davieyadam_g: where did you learn that20:06
Davieyadam_g: just by doing it?  or reading something?20:06
sarthorarrrghhh: how to tell to word to go to my local computer for http://mydomain.com?20:06
Davieyadam_g: incidentally, django-compress seems to have been superseeded by django-pipeline20:07
Daviey\o/20:07
arrrghhhsarthor, how to tell to word?20:07
sarthorarrrghhh: I do not have static IP.20:07
adam_gDaviey: that is true of javascript, but not css. i guess lessc can also minify it20:07
sarthorarrrghhh: world*20:07
adam_gDaviey: for css it uses an external tool (configurable, in this case node-less) to convert something into CSS and caches it by hash, then serves that $hash.css statically until it needs to be regenerated.20:08
adam_ger20:08
arrrghhhsarthor, a static IP is going to be very helpful - but if you don't have one, you'll have to get a service like dyndns (which you said you have?) to update your IP20:08
adam_gDaviey: just spent time yesterday putting together that patch and trying to understand what we're trying to solve :)20:08
arrrghhhso whenever your IP changes, the dyndns client will update your web site20:08
sarthorarrrghhh: I send you private messange20:09
sarthorI sent *20:09
Davieyadam_g: so, django-pipeline supports node/less AND YUI.. I wonder if it's drop in alternative ?20:11
adam_gDaviey: eek20:12
DavieyOh wait, that is a compiler, rather than a compressor20:12
Davieyadam_g: http://django-pipeline.readthedocs.org/en/latest/index.html20:12
arrrghhhsarthor, FYI, you should typically ask first before PMing someone - ask if it's OK that you PM them.20:12
DavieySo.. CSS needs a *compiler* to make it use uncompressed stuff.20:12
adam_gDaviey: yes, there are others out there but i believe we're dealing with a lessc specific source20:13
sarthorarrrghhh: I am sorry, Next I will not do like that. I am sorry20:13
arrrghhhno worries.  just IRC etiquette.20:13
Davieyadam_g: ok20:15
adam_gDaviey: talking to someone who knows more than me. might be an alternative to carrying a patch, but either way i think we'll still need to ship something compiled20:16
Davieyadam_g: ok, yeah.. I wonder if we can do something smart to notify us if we need refresh it?20:18
Davieyie, as part of the CI.. do a diff of our static one.. and see if it is the same?20:18
Davieyadam_g: we should probably also start looking at python-django-pipeline to do this, rather than -compressor which seems to be deprecated20:19
stgraberhallyn: http://paste.ubuntu.com/1189612/ now in python :)20:29
hallynstgraber: wow - awesome - fast!20:33
hallynhas beta been released yet?  i'm itchin to push lxc and qemu-kvm :)20:34
stgraberhallyn: yep, archive is unfrozen20:34
hallynyay20:38
stgraberhallyn: oh, just realized that with the code I'm playing with, my "ephemeral" containers can be persistent ;)20:44
stgraberhallyn: as in, you can start them using overlayfs, then shut them down, then start them again20:44
stgraberthe overlay will just re-assemble at start up time20:45
* stgraber adds a --keep-data option to bypass the post-stop hook wiping the container20:45
hallynstgraber: excellent20:55
uvirtbotNew bug: #1047040 in qemu-kvm (main) "kvm crashed - memory corruption" [Undecided,New] https://launchpad.net/bugs/104704020:56
stgraberhallyn: hmm, looks like I'll need that new LXC to test my code ;) I kinda need pre-mount hooks for what I want to do21:01
hallynstgraber: pushed about 5 mins ago, should be building21:01
=== gary_poster is now known as gary_poster|away
stgraberhallyn: cool21:03
stgraberhallyn: I'm going to need to allow overlayfs and aufs in the start-container profile, though that should be safe anyway21:35
hallynstgraber: np.  you can see from my blog post i also needed to add ecryptfs to that profile21:41
hallynstgraber: the lxc blueprint has me down for a server guide section on the API.  Uh, should i postpone that, or put your name by it?21:43
stgraberhallyn: I wouldn't spend too much time on documenting the API at this point, though mentioning that it exists and that a python module is now shipped with an example (in /usr/share/doc/python3-lxc/examples/) might still be worth doing21:46
hallynall right21:47
hallynHopefully I'll have time next week to work on the server guide.21:47
ScottKhttp://en.wikipedia.org/wiki/File:LART.png21:53
thatotherguyI'm a newb with apache22:10
thatotherguyI'm trying to setup an alias for a site that I just restored from backup and I don't have a /pub directory but the original server config did.  Do I just create a /pub directory and use alias in apache config to redirect to the real new location?22:12
thatotherguyI redirected but didn't create a /pub directory22:17
thatotherguyseems to work but have been getting transient errors22:17
stgraberhallyn: still around?22:21
stgraberhallyn: I'm getting something really weird here22:21
hallynstgraber: i was walking away :)22:22
thatotherguyAnyone good with apache22:22
stgraberhallyn: I'm not sure of exactly what the issue is, but after I set lxc.rootfs and a bunch of other keys, call save_config() and start(), I'm getting part of the path duplicated22:23
stgraberhallyn: basically /var/lib/lxc/tpl-precise-amd64-jb0qqn becomes /var/lib/lxc/tpl-precise-amd64-jb0qqn-jb0qqn22:23
hallynin the written-out configuration file?22:25
stgraberhallyn: if I drop my start() call and just use lxc-start manually, it works fine. So it looks like start() corrupts the config and writes it to disk again22:25
stgraberbefore starting with the wrong config22:25
hallynhm, crud22:26
stgraberhallyn: http://paste.ubuntu.com/1189813/ is what I'm using22:27
stgraberhallyn: running it will get you a failure with the "corrupted" /config in /var/lib/lxc/<name>22:27
thatotherguyAnyone good with apache22:27
stgraberhallyn: if you comment the part of the code calling start() and check the files, they'll be fine and lxc-start -n <name> will work fine22:27
=== arrrghhh is now known as arrrghhhAWAY
hallynhm, i don't see why22:30
hallynwas worrying that &conf was being re-initialized, bu it's not22:30
stgraberhallyn: can you reproduce the issue on your side? (want to check that it's not my system being completely broken somehow ;))22:32
hallynstgraber: feh, why do i have to enter my credentials to get the plaintext version22:33
hallynand then your name crashed python22:34
stgraberhallyn: you need python322:34
stgraberand if it still crashes under python3, then use LANG=C.UTF-8 ;)22:35
stgraberthough I'd think your current LANG is en_US.UTF-8 which is perfectly capable of parsing my name with python3 :)22:35
uvirtbotNew bug: #1047080 in geronimo-jms-1.1-spec (main) "Jar names differ between Debian and Ubuntu builds" [Undecided,New] https://launchpad.net/bugs/104708022:35
hallynstgraber: didn't reproduce, actually22:43
hallynstgraber: though permissions are too tight...22:44
hallynoh wiat22:44
hallynright it does fail to start. the config file is fine, but presumably in memory it's bad?22:45
hallynstgraber: I gotta run soon, but will try to figure this out22:46
stgraberhallyn: what do you have for lxc.rootfs in config?22:46
hallynlxc.rootfs in config *file* is fine22:46
hallynbut it did fail to start22:46
hallynlxc.rootfs = /var/lib/lxc/p1-6jgceb/rootfs22:46
stgraberoh, I guess you're missing hte apparmor entries22:46
stgraber  # required by lxc-start-ephemeral22:47
stgraber  mount fstype=overlayfs,22:47
stgraber  mount fstype=aufs,22:47
stgraberin /etc/apparmor.d/abstractions/lxc/start-container22:47
stgraber+ sudo /etc/init.d/apparmor reload22:47
hallynd'oh22:47
joemyfriendI'm having problem with my apache alias config22:48
joemyfriendAnyone able to help?22:48
=== cpg is now known as cpg|away
=== cpg|away is now known as cpg
hallynstgraber: d'oh!22:49
hallynnetns bug22:49
hallyni'll have to reboot before i try again,22:49
stgrabergah...22:50
hallyngah indeed22:50
stgraberthat bug is getting annoying :)22:50
hallynyeah22:50
hallynit's gonna be a rough month - lots of subtle bugs to squas in q22:50
hallynall right, i'm leaving for some dinner.  i'll reboot while gone, and retry tonight22:50
hallynso far really seeing no reason this should be happening to you :(22:50
hallynttyl22:51
=== cpg is now known as cpg|away
=== cpg|away is now known as cpg
oApocalypseminidlna support needed23:06
oApocalypsei have it up and running on my ubuntu server version 1223:08
oApocalypsefour folders shared in the config file23:08
oApocalypseonly the folder that is on my os drive is showing up the other three are on my external hard drive everytime i try to restart the service it reads directory not found23:09
oApocalypseyou guys make me sad there are far to many people here not to help a poor uneducated linux user23:11
protoCall7Hi all, on my new Ubuntu 12.04 server machines, I've noticed a ton of messages in the logs from SSHD: Sep  6 15:02:22 rarch-002 sshd[2919]: Connection closed by 10.30.0.123 [preauth]23:16
protoCall7we're not noticing any problems with sshd, can anyone verify that these are normal?23:16
protoCall7I would like to add them to log check.ignore if there isn't a problem23:17
oApocalypselol no one is here23:17
protoCall7lol there are plenty of people here, they're just occupied with other things, and no one without a direct answer for your question is going to BS you23:18
protoCall7just be patient and you'll get an answer ;)23:18
protoCall7sometimes it takes me a few hours, so I just ask, and get back to working on other things23:18
oApocalypseyea ive been watching tv :P23:19
oApocalypseyou know anything about dlna services23:19
protoCall7as far as your problem above, I haven't used dlna, but are you only restarting the minidlna service, or the whole machine?23:19
oApocalypserestarting the service23:20
protoCall7interesting.  as I said, I don't know anything about that service, but my first guess was going to be that you needed to put entries into /etc/fstab for the external drive, but unless that service is unmounting the drives for some reason, that wouldn't be the problem :-/23:21
oApocalypseyea i have the drives shared currently23:21
oApocalypsethrough samba23:21
oApocalypseive been watching said items off the hd from multiple computers23:21
protoCall7:)23:22
oApocalypseand its not giving me permission issues23:22
protoCall7I can see where losing those drives every time you restart the service would get mighty annoying though23:22
oApocalypsequite23:22
oApocalypsebut the shared drives are still reachable just not through minidlna23:23
protoCall7thats a strange one.  Once you're in that state, what do you have to do to get minidlna to see them again?23:24
oApocalypseproblem is they dont see them at all23:24
oApocalypseit sees the one folder off my os drive23:24
oApocalypsebut none of the three on my external23:24
oApocalypsemaybe its a discrepency in the path that i am providing23:25
oApocalypsebrb23:25
oApocalypseif its on another drive would i have to designate?23:29
protoCall7ok, so I'm a bit confused on one point here.  Is this external drive hooked up to the same machine that is running minidlna (e.g. a USB ext hdd), or is this like a samba share from another box?23:30
oApocalypsesame box23:30
protoCall7ok, and is the disk mounted?23:30
oApocalypsehd is in the server23:30
oApocalypsenot external23:31
oApocalypsemisspoke23:31
protoCall7ok, so to be clear, the external drive is connected to the same machine that minidlna is running on, however minidlna is not sharing the data from the external drive, only the one folder you have configured on the internal drive?23:32
oApocalypsecorrect23:32
protoCall7okay, and is the external drive mounted?23:33
oApocalypseand i am truly a novice at this23:33
oApocalypseyes23:33
protoCall7ok, what path is it mounted to?  likeā€¦ /mnt/externaldrive or something of that sort?23:33
oApocalypsea command i can put in to show what ur looking for23:34
oApocalypse?23:34
protoCall7just type mount and copy me the output to start :)23:34
protoCall7no arguments or anything23:35
oApocalypsedev/mapper/Exousia-root on / type ext4 (rw,errors=remount-ro)23:35
oApocalypseproc on /proc type proc (rw,noexec,nosuid,nodev)23:35
oApocalypsesysfs on /sys type sysfs (rw,noexec,nosuid,nodev)23:35
oApocalypsenone on /sys/fs/fuse/connections type fusectl (rw)23:36
oApocalypsenone on /sys/kernel/debug type debugfs (rw)23:36
oApocalypsenone on /sys/kernel/security type securityfs (rw)23:36
oApocalypseudev on /dev type devtmpfs (rw,mode=0755)23:36
oApocalypsedevpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=0620)23:36
oApocalypsetmpfs on /run type tmpfs (rw,noexec,nosuid,size=10%,mode=0755)23:36
oApocalypsenone on /run/lock type tmpfs (rw,noexec,nosuid,nodev,size=5242880)23:36
oApocalypsenone on /run/shm type tmpfs (rw,nosuid,nodev)23:36
oApocalypse/dev/sda1 on /boot type ext2 (rw)23:36
oApocalypse/dev/sdb1 on /Shared type ext4 (rw)23:36
lifeless!pastebin | oApocalypse23:36
ubottuoApocalypse: For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imagebin.org/?page=add | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic.23:36
oApocalypsesorry :(23:36
protoCall7ok, so /Shared is your external drive, correct?23:36
oApocalypseyes23:37
protoCall7okay, and can you paste the relevant lines from the configuration file for minidlna where you've configured that share? (using paste bin this time ;))23:37
oApocalypselol23:37
oApocalypsedont wanna get yelled at again23:37
oApocalypsehttp://paste.ubuntu.com/1189890/23:40
oApocalypsedoes that seem correct23:43
protoCall7except for line 1623:43
protoCall7this won't fly because of the space: /Shared/Share/Torrents/TV Shows23:44
oApocalypsego on23:44
protoCall7try writing it as follows:  /Shared/Share/Torrents/TV\ Shows23:44
protoCall7to escape that space23:44
oApocalypseshould i just change the name of the folder to TV_Shows23:45
protoCall7that would make it easier any time you need to use that directory from the command line, that's for sure23:45
oApocalypseok23:46
oApocalypsebut why wouldnt line 14 and 15 work23:46
protoCall7those look right to me, by the example given in the config.  Is there anything relevant in the logs?23:46
oApocalypsethere are logs?23:47
oApocalypselol want me to give you the error message ?23:47
protoCall7yes pls :)23:47
oApocalypsehttp://paste.ubuntu.com/1189900/23:50
protoCall7ls -lsa /Shared/Share/Torrents23:51
protoCall7please23:51
oApocalypsehttp://paste.ubuntu.com/1189902/23:53
protoCall7sudo chmod +r /Shared/Share/Torrents/Movies23:54
protoCall7then restart that service again and see if it still complains about Movies23:54
protoCall7my guess is that minidlna is starting as a non-priviledged user, and doesn't have read permissions on that directory23:55
oApocalypsestill issue23:55
protoCall7Ohh, the directory above it is accessible by root only as well23:55
protoCall7ps aux | grep dlna23:55
oApocalypsethat does?23:56
protoCall7lets just make sure that it's running as a non-priv user before i go having you mess with permissions23:56
oApocalypseminidlna runs an unknown user23:57
protoCall7that is going to show all running processes on the machine along with who is using them23:57
protoCall7yup, so if its not running as root, it doesn't have read permissions to access those dirs23:57
protoCall7you could always do a:23:58
protoCall7sudo chmod -R +r /Shared23:58
protoCall7which will enable read access for that entire directory and everything in it23:58
protoCall7if you need to limit read access to that dir, you'll have to be a bit more selective23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!