[07:15] <lordievader> Good morning.
[07:44] <thewall> Hello. I need help with ubuntu cloud-archive packages for Juno. Can anyone please help me? :)
[10:32] <ruiorey> hi, we have a problem in our server, we wrongly manually replaced some files in usr/lib and as result we had several errors we have been solving. the actual one is error while loading shared libraries: libpthread.so.0: cannot open shared object file: No such file or directory...we try to run apt-get and get errors installing removing etc...because almost all processes depend on it. any suggestion would be very welcome
[10:33] <bekks> Restore your backup.
[10:34] <ruiorey> no backup.. :/
[10:34] <bekks> So you have no data worth to be kept on your server. Reinstall it.
[10:35] <maswan> Set up a system with identical package set installed. Boot the server off rescue media. Copy relevant files from the temporary one.
[10:35] <ruiorey> we have
[10:35] <bekks> ruiorey: You dont, since you have no backup.
[10:36] <bekks> ruiorey: All you can do now is boot a live cd, copy all files you still need, and reinstall the server.
[10:40] <ruiorey> maswan, do you think direct copy from a similar system will work?...in this case libpthread.so.0 just needs to be replaced?
[10:40] <bekks> ruiorey: No. :)
[10:40] <maswan> ruiorey: It might work. For a while.
[10:40] <bekks> ruiorey: Copy all files you want to keep, and reinstall the server.
[10:40] <maswan> ruiorey: But you'll never get to the point where the server will be a reliable server
[10:41] <ruiorey> ok thanks maswan
[10:41] <ruiorey> and why is not going to be reliable?
[10:41] <maswan> Because how can you make sure that you got all the files, and got exactly the right versions?
[10:41] <ruiorey> ok
[10:42] <ruiorey> thanks
[10:44] <arcsky> shortcut , ctrl + w deletes a word does anyone know if you can delete a word if you are infront that word instead if the backend?
[10:47] <TJ-> arcsky: you're talking about readline? In bash "kill-word" is meta+d
[10:50] <arcsky> TJ-: ctrl+d deletes just one char
[10:51] <histo> arcsky: in vim?
[10:52] <TJ-> arcsky: I said Meta+d not Ctrl+d
[11:08] <arcsky> histo: in shell
[11:09] <histo> arcsky: yeah alt+d
[11:11] <arcsky> histo: perfect thanks!
[13:53] <fornax> I I'm running docker on my ubuntu server and want to bind docker to a specific virtual nic. Can one explain me how to create a virtual nic and add it to my bridge br0? The nic will get a ip from my providers dhcp server and shall allow that docker only exposes its ports to the nics ip
[14:44] <bekks> fornax: AFAIK, you cannot use DHCP on a bridge.
[14:45] <jrwren> I missed context, but I've been using DHCP on bridge for a decade+
[14:46] <jrwren> i don't know docker well enough to help. sorry.
[15:10] <lordievader> bekks: Why not?
[15:12] <bekks> lordievader: because a bridge connects two networks, and both networks may offer dhcp, which would lead to a race condition.
[15:13] <lordievader> Doesn't a bridge have one gateway? I understand your reasoning, however I'm using dhcp on my bridge. True, it only has one physical connection.
[15:24] <jrwren> that is nonsense.
[15:25] <jrwren> every port on a switch is a bridge. we use dhcp with switches all the time.
[15:32]  * pmatulis pulls up a chair
[15:38]  * teward yawns.
[15:38] <teward> is the server team still meeting today?
[15:46] <fornax> bekks: I already use DHCP on my bridge and 2 lxc containers. But now I want to create one more simple virtual ethernet device (veth) manually inside an ubuntu config file to get docker bound to this veth
[15:59] <mgz> openstack charmers, can someone take a look at bug 1474291 - if I'm parsing it correctly there's an issue in the ceilometer charm rather than juju
[16:00] <mgz> the config-changed hook should be idempotent, juju can and will call config changed when you're not expecting it
[16:04] <qman__> bekks, lordievader: bridges can absolutely be configured dhcp, though it may not work properly if they're not on the same network
[16:08] <smoser> smb, https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1473527
[16:09] <smoser> my general question to you, is ... assuming functional /dev/console (ie, one tied to an actual serial device) ..
[16:09] <smoser> should you be able to have multiple openers and concurrent writers ?
[16:10] <smb> smoser, Without further checking I would think multiple writers unlikely
[16:11] <smoser> i was unable to recreeate failure
[16:11] <smoser> well, in that serial-shell-looper attached there.
[16:11] <smoser> that basically opens up N sub-processes that just open write close to /dev/console
[16:13] <smoser> smb, so cloud-init has a general desire to write messages to the console... in an effort to have things logged. if in fact /dev/console (or /dev/ttyS* or whatever console=) are not multiple-writer friendly, how would you or anyone suggst such a thing occur
[16:20] <smb> smoser, well that was a first guess. might be the console device is different. But I would need to look into the code to be sure
[16:24] <ruiorey> bekks , maswan it is working thanks for the help..we installed a server version equal in a vm, downgraded the kernel on the vm and replaced the files in the server :)
[16:34] <bekks> ruiorey: And how did you ensure that you replaced ALL files that were corrupted?
[16:36] <ruiorey> i'm not sure but all services are running fine
[16:52] <smoser> smb, could you take a glance at that? i'm at least interested in knowing if i can hope to arbitrarily open /dev/console or its convigured device and write to it.
[16:52] <smoser> or if doing that is going to always be a race condition without some intermediary
[17:06] <linuxgeek_> hi, i have installed 14.04.1 on a vm
[17:07] <linuxgeek_> two nics eth0 and eth1 is added. both have ip's on different subnet and a different gateway
[17:08] <linuxgeek_> at any point of time, only one ip is accessible
[17:08] <linuxgeek_> how can i configure the routes or gateway so both the ip's is accessible
[17:43] <smb> smoser, So I think the answer to console is that may work but is not guaranteed. The other question would by why you insist on directly opening the device again. I assume that ssh-authkey-fingerprints is called from the systemd service which already re-directed output for you.
[17:43] <smoser> smb, 2 reasons
[17:43] <smoser> a.) i dont want to be dependent on systemd
[17:43] <smoser> b.) systemd prefixes output with job name
[17:44] <smoser> c.) i guess... how does one know that systemd wont fail to write there ?
[17:44] <smb> b is not a reason. that's an opinion
[17:44] <smoser> b most certainly is not opinion. it is fact.
[17:44] <smoser> maybe opinion that i dont like that.
[17:44] <smoser> but it is true for sure.
[17:45] <smb> Ok, agreed that it does. But some would rather say thats good
[17:45] <smoser> the other eason woudl be that i'd like to go to all console= parms on the command line. not just the one that gets assigned to /dev/console
[17:52] <smb> smoser, Which probably would require you to write code which emits stuff round-robin. And likely that should only open, write and close, plus handing open failure and do retries. And hope nothing else hogs the device.
[17:54] <smb> smoser, Unfortunately I cannot remember details but that problem feel familiar from somethign invilving upstart and plymouth and not being resolved there either...
[17:54] <smoser> yeah. :-(
[17:54] <smoser> and then there is the fun where writes to a non-existant /dev/ttySX work for a while (as they're buffered)
[17:55] <smoser> and then fail some time later.
[17:56] <smoser> thanks smb.
[20:21] <coreycb> jamespage, we need python-jsonschema for trove in the liberty cloud archive, that fixes those test failures
[23:06] <sarthor> HI, using ubuntu 14.04 mysql-server taking long time to start while ubuntu restarting.. help.
[23:22] <RoyK> sarthor: did the server crash before this bootup?
[23:23] <sarthor> RoyK no.
[23:23] <RoyK> sarthor: perhaps try #mysql - I don't like mysql - it's a pain - use postgresql if you want something useful (if you can)
[23:25] <sarthor> RoyK OK. Thank you,
[23:29] <teward> where do I report an oddity in the landscape dedicated server software?
[23:29] <sarnold> sarthor: check the log files.. often issues just show up there..
[23:30] <sarnold> teward: is it a recent / transient issue? or osmething that's been bugging you for a while?
[23:30] <teward> sarnold: something I only discovered now, not sure if it's something already known or not.
[23:30] <sarthor> sarnold, OK. Checking.. Thanks for reply sarnold
[23:30] <teward> (an oddity in license-to-use detection)
[23:30] <teward> s/license-to-use/registration-class-to-use/
[23:31] <sarnold> teward: aha; bug report might be okay, talking with a support rep or sales rep or whatever might be better. I'm not sure where to send you..
[23:31] <teward> sarnold: probably a bug report
[23:31] <teward> it's not with the ubuntu advantage program, it's the free personal-use one
[23:31] <teward> so not Cloud LDS, Standalone LDS
[23:32] <teward> sarnold: just an oddity - Lubuntu VM, Product Identifier shows as KVM, and it's detected as a computer, not a VM.  Which is... odd, I think
[23:35] <teward> sarnold: http://i.imgur.com/yQ9H5X0.png  <-- a snippet of what I see on Landscape.  I'll probably file a bug, but if you want to throw it towards their support reps feel free.  I get 0 support because i'm not paying for it (it's the standalone personal use)
[23:36] <sarnold> well, if you see it, chances are good someone else may see it too
[23:37] <teward> sarnold: it's odd, because I"m borrowing phillw's Lubuntu VM.  It's on KVM on a CentOS host, so maybe it's intentional to detect it as that?  (perhaps how open stack is supposed to mimick computers, theoretically?)
[23:37] <teward> although granted landscape has a dedicated openstack part