| smoser | genii-around, try nomodeset | 00:05 |
|---|---|---|
| smoser | as a kernel option. | 00:06 |
| genii-around | smoser: I'm looking into the video card, hard to tell what driver it normally would use | 00:07 |
| genii-around | http://pci-ids.ucw.cz/read/PC/1a03/2000 is not enlightening me | 00:08 |
| smoser | your problem is for server? | 00:08 |
| smoser | or desktop | 00:08 |
| firc | server | 00:10 |
| genii-around | smoser: They want server but video blanks out, even with vga set at boot. Desktop apparently loads fine but no raid support | 00:10 |
| genii-around | ( and yet their system is on the list of certified systems) | 00:11 |
| genii-around | firc: Did you try the nomodeset ? | 00:11 |
| firc | yeah, just did. same result. Blank screen | 00:12 |
| smoser | genii-around, try passing cmdline 'nomodeset' | 00:12 |
| firc | did that too. (F6) and added nomodeset | 00:13 |
| === n0ts_off is now known as n0ts | ||
| firc | I have to go now. But thanks for the help genii-around and smoser. I'll try this again tomorrow. At least now I know its the video driver causing the issue | 00:16 |
| genii-around | firc: I'll be here all week 10am onwards EDT | 00:16 |
| firc | great. thanks! | 00:16 |
| === matsubara-dinner is now known as matsubara | ||
| === shantorn is now known as kd7jwc | ||
| === Shenne is now known as Shenni | ||
| delinquentme | Can curl be used to get the IP of the server which a particular domain name resolves to? | 02:17 |
| rext7 | just use nslookup | 02:21 |
| rext7 | nslookup www.whatever.com | 02:21 |
| sarnold | delinquentme: I don't see any command line options that sound like they'd do it in the --help listing... but you could use 'host' (which may return several potential IPs...) or you could use ping, which will do the lookup and then route a packet to the host... | 02:21 |
| sarnold | nslookup, blast from the past :) | 02:21 |
| rext7 | or "host www.whatever.com" | 02:21 |
| delinquentme | sarnold, turns out that googling <url> IP | 02:21 |
| rext7 | dig | 02:21 |
| delinquentme | does just fine | 02:21 |
| rext7 | whatever | 02:21 |
| sarnold | delinquentme: oof. if you don't mind using a web browser and an unrelated external service... | 02:22 |
| sarnold | .. and one that didn't work for me on the first one I tried. :) | 02:22 |
| delinquentme | sarnold, I tried ping .. i guess I was missing some option? | 02:22 |
| sarnold | oof, well, neither ping nor host nor nslookup nor dig will take an url, just the hostname | 02:23 |
| sarnold | $ ping wiki.ubuntu.com | 02:23 |
| sarnold | PING wiki.ubuntu.com (91.189.89.161) 56(84) bytes of data. | 02:23 |
| sarnold | ... :) | 02:24 |
| delinquentme | AHHH the slashes | 02:24 |
| delinquentme | ping dislikes them. | 02:24 |
| sarnold | yeah. turning an URL into a hostname may take a bit of ugly regex work :( | 02:25 |
| rext7 | echo "http://www.yahoo.com" | awk -F\/\/ '{print $2}' | 02:25 |
| rext7 | maybe? | 02:26 |
| sarnold | rext7: now handle :ports :) | 02:27 |
| sarnold | and username@passwords | 02:27 |
| sarnold | anyway, dinner for real now :) | 02:27 |
| Sprocks | has anyone set up an IRC server with nickserv and chanserv on ubuntu? | 02:40 |
| === n0ts is now known as n0ts_off | ||
| === n0ts_off is now known as n0ts | ||
| === n0ts is now known as n0ts_off | ||
| === n0ts_off is now known as n0ts | ||
| Gallomimia | hi. i just noticed avahi-daemon running on my ubuntu 12.04 server. It's a vds, and after researching what it is.... i really have to wonder where it came from (i just did some apt updates and rebooted) | 04:59 |
| CrypticS_ | it's a zero-config daemon | 04:59 |
| CrypticS_ | like bonjour for mac/windows | 04:59 |
| Gallomimia | this is a virtual server in a data center | 05:00 |
| Gallomimia | a question i have to ask, with no possible way to answer, is was it there before. | 05:00 |
| Gallomimia | the next question is, do i need it | 05:00 |
| Gallomimia | hm. two instances of it... | 05:01 |
| CrypticS_ | read the docs to answer the second question | 05:01 |
| === CrypticSquared is now known as numberz | ||
| === numberz is now known as CrypticSquared | ||
| Lavvy | Hello guys, someone should tell me how to locate openstack in ubuntu server install, cos ubuntu says it shipps ubuntu 12.xx with openstack. Please am newbei in ubuntu | 06:24 |
| sarnold | Lavvy: in general, 'apt-cache search foo' will search the package repositories you have configured for any packages matching 'foo' in the name or description; apt-cache search openstack shows a lot of related packages, hopefully you'll be able to sort out what you need from there... | 06:27 |
| Lavvy | Ok tanx | 06:28 |
| cornfeed | anyone around that has experience with setting up a serial console? | 06:34 |
| === smb` is now known as smb | ||
| === doko_ is now known as doko | ||
| SinZ | damn, ubuntu server 12.04 doesnt contain wireless keyboard drivers in the installer, meaning I cant even choose the language | 08:57 |
| SinZ | nvm, appears it was just that wireless keyboard | 09:06 |
| SinZ | The absence of WPA2-PSK worries me | 09:09 |
| === jtv1 is now known as jtv | ||
| AnAnt | Hello, I got a problem on slapd when upgrading from lucid to precise: it says that slapcat cannot be run | 09:20 |
| AnAnt | it seems that bug fix on #990742 didn't fix it ! | 09:21 |
| === zz_KristianDK is now known as KristianDK | ||
| AnAnt | ah, fixed it by installing lucid's libsasl2-2 & libsasl2-modules ! | 09:30 |
| harovali | I was in need to run linux kernel 3.2.x in Lucid, so I copied a sources.list from a precise 12.04 LTS installation, then I ran apt-get update and then apt-get install linux-image-3.2.0-31-generic-pae. This functioned well. The machine rebooted fine. However, the network did not get up. I went back to the original sources.list and removed the 3.2 kernel, and now I'm fine back at the Lucid backported kernel. What can ha | 10:22 |
| === cpg is now known as cpg|away | ||
| === n0ts is now known as n0ts_off | ||
| === n0ts_off is now known as n0ts | ||
| === n0ts is now known as n0ts_off | ||
| === n0ts_off is now known as n0ts | ||
| tobin | Anyone around that can help troubleshoot an issue with the latest precise cloud image? I'm trying to re-package the amd64 ebs backend image as a custom AMI. I keep hitting issues with grub. | 12:52 |
| === n0ts is now known as n0ts_off | ||
| === Ursinha is now known as Ursinha-afk | ||
| genii-around | Was the fellow firc from yesterday back in yet? He had the LSI megaraid on Dell Poweredge C6100 ( Ubuntu certified system ) that the video was not working with the installer even with nomodeset or using vga | 13:55 |
| * genii-around makes more coffee | 13:58 | |
| ppetraki | can I just create an empty container using juju so I can prototype a charm in that environment? | 14:10 |
| * ppetraki whoops wrong channel | 14:11 | |
| === n0ts_off is now known as n0ts | ||
| Belgarath | hi, anybody managed to install bfa brocade fiber channel scsi cards 415/815 on ubuntu? | 14:29 |
| Belgarath | I see the drivers loading but the FC stays dead | 14:29 |
| === shantorn_ is now known as shantorn | ||
| bubu\a | Hi guys, ubuntu 10.04 server I want to put in http_proxy into a user env. Where is best place to do this? ~/.bashrc ? | 14:51 |
| bubu\a | and if i put it in ~/.bashrc for the user | 14:52 |
| Pici | bubu\a: for one user, or for all users? | 14:52 |
| bubu\a | do I need to use http_proxy or export http_proxy ? | 14:52 |
| bubu\a | for one user.. | 14:52 |
| jpds | bubu\a: Either one. | 14:52 |
| bubu\a | what is the diff? | 14:53 |
| jpds | bubu\a: export will send the variable to sub-shells. | 14:53 |
| bubu\a | so if that user is using a ruby application to try and reach out to rubygems.org over a proxy....and the ruby application is configured to take the env variables for that user...it should be all good..? | 14:54 |
| jpds | bubu\a: Should be. | 14:54 |
| bubu\a | and if i wanted a specific address/IP not to be proxied | 14:55 |
| bubu\a | I can use the export no_proxy="ipnottoproxy" | 14:55 |
| bubu\a | ? | 14:55 |
| jpds | bubu\a: Yes. | 14:56 |
| bubu\a | ok cool | 14:56 |
| bubu\a | and when I use envprint | grep proxy | 14:56 |
| bubu\a | and I can see my http_proxy, https_proxy and no_proxy setting | 14:57 |
| bubu\a | looks like they have been applied OK .. | 14:57 |
| jpds | bubu\a: Only one way to find out if they work. | 15:02 |
| === n0ts is now known as n0ts_off | ||
| pitti | hello | 15:44 |
| pitti | I was setting up CanoniStack and juju, and "juju bootstrap" and "juju ssh 0" works fine | 15:44 |
| pitti | but if I try "juju deploy" with anything (tried postgresql and jibel's jhbuild charm), the new "instance-id" keeps being at "pending", and nothing happens | 15:45 |
| pitti | it's sitting like that for > 1 hour without any output in juju debug-log | 15:45 |
| pitti | any idea how I can debug that? | 15:46 |
| melmoth | pitti can you see the vm with nova list ? | 15:47 |
| melmoth | can you log in it with ssh ? | 15:47 |
| pitti | http://paste.ubuntu.com/1285076/ FYI | 15:47 |
| pitti | melmoth: ah, I don't have nova installed, but I can't see it in euca-describe-instances | 15:47 |
| melmoth | try to ssh in it. If you can, then you should be able to see what s going on with the charm install in /var/lib/juju/something | 15:48 |
| * pitti installs python-novaclient | 15:48 | |
| pitti | melmoth: there is no instance | 15:48 |
| melmoth | oh. | 15:48 |
| melmoth | not normal. | 15:48 |
| pitti | just the default one created by bootstrap ("0") | 15:48 |
| pitti | but any other instance that deploy wants to create is never instantiated apparently | 15:48 |
| melmoth | try to launch a vm manually, may be there s a problem there ? | 15:49 |
| pitti | melmoth: "nova list" also just has the default ("0") one, just like euca-describe-instances and juju status | 15:49 |
| melmoth | have also a look at euca-describe-groups | 15:49 |
| melmoth | i have met people who had so many fallout from previous bootstrap that they hitted the quota for security group | 15:49 |
| pitti | melmoth: http://paste.ubuntu.com/1285273/ | 15:50 |
| melmoth | that does not look like more than the default quota (dont know the numbe ron top of my head, but it more than that) | 15:50 |
| pitti | melmoth: disclaimer, I just started with that whole juju thing some two hours ago, so I'm still learning | 15:50 |
| pitti | melmoth: launch a VM manually> is that "euca-something"? | 15:51 |
| melmoth | pitti does it works if you nova boot an instance ? euca-run-instances -k $NOVA_USERNAME -t m1.tiny ami-000000bf | 15:51 |
| pitti | I don't see an euca-create-instance or similar | 15:51 |
| === Ursinha-afk is now known as Ursinha | ||
| pitti | melmoth: yes, I immediately get a new "pending" one | 15:52 |
| melmoth | if it works, check the ami you use in the environment.yaml is the same as the one you can boot instance manually | 15:52 |
| melmoth | if it is, try juju deploy with --verbose (or is it --debug ?) | 15:52 |
| melmoth | just in case there s some hint | 15:52 |
| pitti | melmoth: my environments.yaml doesn't have an ami name | 15:53 |
| melmoth | ahhh. This may be a problem. | 15:53 |
| melmoth | juju needs to know wich image to pick | 15:53 |
| pitti | melmoth: but the default one that "juju bootstrap" created is the same, ami-000000bf | 15:53 |
| pitti | melmoth: hm, how come that "boostrap" can figure it out, but not "deploy"? | 15:53 |
| melmoth | no idea. | 15:54 |
| melmoth | here, i have a default-image-id: something in my environment.yaml | 15:54 |
| pitti | default-image-id: bb636e4f-79d7-4d6b-b13b-c7d53419fd5a | 15:54 |
| pitti | melmoth: ^ I do have this | 15:54 |
| pitti | (sorry, I thought it's something with ami-...) | 15:55 |
| melmoth | hmmm, does not look like a euca- image id (looks like a nova one) | 15:55 |
| melmoth | but may be you are not using the ec2 type (type: ec2) so in this case it could be ok | 15:55 |
| melmoth | i dont know, i always use the ec2 type | 15:55 |
| === hazmat is now known as kapilt | ||
| === AaronMickDee_ is now known as AaronMickDee | ||
| melmoth | https://juju.ubuntu.com/docs/provider-configuration-openstack.html indeed, i think you are using the openstack provider | 15:56 |
| pitti | melmoth: right, I'm using openstack_s3 (that's canonistack, not EC2) | 15:56 |
| pitti | not amazon ec2, anyway | 15:57 |
| === AaronMickDee_ is now known as AaronMickDee | ||
| pitti | melmoth: I'll clean up the pending service and my test machine, and try again with --verbose | 15:57 |
| hallyn | smb: i could still be doing somethign wrong, but using mostly the same juju charm, using canonistack, in precise using openvswitch-datapath-dkms gre tunnels work for me, while in quantal using the upstream openvswitch module, they do not. | 16:03 |
| hallyn | (i'll get some tcpdump traces this afternoon to see if i can figure out where packets get lost) | 16:04 |
| hallyn | has anyone here used openvswitch gre tunnels in quantal (recently)? | 16:04 |
| smb | hallyn, You catch me without any context loaded and I am not sure I want to unload my current one... :) | 16:05 |
| pitti | melmoth: so, I just cleaned up everything and ran it again, and this time I see it in nova list, but still pending in juju status: http://paste.ubuntu.com/1285302/ | 16:05 |
| hallyn | smb: sorry, i thought i'd pinged you on openvswitch when i was having trouble before (which turned out to be amazon-specific) | 16:07 |
| hallyn | smb: keep your current context, i'll ping you when i get more data :) thx | 16:07 |
| melmoth | can if ssh in it ? If not what does nova console-logs fd129f09-4606-4b25-8c6f-1ef47cca2442 says ? | 16:07 |
| melmoth | pitti,once the machine has your ssh key (you ll see that on the console output), you can ssh in it | 16:08 |
| melmoth | from there, you can try to see if there some problem with the actual charm being deployed | 16:08 |
| smb | hallyn, Oh, hm, maybe but it might have been lost to me. Thanks, yeah best with a bug report (maybe subscribe me there and hopefully I notice any updates) | 16:08 |
| hallyn | smb: since module is now upstream, bug against linux is appropriate? | 16:09 |
| pitti | melmoth: ah, I can't yet, it seems (no public IP) | 16:09 |
| smb | hallyn, If it was upstream in quantal yeah, might be a start at least | 16:10 |
| melmoth | you should be able to ssh to the private one as well (with some ssh/config trick) | 16:10 |
| hallyn | smb: thx | 16:11 |
| melmoth | pitti https://pastebin.canonical.com/76737/ | 16:11 |
| pitti | melmoth: ah, thanks; I'm in | 16:13 |
| melmoth | \o/ | 16:13 |
| pitti | melmoth: so it seems juju status doesn't see this, as machine "5" is still "instance-id: pending"? | 16:13 |
| melmoth | may be the charm is still installing | 16:13 |
| pitti | melmoth: there's nothing running on that server | 16:13 |
| pitti | melmoth: and on debug-log I didn't see anything the whole time, including the 'deploy' command | 16:13 |
| melmoth | do you have a /var/lib/juju ? | 16:14 |
| pitti | melmoth: deploy --verbose -> http://paste.ubuntu.com/1285293/ | 16:14 |
| pitti | melmoth: /var/lib/juju> no | 16:14 |
| pitti | dpkg -l *juju* -> none | 16:14 |
| melmoth | a /var/log/cloud-init something ? | 16:14 |
| pitti | yes, cloud-init.log | 16:15 |
| melmoth | look in there for any error | 16:15 |
| pitti | http://paste.ubuntu.com/1285321/ | 16:15 |
| pitti | doesn't look like anything error-ish | 16:15 |
| melmoth | hmmm | 16:16 |
| pitti | melmoth: I can try again with another charm, would that help? | 16:17 |
| melmoth | it cannot hurt :) | 16:17 |
| pitti | postgresql didn't seem too happy either, is there a common one which really Ought To Work ™? | 16:17 |
| melmoth | mysql ? | 16:17 |
| melmoth | pitti can you add this key on 10.55.60.16 /home/ubuntu/.ssh/authorized_keys (and be sure it s chmoded 0600) ? | 16:18 |
| melmoth | https://pastebin.canonical.com/76738/ | 16:18 |
| melmoth | so i can log in, just in case | 16:18 |
| pitti | melmoth: done | 16:19 |
| melmoth | ok, i m in. let s dig | 16:20 |
| pitti | melmoth: oh fun, I issued "terminate-machine 5" several minutes ago, and it's still up | 16:21 |
| melmoth | huhu | 16:21 |
| pitti | melmoth: (I did this after we discussed "try again with mysql") | 16:22 |
| * pitti runs that anyway, shouldn't get in the way | 16:22 | |
| melmoth | i dont understand why there s no trace of juju whatshoewever | 16:22 |
| pitti | melmoth: in case it matters, it's the "juju" package from current (i. e. final) quantal | 16:23 |
| pitti | not precise | 16:23 |
| pitti | ran "mysql"; first problem again, machine 7: instance: pending, and nothing in "nova list" | 16:24 |
| pitti | nor in euca-describe-instances | 16:24 |
| melmoth | weird. | 16:24 |
| melmoth | i m wondering if i m logged in on a machine juju started or may be the one you launched manually instead | 16:24 |
| melmoth | because it was able to contact the metadata server but it does not even looks like it try to install juju | 16:25 |
| pitti | melmoth: it should be a juju-started one; I don't remember starting a machine manually, how would I do that? | 16:25 |
| melmoth | euca-run-instances or nova boot | 16:25 |
| pitti | oooh, sorry | 16:25 |
| pitti | right | 16:25 |
| pitti | that was it | 16:25 |
| melmoth | huhuh | 16:26 |
| melmoth | ok. logging off this one then. | 16:26 |
| pitti | melmoth: alright, so at least it's consistent -- juju deploy doesn't create machines, and instance-id: stays at "pending" | 16:26 |
| pitti | melmoth: killing that machine, to reduce confusion | 16:26 |
| pitti | gone | 16:27 |
| melmoth | so, looks like your juju cannot start instances | 16:27 |
| melmoth | let s try to boot one with nova, and the same image id | 16:27 |
| melmoth | nova boot --flavor 1 --image 45203e5e-7a6a-4493-aa24-ea783460cbf9 --key_name mykey myinstance | 16:28 |
| melmoth | something like that ^ | 16:28 |
| melmoth | just not sure what you should use for the mykey | 16:28 |
| pitti | juju status is clean again, terminated mysql service and machine | 16:28 |
| pitti | $ nova boot --flavor 1 --image 45203e5e-7a6a-4493-aa24-ea783460cbf9 myinstance | 16:29 |
| pitti | ERROR: No image with a name or ID of '45203e5e-7a6a-4493-aa24-ea783460cbf9' exists. | 16:29 |
| pitti | trying the one in my envirionments.yaml (bb636e4f-79d7-4d6b-b13b-c7d53419fd5a) | 16:29 |
| pitti | that seems to have succeeded | 16:30 |
| pitti | | 9697051e-adce-4df3-bceb-4101f12eb77b | myinstance | ACTIVE | canonistack=10.55.60.16 | 16:30 |
| pitti | melmoth: I didn't specify --key_name, it got it fine from the environment | 16:30 |
| === KristianDK is now known as zz_KristianDK | ||
| pitti | melmoth: I can't seem to ssh in, but at least the machine seems to exist | 16:32 |
| melmoth | and the machine booted with the same image you set in your environment.yaml ? | 16:32 |
| melmoth | hmmm | 16:32 |
| melmoth | i must admit i have no real clue what s going on there | 16:32 |
| pitti | melmoth: maybe I'll just tear it all down and restart from scratch | 16:35 |
| pitti | melmoth: thanks a lot for hand-holding! | 16:36 |
| melmoth | pitti, hey if i were you, i wold try the ec2 provider | 16:36 |
| melmoth | https://wiki.canonical.com/InformationInfrastructure/IS/CanonicalOpenstack | 16:36 |
| melmoth | just take the sample environment.yaml from there, adapt it to your setting, and try again | 16:36 |
| melmoth | good luck :) | 16:36 |
| pitti | melmoth: can do, if that's the preferred method | 16:37 |
| melmoth | i have no idea. but it s a "work for me" one :) | 16:37 |
| pitti | melmoth: I was following https://wiki.canonical.com/InformationInfrastructure/IS/CanonicalOpenstack/QuickStart | 16:37 |
| pitti | melmoth: I can use that with juju as well, presumably | 16:38 |
| melmoth | hmmm, yep, here it s says to use the openstack provider | 16:38 |
| * melmoth is kind of lost | 16:38 | |
| melmoth | heyy | 16:38 |
| melmoth | NOTE: Some of the above options have been deprecated. See, https://juju.ubuntu.com/docs/provider-configuration-ec2.html | 16:38 |
| pitti | ok, I'll terminate the default bootstrap machine as well, and remove everything, then start again | 16:38 |
| SpamapS | ec2 is still the most reliable environment for juju, yes | 16:48 |
| pitti | melmoth: much much better indeed | 16:50 |
| pitti | melmoth: it created the machine, I can ssh in, and I see apt-get grinding | 16:50 |
| melmoth | good sign | 16:50 |
| pitti | units: | 16:50 |
| pitti | jhbuild/0: | 16:50 |
| pitti | agent-state: install-error | 16:50 |
| pitti | ok, that's now a charm bug | 16:50 |
| pitti | but the VM looks okay, juju is isntalled and all that | 16:51 |
| melmoth | yep, now you can see whats going on in /var/lib/juju/instances something | 16:51 |
| pitti | 2012-10-17 16:50:18,474: hook.output@ERROR: Unable to locate package libudisks2-dev | 16:52 |
| pitti | yep | 16:52 |
| pitti | melmoth: thanks, now I know where the logs are | 16:52 |
| pitti | melmoth: ok, this is territory I know again :) | 16:52 |
| melmoth | good :) | 16:52 |
| pitti | melmoth: so it seems I was running into an unfortunate bug when using the openstack type? | 16:52 |
| melmoth | most probably | 16:52 |
| pitti | melmoth: ok, thanks a lot for your time and help! | 16:53 |
| * pitti needs to run now | 16:54 | |
| === zz_KristianDK is now known as KristianDK | ||
| === matsubara is now known as matsubara-lunch | ||
| hallyn | SpamapS: is there any guidance regarding apparmor profile switching in juju? Or is that TBD? | 17:20 |
| hallyn | (juju-lxc) | 17:20 |
| === kapilt is now known as hazmat | ||
| subman | Is it possible to setup an installation via netboot when my dhcp server is my router? | 17:47 |
| SpamapS | hallyn: I'm not sure what you're asking. | 17:54 |
| SpamapS | hallyn: you'd be able to tell me better than I can tell you, how apparmor works | 17:54 |
| hallyn | SpamapS: I want to use juju-lxc to fire up containers that will run lxc. that requires a modified apparmor policy. Wondering how best to hook that in (if at all). | 17:59 |
| hallyn | SpamapS: I can just have juju spawn them, write a new policy, and hack the config file after the fact (and restart the container) to use the new policy. That's fine for me, of course. | 18:00 |
| hallyn | my question was just whether it's somethig you've talked about at all. | 18:00 |
| === shantorn_ is now known as shantorn | ||
| hallyn | Similarly (I don't need this right now, but) adding cgroup.devices.allow entries | 18:00 |
| hallyn | If you haven't talked about those at all, then perhaps it's best discussed at UDS | 18:01 |
| hallyn | (dunno if it'd be cleanly doable in environments.yaml) | 18:01 |
| === subman is now known as carling | ||
| === carling is now known as subman | ||
| === matsubara-lunch is now known as matsubara | ||
| pgnd | When ssh'ing to any of my Ubuntu Server 10LTS boxes around here, @ login I see an 'informative' message (http://pastebin.com/NK4reH41). I just built up a 12LTS box; when I ssh in, I get simply a shell prompt -- no message. I know *I* never config'd that 'informative message', so am guessing that there's a package missing? Any hints how to get that message back in place? | 18:33 |
| sarnold | pgnd: that's probably from installing landscape | 18:34 |
| sarnold | pgnd: if your new machine shouldn't be on landscape, you can still get some kind of useful message with pam_motd and the update-motd package | 18:35 |
| subman | Is it possible to setup an installation via netboot when my dhcp server is my router? | 18:36 |
| gabrtv | seeing Hash Sum mismatch on apt repositories | 18:37 |
| gabrtv | W: Failed to fetch bzip2:/var/lib/apt/lists/partial/security.ubuntu.com_ubuntu_dists_precise-security_main_source_Sources Hash Sum mismatch | 18:37 |
| gabrtv | http://pastebin.com/CrM2pBZa | 18:37 |
| pgnd | sarnold: minor confusion then --> http://pastebin.com/raw.php?i=kdXzhTN3 | 18:38 |
| sarnold | gabrtv: thanks, being investigated | 18:38 |
| pgnd | googling on pam_motd ... | 18:38 |
| sarnold | pgnd: oh :) perhaps apt-get purge update-motd? | 18:38 |
| pgnd | sarnold: purged, logged out, ssh'd in -- still no message | 18:39 |
| sarnold | pgnd: that may be where pam_motd comes in... | 18:40 |
| sarnold | of course, there's hushlogin and all that. there's a fair amount that goes in at login time. | 18:40 |
| SpamapS | pam_motd just shows /etc/motd | 18:41 |
| SpamapS | /etc/pam.d/sshd:session optional pam_motd.so motd=/run/motd.dynamic noupdate | 18:41 |
| SpamapS | pgnd: things that update that file are where the helpful message comes from. I think landscape-client is one of those things | 18:41 |
| pgnd | SpamapS: ah, so perhaps diff between 10LTS & 12LTS ... installing -client ... | 18:43 |
| pgnd | SpamapS: nope. still no message. | 18:44 |
| pgnd | SpamapS: fwiw, the 12LTS box has no "/run/motd.dynamic" ... | 18:45 |
| SpamapS | pgnd: right, Its updated in a somewhat lazy fashion | 18:48 |
| SpamapS | pgnd: but I admit to not knowing what program actually updates it | 18:48 |
| SpamapS | I believe it was at one time updated at login time, but that proved to annoy people because the logins would be slow sometimes | 18:49 |
| RoyK | some of you might like this - just wrote/patched it - a small thing to find duplicate files in a directory tree http://karlsbakk.net/finddup/ | 18:55 |
| pgnd | SpamapS: It appears that "/usr/bin/landscape-sysinfo" *gathers* the info in that message, but still dunno what mv's that info somewhere that the shell login process grabs/displays it :-/ | 18:58 |
| LordOfTime | SpamapS, thanks for your expediency on processing that sponsoring of the uplpoad for that nginx bug :) | 18:59 |
| LordOfTime | and isnt Q releasing tomorrow? | 18:59 |
| SpamapS | LordOfTime: yeah | 19:01 |
| pgnd | ah, lrwxrwxrwx 1 root root 46 Oct 17 11:56 /etc/update-motd.d/50-landscape-sysinfo -> /usr/share/landscape/landscape-sysinfo.wrapper*. it's *there* -- so what do I restart .... ? | 19:02 |
| LordOfTime | thought so, i assume the SRU team will be less busy once that happens? | 19:02 |
| LordOfTime | (at least slightly) | 19:02 |
| SpamapS | LordOfTime: actually I'm about to attack the precise-proposed queue right now :) | 19:03 |
| * Nafallo lends SpamapS's his axe | 19:03 | |
| SpamapS | LordOfTime: its a fair assumption that some of the release team will be more free next week yes. Tho some will be busy with UDS plans/prep | 19:03 |
| LordOfTime | true. | 19:03 |
| LordOfTime | post-UDS, it should be quieter :) | 19:03 |
| pgnd | bingo. 12's /etc/motd is a file. rm /etc/motd; ln -sf /var/run/motd /etc/motd fixes the problem. message is back. | 19:08 |
| === LordOfTime is now known as TheLordOfTime | ||
| blkperl | anyone else noticing slow speeds to ppa.launchpad.net? | 19:37 |
| TheLordOfTime | blkperl, #launchpad. BUT.... | 19:39 |
| TheLordOfTime | they're aware of currently id'd issues | 19:39 |
| TheLordOfTime | there's a current issue with the uploader afaict | 19:39 |
| TheLordOfTime | but also, given Q releases tomorrow at the earliest, i'd say that lag is expected as people push Q updates. | 19:39 |
| blkperl | alright thanks | 19:40 |
| === cpg|away is now known as cpg | ||
| === cpg is now known as cpg|away | ||
| FunnyLookinHat | Is ppa:chris-lea/node.js still the best means to obtain node.js ? | 19:56 |
| SpamapS | FunnyLookinHat: Its certainly very up to date. But the distro isn't so far behind anymore IIRC | 19:56 |
| gabrtv | FunnyLookinHat: that's what we're still using | 19:56 |
| FunnyLookinHat | Ah ok. | 19:56 |
| FunnyLookinHat | as of 11.10 was the last time I had to | 19:56 |
| FunnyLookinHat | As long as the PPA is stable that makes sense to me. | 19:57 |
| SpamapS | looksl ike Distro has 0.6.19, and chris has 0.8.12 | 19:57 |
| FunnyLookinHat | yeah - that's pretty far off | 19:57 |
| gabrtv | 0.6.x is ancient in node land | 19:58 |
| FunnyLookinHat | it's easy to be bleeding edge with little adoption ;0 | 19:58 |
| gabrtv | heh | 19:58 |
| FunnyLookinHat | jk jk | 19:58 |
| FunnyLookinHat | Well cool - that helps a lot - thanks SpamapS and gabrtv | 19:58 |
| SpamapS | Hrm, sadly, I don't see a 0.8 upload pending for the Debian packages :-P | 20:00 |
| SpamapS | not even in experimental | 20:00 |
| gabrtv | i get that the node runtime and packages are a fast moving target, but that's still surprising to me | 20:01 |
| FunnyLookinHat | Yeah that seems strange to me as well... | 20:01 |
| _cronus__ | Hello, does anyone know where to find the list of the metapackages created by the server team? ie dovecot-postfix | 20:01 |
| SpamapS | to me the problem with the PPA isn't that its bleeding edge.. its that once you decide you want off the crazy train.. you have to remove the PPA, and then.. you're up a creek for security | 20:03 |
| FunnyLookinHat | Right. | 20:03 |
| SpamapS | Its possible that the node maintainers are frustrated about the Debian policy change which required them to rename /usr/bin/node to /usr/bin/nodejs | 20:03 |
| ScottK | gabrtv: There was a huge delay in node.js in Debian due to the fact that it used /usr/bin/node, which was already used by another package (a fact that was known to the node.js upstream and they decided not to care). | 20:03 |
| gabrtv | hah, very interesting | 20:04 |
| ScottK | So any delay can probably be reasonably be put on the node.js upstream being deliberately anti-social. | 20:04 |
| ScottK | I'm sure they have a different view, of course. | 20:04 |
| SpamapS | ScottK: I forget, did they get the rights to have a node-legacy package that has a symlink at /usr/bin/node ? | 20:04 |
| ScottK | Yes, but no package in the archive is allowed to depend on it. | 20:05 |
| gabrtv | http://lists.debian.org/debian-devel-announce/2012/07/msg00002.html | 20:05 |
| SpamapS | ScottK: yeah, that seems pretty reasonable | 20:05 |
| ScottK | Yeah. That's the one. | 20:05 |
| SpamapS | Like, don't ask the archive to be broken, but if you, as a user, don't want to change all your shebangs and automated scripts.. here's some relief | 20:05 |
| gabrtv | looks like there's been no movement on implementing the technical committe's judgement: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=614907 | 20:09 |
| uvirtbot | Debian bug 614907 in tech-ctte "node: name conflicts with node.js interpreter" [Serious,Open] | 20:09 |
| SpamapS | its rather important to me, since the juju webui is nodejs based | 20:14 |
| keithzg_ | It looks to my untrained eye from that bug report that it must've been implemented, as the bug was archived on the 16th of August. | 20:16 |
| keithzg_ | Further, the packages recommended to exist by the committee seem to exist now, ex. http://packages.debian.org/sid/nodejs-legacy | 20:17 |
| SpamapS | gabrtv: thats not the bug with the recommendation btw http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=681360 | 20:19 |
| uvirtbot | Debian bug 681360 in nodejs "Provide /usr/bin/nodejs in nodejs with /usr/bin/node legacy package" [Serious,Fixed] | 20:19 |
| SpamapS | fixes in 0.6.19~dfsg1-3 | 20:19 |
| gabrtv | yep, thx.. still no 0.8.x upgrade i can find | 20:22 |
| SpamapS | gabrtv: yeah things have gone a bit dark.. they may be focused on RC bugs as well | 20:35 |
| harovali | I'm trying to decide among ext4, xfs, and reiser4; would some kind soul help me decide? | 20:43 |
| harovali | It's a 320GB disk | 20:44 |
| harovali | I'm trying to boost performance of my app, which has pretty heavy IO | 20:44 |
| SpamapS | harovali: reiser4 has gone unmaintained since its founder went to jail for 2nd degree murder... so I'd write that one off | 20:44 |
| harovali | SpamapS: and what about XFS ? Does it look like an attractive option looking for good IO Performance ? | 20:45 |
| rbasak | What kind of IO performance? | 20:45 |
| SpamapS | harovali: XFS is typically very good at everything except removing lots of files | 20:45 |
| harovali | SpamapS: interesting | 20:46 |
| sarnold | I may be a bit of a dinosaur, but I still prefer ext3. | 20:46 |
| SpamapS | I have heard that its especially good with multiple writers on large files.. which makes it particularly attractive for use in databases like MySQL and postgres | 20:46 |
| harovali | sarnold: I've been as a dinosaur as you | 20:46 |
| SpamapS | There are some workloads where ext4 makes things worse.. its true | 20:47 |
| patdk-wk | how much *performance* do you want from a single 320gb disk? | 20:47 |
| sarnold | (heck, even ext3 took some convincing, I got there years after everyone else had written off ext2...) | 20:47 |
| SpamapS | sarnold: well thats kind of lol.. ext3 was a huge jump in usability | 20:47 |
| patdk-wk | no matter what, your not going break a 80iop per sata disk | 20:47 |
| Nafallo | reiser has a tendency to murder data, as well... I've heard to many horror stories to trust it. | 20:47 |
| patdk-wk | nafallo, a few times to me :( | 20:47 |
| patdk-wk | but that was during reiser3 | 20:47 |
| harovali | Nafallo: interesting | 20:47 |
| SpamapS | I've never lost data to ext3, ext4, or XFS. Have lost data to all of the other options | 20:47 |
| patdk-wk | I have rarely lost data to ext3 (once, and wasn't really ext3 fault) | 20:48 |
| sarnold | Nafallo: indeed. hans ate 2/3 of my tree once. I was not impressed. | 20:48 |
| Nafallo | well, the kewl kidz use btrfs these days ;-) | 20:49 |
| Nafallo | and it's supposed to have speed improvements in 12.10 | 20:49 |
| Nafallo | just throwing it out here to see people's reactions :-) | 20:49 |
| patdk-wk | nafallo, ran that for a year | 20:50 |
| Nafallo | I lost data to ext4, but that was before it was "stable" | 20:50 |
| patdk-wk | I lost data cause of a bad wifi driver | 20:50 |
| Nafallo | lol | 20:50 |
| patdk-wk | as soon as I booted without the wifi enabled, no data corruption | 20:51 |
| patdk-wk | downloaded new kernel, all better :) | 20:51 |
| harovali | Nafallo: that too has been mentioned as pretty interesting | 20:51 |
| Nafallo | harovali: it's on your own risk currently. I'm not even sure if 12.10 will have fsck :-) | 20:51 |
| Nafallo | I /believe/ they've implemented offline btrfsck now though | 20:52 |
| Nafallo | so chances are it's in. | 20:52 |
| === cpg|away is now known as cpg | ||
| Daphko | Hello, i have some questions about server security | 21:43 |
| Daphko | can anybody help me? | 21:43 |
| RoyK | !ask | Daphko | 21:44 |
| ubottu | Daphko: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience | 21:44 |
| genii-around | Daphko: It's difficult to answer questions which have not been asked. | 21:47 |
| Daphko | ok i am not sure if anybody attack my server | 21:48 |
| Daphko | i checked my auth.log file | 21:48 |
| Daphko | and i want to ask anybody which knows much more then i about server security | 21:48 |
| Daphko | !paste | 21:48 |
| ubottu | 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. | 21:48 |
| addisonj | hrm... having a rough time getting dhcp working on brand new hp dl380 gen8, using the 4 NIC network card, dhclient3 is just refusing to do anything | 21:49 |
| Daphko | http://paste.ubuntu.com/1285933/ | 21:49 |
| Daphko | here are a lot of lines from my auth.log file | 21:49 |
| Daphko | i think a bot attacks me but not sure!? | 21:50 |
| addisonj | Daphko: looks like something is going on, is it a physical host or VPS? | 21:51 |
| sarnold | Daphko: yes, you are being attacked. it is probably not targetted at _you_, in the sense that botnets routinely scan for ssh servers to connect to | 21:51 |
| Daphko | linux vserver with ubuntu | 21:51 |
| sarnold | Daphko: they just guess usernames and passwords and get in often enough that it is worthwhile for them to try. | 21:51 |
| genii-around | Might want fail2ban to deter them a bit | 21:52 |
| Daphko | okay i change the ssh port and permiterootlogin | 21:52 |
| Daphko | genii-around: i have thought about to install fail2ban | 21:53 |
| sarnold | Daphko: simple steps: make sure your ssh does not accept passwords. (Require keys.) Configure /etc/hosts.allow to only allow login from network ranges you expect you'll use. Look into fail2ban or denyhosts to block these goofballs specifically. | 21:53 |
| sarnold | Daphko: and a super-cheap step you can take immediately is to null-route their traffic: http://en.wikipedia.org/wiki/Null_route | 21:54 |
| sarnold | Daphko: e.g., 'ip route add blackhole 208.109.120.151/32' | 21:54 |
| Daphko | ok thx | 21:57 |
| === Ursinha is now known as Ursinha-afk | ||
| ninjai | Can somebody explain why I'm unable to upgrade my server from 10.04 to 12.04? It says "No new release found" when I enter do-release-upgrade. What gives? | 22:16 |
| === Shehrazad is now known as ElixirVitae | ||
| genii-around | ninjai: What is the result of: grep = /etc/update-manager/release-upgrades ? | 22:17 |
| ninjai | genii-around, Prompt=lts | 22:18 |
| genii-around | ninjai: Have you tried with -p ? | 22:19 |
| ninjai | genii-around, same result | 22:20 |
| * genii-around reads the long and tedious bug 990740 | 22:26 | |
| uvirtbot | Launchpad bug 990740 in python-defaults "upgrading from lucid to precise fails" [High,Invalid] https://launchpad.net/bugs/990740 | 22:26 |
| ninjai | oh joy. Thanks! | 22:27 |
| ninjai | actually wait | 22:27 |
| ninjai | my upgrade never "failed" | 22:27 |
| ninjai | it just doesn't see an upgrade for me | 22:27 |
| genii-around | ninjai: Have you done dist-upgrade before running do-release-upgrade ? | 22:29 |
| ninjai | no | 22:29 |
| ninjai | doesn't work either | 22:30 |
| genii-around | ninjai: By "doesn't work" you mean dist-upgrade fails, or that it completes successfully but that subsequently running the do-release-upgrade fails as before? | 22:31 |
| ninjai | it just says there's nothing to upgrade | 22:31 |
| ninjai | 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. | 22:31 |
| ninjai | I think I found the problem | 22:32 |
| ninjai | URI = http://127.0.0.1/meta-release | 22:32 |
| ninjai | URI_LTS = http://127.0.0.1/meta-release | 22:32 |
| ninjai | what should those be? | 22:32 |
| ninjai | should they be the same as shown here? http://askubuntu.com/questions/124999/distribution-upgrade-problem-no-new-release-found | 22:33 |
| ninjai | genii-around, that fixed my problem. | 22:34 |
| protoCall7 | Hi all, I am trying to get an Ubuntu 12.04.1 domU to boot in Xen 4.0.1, and getting a bunch of errors regarding upstart services http://pastebin.com/hafwVm2s | 22:53 |
| protoCall7 | could someone possibly point me in the right direction for resolving these? | 22:53 |
| protoCall7 | the services listed are symlinks from /etc/rcX.d/ to /etc/init.d/<servicename>, which in turn are symlinks to upstart | 22:53 |
| ChmEarl | protoCall7, these are not errors, but a little warning that there is a new, cleaner way to do it | 22:55 |
| protoCall7 | ChmEarl: Interesting… Thanks. The system hangs on the sssd start, but that might be unrelated. This poor VM is having a LOT of problems :-/ | 22:56 |
| protoCall7 | every other boot, the disk is saying that it's unclean and mounting read-only, even though fsck is giving it a clean bill of health before I boot, network comes up, but has no connectivity, etc etc lol. I'm just taking them one at a time | 22:56 |
| ChmEarl | protoCall7, echo '127.0.0.2 myboxname' >> /etc/hosts | 22:57 |
| protoCall7 | ChmEarl: I have seen those warnings before when starting services from /etc/init.d by hand, but i've never seen the initctl: line… is that also just a part of the warning? | 22:57 |
| protoCall7 | out of curiosity, what is adding 127.0.0.2 to hosts doing? I've seen 127.0.0.1 as lo before, but not .2 | 22:58 |
| ChmEarl | the initctl line: 20sshd -> sshd (should fix it) | 22:59 |
| sarnold | all of 127.0.0.0/8 routes to localhost | 22:59 |
| protoCall7 | ChmEarl: You lost me there, are you saying symlink S20sssd to /etc/init.d/sssd? | 23:01 |
| protoCall7 | or rename the file all together | 23:02 |
| ChmEarl | protoCall7, about unclean mount... mount / with option: barrier=0 | 23:05 |
| protoCall7 | k, I'll give that a shot *crosses fingers* | 23:06 |
| ChmEarl | edit /etc/fstab | 23:06 |
| protoCall7 | yup, just threw it in there, rebooting now | 23:06 |
| protoCall7 | glad this isn't a production machine ;) | 23:07 |
| protoCall7 | much better: /dev/xvda1: clean, 87751/589824 files, 824776/2357248 blocks | 23:07 |
| sarnold | ChmEarl: why does that work? | 23:07 |
| ChmEarl | sarnold, idk | 23:08 |
| sarnold | hehe | 23:08 |
| * ChmEarl hangs head.. so much about Linux I don't understand | 23:08 | |
| protoCall7 | sarnold: I found this blog post with the same… http://blog.jolexa.net/2012/02/linode-dont-use-barriers-and-ext-4/ | 23:09 |
| protoCall7 | sarnold: looks like a possible bug, perhaps? | 23:09 |
| sarnold | protoCall7: not much else would make sense to me, but ... yeah. hrm. | 23:10 |
| protoCall7 | sarnold: in the comments section, there was a request for the author of the post to file a bug-report, but he declined to do so... | 23:10 |
| ChmEarl | protoCall7, after a few edits like those in setup, I used my 10GB precise VM to build xen 4.2 from tar.gz... took less than 1hour to install the Dev ENV and compile | 23:15 |
| protoCall7 | not bad at all… This has been a nightmare today trying to get this machine to come up | 23:16 |
| protoCall7 | I've been tempted to start over several times lol | 23:16 |
| ChmEarl | ubuntu-server is sweet when all you want to do is do a few build | 23:18 |
| protoCall7 | *nod* this box is dangerously close to being replaced with a Cent machine (blasphemy, i know) | 23:19 |
| protoCall7 | then again, upstart is getting a lot of adoption these days, so maybe it's just time to go RTFM and learn it... | 23:21 |
| protoCall7 | sry for all of the questions, but is it normal during boot for the system to be starting and then immediately stopping so many things? http://pastebin.com/6e8jbHqW | 23:25 |
| === KristianDK is now known as zz_KristianDK | ||
| sarnold | protoCall7: my /var/log/boot.log only shows three services stopping: System V initialisation compatibility, anac(h)ronistic cron, save kernel messages | 23:27 |
| sarnold | (but that's a desktop image..) | 23:27 |
| protoCall7 | Hmm… interesting, that MIGHT be normal then. I now have the unclean drive and the upstart errors corrected, but I'm still hanging on boot, and can't get network up. I can get into linux single fine though, so… progress all around | 23:28 |
| ChmEarl | protoCall7, your VM is running in 2TB? and has about 1200 packages? | 23:32 |
| ChmEarl | I debootstrap and then run this script: http://paste.ubuntu.com/800901/ | 23:32 |
| ChmEarl | 10GB and about 160 packages | 23:32 |
| protoCall7 | no, this is a 10gb image | 23:33 |
| protoCall7 | yes, that sounds about right | 23:33 |
| protoCall7 | most of its storage is on an NFS mount | 23:33 |
| protoCall7 | let me get back to single, I can get more done when I'm not sitting here waiting on a hung boot | 23:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!