| zul | jamespage/coreycb: i switched over glance to use the milestone-proposed git branch | 00:17 |
|---|---|---|
| Havenstance_ | question | 02:17 |
| sarnold | answer | 02:18 |
| Havenstance_ | Installed ubuntu 13.10 server, during install i had DHCP and the internet worked, now after install I have no DHCP | 02:18 |
| Havenstance_ | any idea's what might be wrong? I checked /etc/networking/interfaces and eth0 is listed there as dhcp | 02:18 |
| Havenstance_ | but when I run a sudo service networking restart I get networking stop/waiting | 02:19 |
| Havenstance_ | never goes beyond it, router is giving out DHCP Addresses fine as I'm talking to you on it now | 02:19 |
| sarnold | Havenstance_: yeah don't run "service networking restart", that busts dbus badly ... | 02:19 |
| Havenstance_ | ok, how would I get it to get a DHCP, that's what I don't understand it had one during install | 02:20 |
| sarnold | Havenstance_: ifdown eth0 ; ifup eth0 ; is the better way to cycle networking | 02:20 |
| Havenstance_ | but once install finished and it rebooted it gets nothing | 02:20 |
| Havenstance_ | ok | 02:20 |
| Havenstance_ | its sending DHCP Discover on 255.255.255.255 | 02:21 |
| Havenstance_ | my netmask should be 255.255.255.0 | 02:21 |
| sarnold | Havenstance_: check logs, there might be something there, /var/log/syslog, /var/log/upstart/network-interface* | 02:21 |
| sarnold | Havenstance_: dhcp discover is sent using a link-local broadcast packet; 255.255.255.255 is correct | 02:21 |
| Havenstance_ | okay :) | 02:21 |
| sarnold | nice paranoia :) | 02:22 |
| Havenstance_ | well, i understand enough about networking but in linux it seems to be enough to get me in trouble lol | 02:22 |
| Havenstance_ | Can't a guy just shitcan windows, that's all I want to do :) | 02:22 |
| Havenstance_ | give bill gates some sign language | 02:22 |
| sarnold | heh, it took me a few years to transition completely away from windows | 02:23 |
| Havenstance_ | i've been running ubuntu on and off for ever now | 02:23 |
| Havenstance_ | I still have disks from 5.04 and the such, when they used to give away the CDs to those of us unlucky people who had Dial Up :) | 02:23 |
| Havenstance_ | 4.10 too apparently :) | 02:24 |
| Havenstance_ | here's something in syslog idk if it means anything or not but it says -- apr 1 22:03:43 uss-enterprise rsyslogd-2039 could not open output pipe '/dev,console' | 02:26 |
| sarnold | /dev,console ?? what an odd mistake.. | 02:27 |
| sarnold | Havenstance_: time for me to bail, good luck :) have fun | 02:28 |
| Havenstance_ | dhcp request of 192.168.1.104 on eth0 to 255.255.255.255 port 67 | 02:29 |
| Havenstance_ | dhcp offer of 192.168.1.104 from 192.168.1.1 | 02:29 |
| Havenstance_ | dhcpack of 192.168.1.104 from 192.168.1.1 | 02:29 |
| Havenstance_ | bound to 192.168.1.104 -- renewal in 33966 seconds | 02:30 |
| Havenstance_ | but ifconfig still shows no dhcp addr | 02:30 |
| jrwren | ubuntu systemd in the future will already benefit : https://plus.google.com/+TomGundersen/posts/eztZWbwmxM8 | 02:31 |
| === Havenstance_ is now known as Havenstance | ||
| koolhead17 | hello world11 | 04:17 |
| ubunty_ | Why do I get this message http://paste.ubuntu.com/7192789/ What could be the issue? http://paste.ubuntu.com/7192787/ | 04:22 |
| ubunty_ | why do I keep getting "sudo: unable to resolve host (none)" during my PXE boot installation? | 05:49 |
| ubunty_ | It will complete the installation but with no internet access which I suspect is caused by this issue | 05:49 |
| TJ- | ubunty_: Sounds like there is no DNS resolver configured | 05:50 |
| ubunty_ | How would I got about fixing this issue? Or I mean how would I get to the point where I am configuring the dns resolver? | 05:51 |
| sarnold | ubunty_: does your /etc/hosts look sane? does your /etc/nsswitch.conf look sane? | 05:52 |
| TJ- | add an entry in /etc/resolv.conf | 05:53 |
| ubunty_ | ok ill try that | 05:53 |
| TJ- | ubunty_: If you did a PXE boot, the interface will be inherited from the kernel, and so you need to take care of such things | 05:53 |
| ubunty_ | and if it is inherited from the kernel how can I make those changes if needed? | 05:54 |
| sarnold | ubunty_: yikes are you still doing this?? http://paste.ubuntu.com/7192787/ | 05:54 |
| ubunty_ | nope got that worked out | 05:54 |
| sarnold | ubunty_: after CIDR was introduced back in the 90s, I think support for non-CIDR style netmasks has long since atrophied, I wouldn't expect "subnet-mask 255.255.0.255; | 05:55 |
| sarnold | to work all that often | 05:55 |
| sarnold | oh good | 05:55 |
| ubunty_ | resolv.conf looks like this http://paste.ubuntu.com/7192972/ | 05:56 |
| TJ- | ubunty_: best to fix the typo | 05:58 |
| sarnold | and there's no point in searching hsd1.tx.comcast.net, you'll never look up hosts under that domain.. | 05:58 |
| ubunty_ | oops but it was a copy paste typo not on my actual resolv.conf file | 05:59 |
| ubunty_ | google.com for the domain? | 05:59 |
| sarnold | none at all should work fine | 06:00 |
| ubunty_ | so take out domain and search? | 06:00 |
| TJ- | If you're doing PXE boot, that infers a local BOOTP/DHCP server, is there not also a local network DNS server? | 06:02 |
| ubunty_ | yes my modem is giving out those DNS | 06:03 |
| ubunty_ | thats what it says on my dhcpclient.leases | 06:04 |
| === ohcrap is now known as cupojoe | ||
| === psivaa is now known as psivaa-afk | ||
| jamespage | zul, Daviey: that nova test failure only happens in virtualized PPA builds | 08:02 |
| jamespage | as found in openstack-ubuntu-testing and icehouse-staging for the CA | 08:02 |
| * jamespage siggs | 08:02 | |
| allaga_ | hey :) | 08:20 |
| === allaga_ is now known as allaga | ||
| === psivaa-afk is now known as psivaa | ||
| lordievader | Good morning. | 08:57 |
| mischief | hello | 08:59 |
| mischief | i'm trying to boot ubuntu 14.04 LTS server from a hard drive | 08:59 |
| mischief | i dd the iso to the disk and reboot, and now the installer can't find the cd image. | 08:59 |
| mischief | the server has a LSI Logic Fusion-MPT SAS card to which i believe the disk is attached to, but i can't find any kernel module for this in the installer | 09:00 |
| cfhowlett | mischief until release, 14.04 support = #ubuntu+1 | 09:00 |
| mischief | oh, it's not released until the 17th :| | 09:01 |
| mischief | well, i have a feeling 12.04 would result in the same issue, but i can try it too | 09:01 |
| mischief | cfhowlett: is there a standard way to get raid drivers at install time? | 09:09 |
| cfhowlett | mischief oh, my. Never done that and so I don't know. sorry. | 09:10 |
| mischief | poop | 09:10 |
| mischief | i really don't want to have to file a ticket at my dedicated server provider :< | 09:10 |
| cfhowlett | mischief stay in channel and ask. someone else will know. | 09:11 |
| mischief | i am getting 13.10 server image now, to try instead | 09:11 |
| mischief | but i think it will end up the same at 14.04 | 09:11 |
| mischief | same problem on 13.10 ;_; | 09:19 |
| zetheroo | Is there a GUI for OpenIPMI? | 09:34 |
| mischief | zetheroo: i wasn't able to find one | 10:10 |
| ivoks | gui for ipmi? | 10:20 |
| ivoks | why would anyone want that? :) | 10:20 |
| mischief | so they dont have to use the stupid java clients ;) | 10:21 |
| ivoks | isn't java client a gui? | 10:22 |
| mischief | yes | 10:22 |
| mischief | i'd rather a native client than java | 10:22 |
| ivoks | why do you need GUI access to the server? | 10:22 |
| ivoks | i mean, what's wrong with SOL? | 10:23 |
| mischief | can you mount iso with that | 10:23 |
| ivoks | you can't do ISO with SOL, true | 10:23 |
| mischief | sorry if i sound dumb, it's my first day | 10:23 |
| ivoks | but PXE is much faster than ISO anyway | 10:23 |
| mischief | i dont have pxe on my host | 10:23 |
| mischief | i mean, pxe exists, but my host doesn't have anything to boot to | 10:24 |
| ivoks | let's go from start | 10:25 |
| ivoks | you have a server isolated on the network to wich you want to install ubuntu 12.04? | 10:25 |
| mischief | uh | 10:26 |
| mischief | yes? | 10:26 |
| mischief | i'm not sure what you mean by isolated on the network. | 10:26 |
| ivoks | PXE is network boot | 10:27 |
| mischief | right | 10:27 |
| ivoks | usually, datacenters have some kind of network installation setyp | 10:27 |
| ivoks | setup | 10:27 |
| ivoks | so that people don't walk around with CDs | 10:27 |
| ivoks | so when i say isolated, i mean network without such infrastructure | 10:28 |
| mischief | well i tried to pxe boot, but the tftp server makes no offers | 10:28 |
| mischief | dhcp works | 10:28 |
| ivoks | ok, then you probably don't have tftp/pxe setup on your network | 10:29 |
| mischief | yea | 10:29 |
| mischief | so i tried to mount an iso with the IPMI shit on my dedi | 10:29 |
| mischief | some 'MEGARAC Aster' by AMI on a dell poweredge | 10:29 |
| mischief | the kvm console it provides is a really poor java program, and i have openjdk on debian | 10:30 |
| mischief | most of it works, *except* mounting an iso | 10:30 |
| mischief | so i tried to write the ubuntu server iso to the second disk, and directly boot to that | 10:30 |
| mischief | it boots, it scans for the cd image.. and it can't find it. | 10:31 |
| ivoks | dell uses DRAC, so you should be able to open web browser and go to DRAC's IP | 10:31 |
| mischief | and why? because the hard drive is on a fucking lsi fusion-mpt sas card and the ubuntu iso has no mptsas driver | 10:31 |
| ivoks | well, watch your language | 10:31 |
| ivoks | ubuntu iso does have mptsas driver | 10:32 |
| ivoks | but it's not part of the kickoff installer | 10:32 |
| ivoks | it gets added during discovery process | 10:32 |
| mischief | sorry, i'm just very frusterated. i've been trying to get openstack running on this dedicated server since last week | 10:33 |
| mischief | the web ui of the management console looks like -> http://blog.milford.io/wp-content/uploads/2012/03/megarac2.png | 10:33 |
| ivoks | yeah... it's ugly, but you have to do it that way | 10:34 |
| ivoks | at least to kick off installer | 10:34 |
| mischief | what way? | 10:34 |
| ivoks | mount the iso within the web ui and boot from cd | 10:35 |
| mischief | there is no way to do it from the web ui! | 10:35 |
| mischief | and it's full of bugs anyway | 10:36 |
| mischief | the web ui runs on https, but frequently tries to redirect me to a url like https://1.2.3.4:80/... | 10:36 |
| mischief | which doesn't work obviously | 10:36 |
| ivoks | and you say this is dell poweredge? | 10:37 |
| mischief | one moment and i can run dmidecode.. | 10:37 |
| ivoks | (that's not what dell's drac ui looks like) | 10:38 |
| ivoks | http://wiki.hetzner.de/index.php/Datei:Idrac_logon_en.jpg | 10:39 |
| ivoks | that's how it looks like ^ | 10:39 |
| mischief | it's definitely a dell system, albeit probably an older one | 10:39 |
| mischief | now i can't even run the java applet, great | 10:40 |
| ivoks | there are also drac cli tools | 10:45 |
| ivoks | but i'm not sure they can handle isos | 10:46 |
| ivoks | Attaching, Auto-Attaching, and Detaching Virtual Media using RACADM | 10:46 |
| ivoks | racadm config -g cfgRacVirtual -o cfgVirMediaAttached 1 | 10:46 |
| zetheroo | so is the short answer that there is none? | 10:46 |
| ivoks | http://support.dell.com/support/systemsinfo/document.aspx?~file=/software/smdrac3/drac5/145/en/ug/racugc1b.htm | 10:47 |
| mischief | it's a dell poweredge c6100 | 10:47 |
| ivoks | ah, this doesn't have drac | 10:48 |
| ivoks | just ipmi | 10:48 |
| mischief | and is apparently incapable of mounting ISO's | 10:48 |
| mischief | even though a virtual cd rom drive appears | 10:49 |
| mischief | so there's really no way for me to reinstall ubuntu, unless i either ask the datacenter to do it, or build my own iso with the mptsas driver in the kernel | 10:49 |
| ivoks | http://www.iptp.net/en/support_ipmi.php | 10:49 |
| ivoks | looks like yours | 10:50 |
| zetheroo | I have been using IPMIview from Supermicro for our Supermicro servers - but this time I am dealing with a HP server which is not being seen by IPMI view ... | 10:50 |
| mischief | it's not the same | 10:50 |
| mischief | this is AMI ipmi, not SuperMicro | 10:50 |
| ivoks | zetheroo: hp uses ilo; not every ilo is ipmi compatible | 10:50 |
| ivoks | mischief: look at the screenshots | 10:51 |
| ivoks | mischief: it's the same, just rebranded | 10:51 |
| zetheroo | I have also checked out the IPMI (iLO) of the HP via the browser but there is hardly anything in there - and no KVM viewer | 10:51 |
| mischief | well, perhaps so. but my interface does not have such options as 'Virtual Media' | 10:51 |
| ivoks | 1) Launch the remote console; | 10:51 |
| ivoks | 2) On the top of window select Media->Virtual Media; | 10:51 |
| mischief | it doesn't exist in the web ui man | 10:52 |
| ivoks | zetheroo: some older ilos are IE only | 10:52 |
| zetheroo | ok, so is there any utility (graphical) that can be used to connect to iLO ? | 10:53 |
| mischief | http://i.imgur.com/VRrbnU9.png | 10:54 |
| mischief | that's the java client with the iso error, next to the web page of the ipmi | 10:54 |
| mischief | i think i should just ask the dc for a new server :) | 10:54 |
| ivoks | mischief: well, that's you answer... can you redirect usb? | 10:54 |
| zetheroo | I could even try a Windows-based on in wine I suppose ... | 10:54 |
| mischief | ivoks: nope | 10:55 |
| ivoks | mischief: ubuntun iso is cd/hdd hybrid; you can boot it as a disk or as a cd | 10:55 |
| mischief | i can't redirect anything from the media menu | 10:55 |
| mischief | cdrom, iso, usb, floppy | 10:55 |
| ivoks | bad luck | 10:55 |
| mischief | and.. i thought i made this clear. i did write the ubuntu server iso to the hard drive and try to boot that | 10:56 |
| mischief | but the installer can't load itself from the cd image, because it can't read the disk, because there is no mptsas module in the installer's kernel | 10:56 |
| mischief | bit of a chicken and egg problem | 10:56 |
| ivoks | mischief: correct, it actually tries to read cdrom to fetch that module | 10:56 |
| ivoks | if you can dd iso to a usb stick, that would solve the problem | 10:57 |
| mischief | what usb stick :-) | 10:57 |
| mischief | this server is 200 miles away | 10:57 |
| ivoks | well, then you are managing a remote server that has pretty poor remote management capabilities | 10:58 |
| ivoks | it can't boot from a virtual media | 10:58 |
| ivoks | and you don't have pxe | 10:58 |
| mischief | i've come to realize that | 10:58 |
| ivoks | how about... | 10:59 |
| ivoks | net boot iso | 10:59 |
| ivoks | hm... | 10:59 |
| ivoks | at what stage does installer fail? | 10:59 |
| mischief | it's a possibility, if it can do dhcp and then load the mptsas module from the 'net | 10:59 |
| mischief | ivoks: scanning for and mounting cdrom | 11:00 |
| mischief | right after picking the language/keyboard pretty much | 11:00 |
| ivoks | you can try net boot iso | 11:00 |
| ivoks | i'm not 100% sure it will work | 11:00 |
| mischief | might as well | 11:00 |
| ivoks | http://www.howtoforge.com/boot-linux-over-http-with-netboot.me | 11:02 |
| ivoks | that might be an option too | 11:02 |
| ivoks | doh | 11:03 |
| ivoks | latest ubuntu is 10.04 | 11:03 |
| mischief | ivoks: nice work man | 11:08 |
| mischief | netboot is going | 11:08 |
| mischief | now i just ask providence to help me install openstack without too many headaches | 11:09 |
| mischief | ivoks: thanks again, saved me a lot of time. i would have either had to wait on the datacenter guys or build my own installer image with the driver, if it had not been for the netboot image | 11:30 |
| ivoks | mischief: np | 11:37 |
| === Havenstance is now known as Haven|ToWork | ||
| zul | jamespage: glance rc1 still waiting to be accepted | 12:32 |
| jamespage | zul, did you see my ping re nova in virt-ppa's | 12:34 |
| zul | jamespage: i did...not good | 12:36 |
| jamespage | zul, still need to decide what todo about db encoding | 12:36 |
| jamespage | zul, the tests must have passed in the lab | 12:36 |
| zul | jamespage: they did | 12:36 |
| jamespage | otherwise it does not get to the PPA | 12:36 |
| mischief | i'm installing openstack on trusty ^.^ | 12:43 |
| jamespage | mischief, good! | 12:46 |
| jamespage | mischief, things to watch for - db table encoding | 12:46 |
| jamespage | mischief, if you are deploying neutron you have to configure neutron to talk to nova for nic plugging notifications | 12:48 |
| mischief | jamespage: any link to tips on that? | 12:55 |
| mischief | jamespage: re table encoding - a bug got filed that has the fix | 12:56 |
| mischief | https://bugs.launchpad.net/ubuntu/+source/keystone/+bug/1300814 | 12:56 |
| uvirtbot | Launchpad bug 1300814 in keystone "Tables "migrate_version" have non utf8 collation" [Undecided,New] | 12:56 |
| jamespage | mischief, kinda | 12:56 |
| mischief | i'm setting up nova as we speak | 12:56 |
| mischief | so neutron is next :^) | 12:56 |
| jamespage | mischief, https://github.com/openstack/neutron/blob/master/etc/neutron.conf#L295 | 12:56 |
| jamespage | zul: do we have an upstream bug for the keystone utf-8 break? | 13:07 |
| zul | jamespage: gimme a sec | 13:08 |
| zul | jamespage: affects heat as well https://bugs.launchpad.net/oslo/+bug/1301036 | 13:09 |
| uvirtbot | Launchpad bug 1301036 in oslo "openstack.common.db.sqlalchemy.migration utf8 table check issue on initial migration" [Critical,In progress] | 13:09 |
| zul | jamespage: im going to test that patch in the openstack-ci lab and re deploy using the charms | 13:10 |
| jamespage | zul, OK | 13:10 |
| zul | jamespage: ok we are just waiting for swift rc1 now | 13:24 |
| jamespage | zul, ack | 13:27 |
| jamespage | zul, I think we might have a bug in the image snapshotting process | 13:35 |
| jamespage | 2014-04-02 13:33:59.067 32697 INFO nova.virt.libvirt.driver [req-c05e37f9-1d3f-49ad-a6a9-57b6d690979e fdb9cd48f1804034a98aebd2918a9bdc d12b5ca2e2ca4329ac3b232052bd6a5e] [instance: cbe3d0c6-ee35-450e-9cbe-614108f13d1a] Snapshot extracted, beginning image upload | 13:35 |
| jamespage | 2014-04-02 13:33:59.297 32697 ERROR oslo.messaging.rpc.dispatcher [-] Exception during message handling: Requested operation is not valid: No active operation on device: drive-virtio-disk0 | 13:35 |
| zul | uh? is there a traceback? | 13:35 |
| jamespage | zul, yes | 13:36 |
| jamespage | zul, http://paste.ubuntu.com/7194270/ | 13:37 |
| jamespage | zul, wanna bug? | 13:37 |
| zul | jamespage: yes please | 13:38 |
| zul | jamespage: interesting | 13:38 |
| zul | jamespage: anything in the /var/log/libvirt ? | 13:39 |
| zul | jamespage: trusty or CA? | 13:46 |
| jamespage | zul, trust | 13:46 |
| jamespage | y | 13:46 |
| jamespage | zul, https://bugs.launchpad.net/ubuntu/+source/nova/+bug/1301393 | 13:46 |
| uvirtbot | Launchpad bug 1301393 in nova "Live image snapshotting failure on icehouse" [Undecided,New] | 13:46 |
| Haven|Work | TJ- could I pick your brain for a moment? | 13:59 |
| jamespage | zul, nothing in the libvirt logs | 14:21 |
| zul | shazbut | 14:21 |
| zul | im just trying to reproduce it now | 14:22 |
| jamespage | zul: adam_g was hankering after new libvirt - wonder whether he's seen anything similar | 14:31 |
| zul | jamespage: like libvirt 1.2.3? or just an ubuntu cloud archive update | 14:32 |
| jamespage | just ca | 14:32 |
| zul | ah last time we touched it was for an libxl update...hallyn is doing a qemu 2.0/libvirt update for it though | 14:32 |
| hallyn | but not just yet | 14:35 |
| tomixxx5 | what does mean entry " | 14:54 |
| tomixxx5 | sorry | 14:54 |
| tomixxx5 | what does mean entry "*" in gateway column when i call "route" | 14:55 |
| zul | jamespage: i havent been able to reproduce it | 15:08 |
| jamespage | zul, hhm | 15:09 |
| zul | granted i tried devstack but the tests ran fine for me | 15:10 |
| ivoks | damn | 15:10 |
| ivoks | i have to kill highlight on 'ante' :) | 15:10 |
| zul | ivoks: try highlighting chuck | 15:11 |
| zul | as in "im going to chuck this thing in here" | 15:11 |
| ivoks | wanted, granted... | 15:11 |
| jamespage | zul: bug 1301154 | 15:29 |
| uvirtbot | Launchpad bug 1301154 in python-openstackclient "python-novaclient needs to bump the epoch like the Debian package" [High,Confirmed] https://launchpad.net/bugs/1301154 | 15:29 |
| jamespage | opinion? | 15:29 |
| jamespage | might as well make it compat | 15:29 |
| zul | jamespage: why not | 15:30 |
| === Overand_ is now known as Overand | ||
| seaninryan | quit | 16:02 |
| jamespage | zul: added some libvirtd.log to bug 1301393 | 16:39 |
| uvirtbot | Launchpad bug 1301393 in nova "Live image snapshotting failure on icehouse" [Undecided,New] https://launchpad.net/bugs/1301393 | 16:39 |
| jamespage | zul, ah | 16:41 |
| zul | hallyn: ^^^ | 16:41 |
| jamespage | I might see the issue | 16:41 |
| jamespage | zul, root disk filling up | 16:42 |
| zul | ah...:) | 16:42 |
| hallyn | that's bug 1301393 you're talking about, jamespage ? | 16:44 |
| uvirtbot | Launchpad bug 1301393 in nova "Live image snapshotting failure on icehouse" [Undecided,New] https://launchpad.net/bugs/1301393 | 16:44 |
| jamespage | hallyn, yes | 16:44 |
| hallyn | k | 16:44 |
| jamespage | hallyn, zul: I've bumped the root-disk size for m1.medium in serverstack | 16:46 |
| zul | jamespage: cool | 16:46 |
| zul | jamespage: i was using /mnt for some devtsack testing | 16:46 |
| hallyn | zul: i'm trying to migrate a vm from a saucy laptop to a trusty one. have you doen that, and had success/ | 16:54 |
| hallyn | zul: seems to hang on sh -c if nc -q 2>&1 | grep "requires an argument" >/dev/null 2>&1 | 16:55 |
| patdk-wk | I have done several precise -> trusty, without issue | 16:55 |
| hallyn | patdk-wk: live migrations? | 16:55 |
| patdk-wk | oh, migrate | 16:55 |
| hallyn | those are supposed to be impossible :) | 16:55 |
| patdk-wk | I read it as, migrate version :) | 16:56 |
| zul | hallyn: i havent | 16:56 |
| jamespage | zul, hallyn: looks OK now - marked invalid | 16:59 |
| hallyn | zul: ok. hm. | 16:59 |
| === irv_ is now known as irv | ||
| === henkjan_ is now known as henkjan | ||
| === robher_ is now known as robher | ||
| adam_g | jamespage, i haven't hit that but i needed the newer libvirt for something else | 19:02 |
| adam_g | jamespage, you might be interested in https://review.openstack.org/#/c/74889/ someone is playing with the icehouse pocket and beating at it in the gate | 19:02 |
| tsilenzio | hello | 20:08 |
| tsilenzio | I have an ubuntu-server (13.10) box running nginx, php-fpm, mysql and postgres, etc. | 20:09 |
| tsilenzio | its 8 cores, 20gb, etc. its a local development box that 4 developers use | 20:09 |
| tsilenzio | right now when I do free -m as a command I get (Mem: 19070 total 18806 used 264 free 0 shared 250 buffers 16524 cached) | 20:11 |
| tsilenzio | is it normal to be so low on free memory? or do i have a memory leak? :s | 20:11 |
| Pici | you have 16GB cached. thats fine. | 20:12 |
| tsilenzio | ah alright, thought i had a memory leak | 20:12 |
| Pici | http://www.linuxatemyram.com/ | 20:12 |
| tsilenzio | why do i have some 6 bytes or 6kb of swap used? :s jw | 20:13 |
| patdk-wk | looks like you have a harddrive leak | 20:24 |
| patdk-wk | leaking all over your cached ram | 20:24 |
| faiss | hi, how to rename p3p2 into eth0 on saucy? | 20:28 |
| sarnold | faiss: check out /etc/udev/rules.d/70-persistent-net.rules | 20:49 |
| bladernr_ | hey, how familar are any of you with tweaking kernel routing? | 20:52 |
| jamespage | adam_g, don't worry - it was just me being stupid (to small a root volume) | 20:52 |
| bladernr_ | there used to be a setting in /proc/sys/net/ maybe in ./ipv4 to force packets to ONLY go out and return via the interface they were supposed to. | 20:52 |
| jamespage | adam_g, I'm fairly comfortable with the icehouse pocket | 20:53 |
| jamespage | its smoking OK for me right now | 20:53 |
| bladernr_ | by default, it's possible to have packets go out eth1 and come in eth0 if both are on the same network | 20:53 |
| bladernr_ | and it's been well over 5 years since I last tried this and I forgot the magic switch. | 20:53 |
| sarnold | bladernr_: check Documentation/networking/ip-sysctl.txt for the rp_filter variable -- that looks like the right thing on first glance | 20:56 |
| bladernr_ | thanks... that's a good start then... | 20:56 |
| === Havenstance2 is now known as haven|Home | ||
| Quintasan | Hi, is there any way to make sure which way the data is mirrored in RAID1? Currently I have /dev/md1 (RAID1) with only one disk which contains my data, today I have created /dev/md0 which is RAID0 made out of two 1TB disks, can I just add /dev/md0 to the /dev/md1 array as a device and the data won't get overwritten? | 21:16 |
| wiredfool | openssl question: Starting a server in a vm at startup, it starts up O(1) seconds after boot and reads 32 bytes from /dev/urandom to seed openssl's random number generator. . In trusty, O(100) seconds after boot, [ 99.783379] random: nonblocking pool is initialized shows up in the logs. | 21:22 |
| wiredfool | Is this as sketchy as it sounds? | 21:22 |
| sarnold | wiredfool: every process that uses openssl's random number generator should seed it themselves when they need it; that specific message comes from the linux kernel when it has finally collected enough entropy for the pool to be safe to use | 21:30 |
| wiredfool | sarnold: so, yes, it's sketchy? | 21:31 |
| wiredfool | for instance, starting a ssl-enabled webserver @ system startup | 21:31 |
| Diegonat | hi.. With iftop I see a connection to a IP. I'd like to understand what program is connecting to that ip. How can I do it? | 21:31 |
| sarnold | wiredfool: it would be best if it could block, but that's not the way the linux /dev/urandom works :( | 21:31 |
| sarnold | wiredfool: however, we've got a new feature that you could turn on, if it isn't on already... let me go find a nice link to pollinate :) | 21:32 |
| wiredfool | is there an upstart event that could trap the kernel saying urandom is ready? | 21:32 |
| sarnold | wiredfool: http://bazaar.launchpad.net/~kirkland/pollen/trunk/view/head:/README | 21:32 |
| kirkland | sarnold: right, if you want it to block, it should be pointed at /dev/random, not urandom | 21:34 |
| sarnold | wiredfool: one of our users put this together, but I didn't get around to trying to integrate it into our distribution in time for trusty: http://www.av8n.com/cgit/cgit.cgi/init-urandom/ | 21:34 |
| Diegonat | hi.. With iftop I see a connection to a IP. I'd like to understand what program is connecting to that ip. How can I do it? | 21:36 |
| wiredfool | sarnold: that sounds like what I want | 21:36 |
| sarnold | Diegonat: netstat -nlp is very useful for this | 21:36 |
| sarnold | Diegonat: (sorry for missing it earlier..) | 21:36 |
| sarnold | wiredfool: yes :) it sounds nice. granted it isn't perfect because the kernel's api really does suck here. :( | 21:37 |
| wiredfool | sarnold: this is what I'm getting from pound, a ssl terminating load balancer: http://pastebin.com/UPU2EFhv | 21:37 |
| Diegonat | thank you sarnold | 21:38 |
| wiredfool | sarnold: looks like open, read, close. | 21:39 |
| Diegonat | mh... sarnold I cannot see anything with your command | 21:39 |
| Diegonat | I still see that ip | 21:39 |
| sarnold | Diegonat: if you run it as root you'll see which process is doing the sending and receiving | 21:40 |
| wiredfool | sarnold: incidentally, when pound starts, /proc/sys/kernel/random/entropy_available == 0 | 21:40 |
| Diegonat | ok i see it | 21:40 |
| sarnold | wiredfool: yeah, it'd be nice if you could delay that read() command until after the entropy pool has been filled some.. | 21:40 |
| Diegonat | but i see an IP | 21:40 |
| Diegonat | a remote IP | 21:40 |
| Diegonat | i want to understand what is connecting to it | 21:41 |
| sarnold | wiredfool: if you set pound to wait to run until after pollinate, you can at least be certain that some entropy has been shoved into the kernel -- it won't change the kernel's conservative entropy measurements, but it -will- provide the pool with unique input | 21:41 |
| sarnold | Diegonat: which program has the socket open? | 21:42 |
| Diegonat | no which program is talking to this remote ip | 21:42 |
| wiredfool | sarnold: I'm wondering if the urandom seed file is early enough in the process that I'm getting the seed from there, it's just not updating the estimate | 21:42 |
| sarnold | wiredfool: hrm? | 21:43 |
| hallyn | mdeslaur: jdstrand: hey, how do you feel about the lxc option in virt-manager and virt-install? and in particular about disablign it? | 22:47 |
| hallyn | zul: ^ | 22:50 |
| mdeslaur | hallyn: I don't really have an opinion on it | 23:01 |
| hallyn | ideally we'd have a config file where we coudl leave lxc out by default and have user override it if they really really want | 23:02 |
| hallyn | but if ppl aregonna stumble into this whiel just playing around, this is pretty much exactly what i didn't want (supporting another, untested lxc) | 23:02 |
| sarnold | I know I've seen at least one person surprised that the libvirt "lxc" wasn't as safe as the lxc lxc.. | 23:03 |
| hallyn | i'd be surprised too if id idn't know | 23:03 |
| === ABC-XYZ_ is now known as ABC-XYZ | ||
| ivoks | hallyn: ping | 23:34 |
| ivoks | serious issue with pacemaker in 12.04 | 23:35 |
| ivoks | https://github.com/ClusterLabs/pacemaker/commit/03f6105592281901cc10550b8ad19af4beb5f72f | 23:35 |
| ivoks | marked as low, but really has a big impact | 23:35 |
| ivoks | pacemaker node might refuse joining existing cluster cause of this | 23:36 |
| ivoks | roaksoax: ^ | 23:36 |
| roaksoax | jamespage: ^ | 23:38 |
| roaksoax | ivoks ^ | 23:38 |
| jdstrand | hallyn: disabling it makes some degree of sense. I have always hoped that libvirt could be made to use our lxc. that said, I thought I saw some patches go by on the list for apparmor integration for their lxc | 23:39 |
| jdstrand | hallyn: maybe from suse? not sure where they came from... maybe within the last month? | 23:39 |
| ivoks | roaksoax: really? i notified you :) | 23:39 |
| zul | hallyn: i dont have an opinon on it | 23:50 |
| zul | jdstrand: they came from suse | 23:50 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
