=== WiuEmPe is now known as Guest88415 === IdleOne is now known as Guest18511 === markthomas_ is now known as markthomas === DavidDuffey is now known as dduffey === huttan_ is now known as huttan === Metacity|uh-oh is now known as Metacity === daker_ is now known as daker === jeremy_carroll_ is now known as jeremy_carroll === slangase` is now known as slangasek === markthomas is now known as markthomas|away === rcj` is now known as rcj === TJ- is now known as Guest8103 === TJ_Remix is now known as TJ- === cipi is now known as CiPi === G_ is now known as G === Guest88415 is now known as wmp === Lcawte|Away is now known as Lcawte === Lcawte is now known as Lcawte|Away [08:32] smb: http://dpdk.org/ml/archives/dev/2015-September/023183.html === Lcawte|Away is now known as Lcawte === lukasa is now known as lukasa_away === lukasa_away is now known as lukasa [13:24] Hi there, I was wondering if anyone could help me with a samba/acl issue i am having please, I have made a post on the forums here http://ubuntuforums.org/showthread.php?t=2294351 which is probably easier to understand than me trying to type it out here, thanks in advance [13:45] Hey, I need some help setting up unattended upgrades on my ubuntu server. I installed and configured it, but it does not get run automatically somehow... [13:45] I used this infopage to set it up https://help.ubuntu.com/community/AutomaticSecurityUpdates [13:45] funkenstrahlen: what makes you say it's not run? [13:48] teward: I checked back after 14 days. could not find any logs in /var/log/unattended-upgrades. Then I ran apt-get upgrade manually. some updates where found. so I ran unattended upgrades manually and a log was created. [13:48] teward: So this is evidence for me it did not run at all [13:49] my apt.conf.d looks like this: https://gist.github.com/funkenstrahlen/2153d78dc2a786bb9c4d [13:49] so periodica intervald of each day is set [13:50] I *think* I might now the problem, but do not know enough about how unattended upgrades work [13:50] I found a script in /etc/cron.daily/apt.disabled [13:51] funkenstrahlen: OS? [13:51] I assume that by default this apt script runs the unattended upgrades? [13:51] or rather the version [13:51] teward: ubuntu server 14.04 lts [13:52] I thought maybe the *.disabled blocks it from getting run? [13:52] but why is it .disabled? there is no word of that in any tutorials [13:53] and because it works fine when I run unattended upgrades manually, I do not think there is a config file error. Its just not getting called at all [13:53] any thoughts on the use of serverpilot? If any of you use it [13:59] teward: still there? any idea? [14:00] prudentmav: never used it but it looks interesting. [14:01] funkenstrahlen: unattended upgrades doesn't just run apt-get upgrade on a cron [14:02] dft I used it on a development environment and it did save a lot of time but just hesitant to use in a production environment [14:02] it picks specific package types [14:04] thebwt: yes I know. still I would like to know why my unattednded upgrades does not get run daily [14:05] prudentmav: I hear you. If serverpilot goes away, what will take to unplug your VPS's from it and take manual control of your wp sites [14:05] yeah that is the big unanswered question for me [14:06] I think I would rather [14:06] ack [14:07] typing spasm [14:07] rather use aws and bootstrapping if I was in the business of building wp sites for all sorts of clients. [14:10] I think I would rather roll my own wp/vps builder using AWS's api. [14:14] I use digitalocean and just tired of sysadmin tasks... I'd rather stick to app and front end dev work [14:39] Hi Guys, I'm running ubuntu server 15.10 with 4.1.0.3-generic kernel. And updated to 4.2.0-7generic. But it has not been live updated. I thought this was new in 4.0 kernel ? Any info on this ? [14:39] siebjee-: I imagine the tooling with apt isn't yet in place [14:39] jpds: is there a way to force this anyway ? [14:40] I'd realy like to see this happening! *_* [14:40] (Also, I don't think it's a good idea to run a non-LTS release on a server) [14:40] Its my private server, so don't really care about LTS on there or not :) [14:40] On production servers i would care about LTS [14:40] siebjee-: there is no such thing as live kernel updates. [14:41] On arch it is possible to update the kernel without down-time === lukasa is now known as lukasa_away === lukasa_away is now known as lukasa [14:53] is there any way of getting dlm-pcmk on 12.04? [14:55] frediz: from #ubuntu-devel: [14:55] 15:54 rbasak: I'm happy with ginger, for kimchi, you're missing a debian/copyright entry for ui/pages/help/gen-index.py which is LGPLv2.1+ and not Apache2 as debian/copyright declares [14:56] rbasak: https://launchpad.net/ubuntu/precise/amd64/dlm-pcmk [14:57] jpds: ? [14:57] rbasak: Err, that was for revolve. [14:57] np [14:58] jpds: right, I'm on trusty though [14:58] can I get those sources? [14:59] revolve: You said 12.04 [14:59] right [14:59] ah 14.04 sorry === notafadss is now known as notafads === Luke__ is now known as Luke === markthomas|away is now known as markthomas [18:32] so... it's definitely something related to the dhcp lease... [18:32] seems client side not server side [18:32] after about 30 minutes it drops.... running dhclient in terminal kicks it back in (don't even need to release the old lease...) [18:33] using dhcpcd as the daemon didn't resolve it... dhclient is back on but no clue why it keeps dropping [18:37] would you recommend hardware or software RAID? === moonligh- is now known as moonlight [18:43] Slugs_: I'd recommend sw raid [18:43] Slugs_: but it all depends the use of it [18:43] yeah alot of people are recommending sw [18:44] Slugs_: hwraid can be fine on the boot medium, since it outrules problems with boot sectors [18:44] I usually go for SW too, so I don't have to worry about finding a matching controller if it dies. [18:44] Slugs_: sw raid (or zfs if you're doing the planning well) is what I'd recommend for data [18:45] RoyK: zfs not ext4? [18:45] !zfs [18:45] For information concerning ZFS and Ubuntu, see: https://wiki.ubuntu.com/ZFS [18:45] Slugs_: zfs is something like mdadm+lvm+checksumming [18:45] wow [18:45] Slugs_: it's very secure, checksumming all over, but it's not very flexible [18:46] Slugs_: I'm using zfs in production, also on my home server, but I'm planning to go back to mdadm, for various reasons [18:47] i see [18:47] Slugs_: if you have a machine with ECC RAM and you don't want to add new disks to the raidset, use ZFS - otherwise, I'd suggest mdadm raid [18:47] well I’m trying to simplly setup 4, 4TB drives in a RAID0 config for performance [18:48] how do you even put "4TB drives" and "performance" into the same sentence [18:48] Slugs_: just don't do that [18:49] are there 15krpm 4TB drives now? [18:49] Slugs_: what sort of performance? iops or sequencial? [18:49] jelly: no [18:49] no btrfs love? [18:50] oh i see, the drives are too big for ‘performance’ im assuming your all saying [18:50] Slugs_: a single 4TB drive can do something like 200MB/s sequencially, far less with seeks, very far less [18:50] Slugs_: it all depends on how you look at performance [18:50] im trying to write about 80 mbps [18:50] constantly [18:50] Slugs_: 4TB drives spin at 7200rpm, meaning they may do 120iops [18:51] heh, that is a lot of iops for 7.2k rpm [18:51] Slugs_: a single drive can do that - a raid5 with four drives can certainly do that unless the CPU is something 10+ years old [18:51] patdk-wk: 'may' do - usually around 80 [18:52] i thoguht raid0 was better then raid5 since its only stipped [18:52] Slugs_: striped, not stripped [18:52] sorry yes [18:52] Slugs_: raid5 on 3 drives has about the same speed as raid0 on 4 drives [18:52] Slugs_: for reads [18:53] Slugs_: a wee bit less for writes [18:53] i need writes more then reads in this setup [18:53] my server hangs during the restart via cron . it fails to shutdown syslog-ng . any ideas? [18:53] Slugs_: and you'll be a sorry bastard if using raid0 with your data and a drive fucks up, and it will, beleive me on that [18:53] Slugs_: is the data important? [18:53] RoyK: come on with the language ! [18:53] RoyK: understood [18:54] Slugs_: if the data isn't important, use raid0 [18:54] Slugs_: if you need massive writes and the data is important, use raid1+0 [18:54] you mean, if uptime isn't important [18:54] patdk-wk: or data [18:54] why? backups cover important data, not raid [18:55] raid covers uptime [18:55] design to your requirements [18:55] patdk-wk: well, if you have a truckload of writing, backups won't be enough to get the important data if a drive is lost 23 hours after last backup in a raid0 setup [18:55] so just to be clear raid 1+0 is good of data *is* important as well [18:56] *if* [18:56] s/of/if [18:56] that is a backup failure then, backup every 5min, I do [18:56] Slugs_: with 1+0 you'll get the best of both worlds, good iops and good sequencial i/o, but you'll have to get a lot of drives [18:57] 4 not enough/ [18:57] ? [18:57] some servers I even have on continuous backup, once it's done, it starts again. [18:58] Slugs_: what are your requirements? what sort of backup do you use? etc? [18:58] raid6 is generally better than raid10 with limited disks [18:58] but it does have iop issues [18:58] Slugs_: are you on 10Gbps or just gigabit? [18:58] but better for protection [18:58] patdk-wk: obviously true [18:58] RoyK: I don’t care about backing it up, im using 1 gb (network) [18:58] raid10 has the issue if one disk breaks, you depending on the other disk to be fine, normally this is the case, and not an issue [18:59] Slugs_: what sort of i/o is this? are you just dumping data to this? [18:59] yes im wriring multicast video to disk [19:00] 80 mpbs [19:00] then use raid[56] [19:00] it's good for sequencial stuff [19:00] 80Mbps is rather a lot [19:00] use mdraid with a large chunk size and use raid-5 or better raid-6 [19:00] the bottleneck would be my drives? [19:01] 7200 rpm [19:01] with a large chunk size, the drives should do 100MB/s each [19:01] or more [19:01] and i can configure this with mdraid? [19:01] man mdadm ;) [19:01] ok [19:02] RoyK: Thank you for the advice [19:02] Slugs_: what is this - cctv stuff? [19:02] and guidance [19:02] no, NPVR [19:02] network PVR [19:02] for off-air channels [19:03] recording live tv [19:03] same thing, really [19:03] sure, just thought you wanted to know ;) [19:03] Slugs_: just use a large chunk size on that raid [19:03] ok [19:03] Slugs_: and use at least raid-5 [19:03] ok [19:03] Thank you again [19:04] Slugs_: your boss will hate you if the raid goes down and the data is lost [19:04] well i already hate him, so we will be even [19:04] Slugs_: also, better use a lot of spindles rather than a few large ones [19:04] yeah, more drives the better [19:04] Slugs_: the most important is the chunk size - 1MB minimum [19:04] it might even make sense to add more smaller capacity drives [19:05] ok [19:05] how many video streams are you going to record at once? [19:05] 4MB or 8MB may be practical [19:06] im recording 40, 3.5 Mbps streams at once [19:06] wow! cool. good luck :) [19:06] Slugs_: and monitor CPU use - it will be used for parity calculation - it should be pretty fast, but at the rate you're mentioning you may need a good CPU (or four) [19:07] 160 Mbps? not 160MBps? [19:07] Slugs_: sar/sysstat may be a good friend for long time monitoring, so may munin be [19:07] 4 core, Intel(R) Xeon(R) CPU E5-2609 0 @ 2.40GHz [19:07] sarnold: 160MBps won't work too well over gigabit ;) [19:08] RoyK: true :) [19:09] Slugs_: should do well [19:09] Slugs_: what is the current cpu usage over the last hour? [19:09] im not recording anything... [19:10] Hi. I'm trying to figure out an issue with unmodified cloud images. If I download and launch a 14.04 cloud image on openstack, when I try to update the kernel or otherwise run update-grub it hangs trying to "modprobe btrfs". Anyone have any idea why it might be doing that? [19:10] Slugs_: then install sysstat, enable it, install munin-node and perhaps munin if you don't have a munin installation somewhere and make nice graphs to see how it's behaving [19:10] awesome thank you [19:11] Slugs_: there are several other monitoring systems available, but munin is easy to install, so I'd recommend it if you don't have any experience in monitoring [19:11] i don’t so thank you [19:12] the more i learn the more i realize i don’t know [19:12] Slugs_: touché [19:12] Slugs_: that's philosophy ;) [19:12] RoyK: Yes ;) === rbanffy_ is now known as rbanffy [19:16] so still thirty minutes even after changing dhclient.conf to send dhcp-lease-time 600; [19:20] rbanffy_: did you restart dhclient? [19:21] restarting it works but i need it to do that automatically [19:21] like i shouldn't even have to do that [19:22] RevertToType: copy aside your dhclient systemd thingy, apt-get purge dhclient, and try out dhcpcd instead. This has just been too much hassle... [19:26] Hello, I'm trying to install ubuntu 15.04 server, I have also tried 15.10 server and the system gets stuck after "boot-efi.mount". I've typed up the messages I get on bootup: http://paste.ubuntu.com/12428947/ [19:26] Both 15.04 and 15.10 hang on the same message. [19:28] sarnold: i basically did that yesterday and dhcpcd didn't seem that good but i'll do a full reboot and try again :( [19:28] RevertToType: oh. :( [19:28] "virtual packages like 'dhcp-client" can't be removed [19:29] oi dpkg [19:29] dpkg is saying it's not installed [19:29] WUHT [19:30] update/upgrade time [19:31] RevertToType: dpkg -l '*dhc*' may show what's really going on.. [19:32] * RevertToType nods [19:32] boshhead: dang that's the worst thing to google for, ever. [19:32] sarnold: yeah there's no relevant results on google [19:32] so im in despair :P [19:34] oh wait.. maybe this is it "It sounds like your initramfs doesn't have the needed module for vfat." - im gonna try adding vfat to /etc/modules if it doesn't exist [19:35] after that ill try replacing grub with elilo or refit [19:36] sarnold: it lists it as unknown for desired action and not-installed [19:37] boshhead: mkinitramfs(8) suggests /etc/initramfs-tools/modules is the place to put modules [19:38] sarnold: oh right the other file is generated. thank you. [19:40] so is debootstrap for vivid just...broken? [19:41] rebooting, which me luck :) === markthomas is now known as markthomas|away [19:55] well, i have jenkins set up, Not insde a container. now i am trying to run a job, the job should clone a repo from github, and then run a bash script. , i am not sure how to tell jenkis where to clone the repo to. [20:01] try #jenkins ? [20:01] Thanks [20:30] this is gonna be the death of me... grabbing a second laptop- out of our bin--- starting from scratch... maybe i just left too much of a mess [20:32] jamespage: Sorry to poke you, did you guys notice any issues around python-cryptography/python-cffi on liberty packages with services like nova or keystone ? === markthomas|away is now known as markthomas [21:13] Slugs_: you'll eaither need to mux that back to some 1-4 streams to disk, cheap 4TB sata disks are not very likely to be able to work with 40 parallel writers [21:14] either ^^ or get a drive array that can actually survive more random io [21:16] if they're only writing for archival purposes and never reading, it might be doable [21:19] (and if you want users to also watch those saved streams, you're looking at SSDs or ramdrives, two orders of magnitude more expensive solutions) [21:20] is Ubuntu (15.04) capable of booting from root on multi-disc btrfs? It requires special btrfs device scan support. [21:21] I'm talking about mounting root, not actually booting, I know /boot can't be btrfs === lukasa is now known as lukasa_away [21:28] How can i create a new user and assign it ssh key ? [21:29] new user is adduser, but how can i assign it neww ssh keys [21:30] netameta, login as that user, and follow the guide ? https://help.ubuntu.com/community/SSH/OpenSSH/Keys [21:31] OerHeks, i dont think i can log in with this user [21:31] not before i have ssh keys for it [21:31] ec2 restrict access to only ssh keys [21:32] unless, its possible to create a user from the main user, and then use the new user and follow that guide [21:32] netameta, oh, not sure how ec2 works, they should auto make ssh keys when adding an user, logically [21:33] yea i have 1 pair for the main user [21:33] so i can log on their instan [21:33] but i need another user [21:33] Is it possible to create a user and then "mount" it ?