=== Lcawte is now known as Lcawte|Away === markthomas is now known as markthomas|away === markthomas|away is now known as markthomas === markthomas is now known as markthomas|away === mgagne is now known as Guest39926 === cripperz is now known as CripperZ === zz_DenBeiren is now known as DenBeiren === kickinz1|afk is now known as kickinz1 === CripperZ is now known as CripperZ` [07:56] Is there a way to specify a preferred dns server which is always used/tried first? I use one internal dns server (which resolves local names and redirects other requests to a public dns) and a public dns for when the nternal one is unreachable. The problem now is that Ubuntu seems to be switching to the second (public) dns I have entered in NM... === Lcawte|Away is now known as Lcawte === kickinz1 is now known as kickinz1|afk === kickinz1|afk is now known as kickinz1 === DenBeiren is now known as zz_DenBeiren === ashleyd is now known as ashd [11:35] Phibs, icehouse is in main for 14.04, so it's not in the cloud archive [11:36] Phibs, icehoues is in the cloud archive for 12.04 though [11:47] i did apt-get update && upgrade, i have still OpenSSL 1.0.1f which are vulnerable [11:53] arcsky: distributions backport security patches. To verify if a particular vulnerability is fixed, you need to examine the package version number. [11:56] rbasak: for example i have bash 4.3. and found this: http://www.cvedetails.com/cve-details.php?t=1&cve_id=+CVE-2014-7169 [11:56] do i have to dig deeper? [11:57] arcsky: go to http://people.canonical.com/~ubuntu-security/cve/ [11:57] arcsky: enter the CVE [11:57] arcsky: that will show you the status in Ubuntu. If fixes have been released, you will see the corresponding package version numbers. Then check that you have at least those package versions installed. [11:57] arcsky: or, just make sure that you have the security repositories in your sources.list and that you keep up to date. [11:58] rbasak: nice thanks for that url! [11:59] you should have something like, 1.0.1f-1ubuntu2.11 that is not vaunerable, not 1.0.1f === zz_DenBeiren is now known as DenBeiren [12:02] 1.0.1f-1ubuntu2.11 i have te [12:02] ye* [12:02] The requested URL /~ubuntu-security/cve/pkg/1.0.1f-1ubuntu2.11.html was not found on this server. === DenBeiren is now known as zz_DenBeiren [13:12] anyone running a newer version of php on 14.04 ? [13:13] standard is 5.5.9 [13:21] YamakasY: Running php 5.6.7-1 currently using https://launchpad.net/~ondrej [13:28] BlackDex: works OK ? [13:28] YamakasY: Yea works nice, no problems with it. [13:29] I have it combined with nginx and fpm [13:29] BlackDex: mhh my ubuntu apt-get upgrade doesn't see them [13:29] have you added it with add-apt-repository ? [13:30] and then did a apt-get update? [13:30] yes [13:30] Strange. With my server it whas just as simple as that [13:31] might be a puppet issue [13:31] Ah!. LoL that could be. That puppet fixed versioned them [13:41] YamakasY: check it with something like this apt-cache policy php5 [13:50] BlackDex: yap might be indeed [13:51] BlackDex: can I change that ? [13:53] YamakasY: The best way to change that is to change the puppet config [13:54] And even let puppet add the other repo so that it will be correctly configured when puppefied again [14:02] BlackDex: I know, but this is just a wrapup test [14:05] Ah! then just do this. `apt-get install php5=5.6.7+dfsg-1+deb.sury.org~trusty+1 [14:05] That will overrule the pin [14:07] To check for which versions you can install do: `apt-cache madison php5` [14:08] heh, puppet makes my life too easy [14:09] doesn't override [14:09] it says it installs php5 but it doesn't [14:28] got access gateway replaced. today. Got home today after work it seems ip address on linux box connected to access gateway wasn't updated, and needed to do service networking restart to force pc/ubuntu to update eth0 (interface connected to access gateway) not sure if it's hardware(mobo) or ubuntu that doesn't update itself auto ? what to check out whats not working properly ? [14:30] Blueking: Does it do DHCP? Long lease time out? [14:30] eth0 should get ip from access gateway (ISP' property), dhcpd runs on eth1 === leTali is now known as Talyian [14:31] lordievader [14:33] Can still be a long lease time. [14:33] long lease time given by access gateway or lordievader ? [14:34] If that is the one who gives out the leases, yes. Could be a reason. I'm not saying that is the reason. Just a suggestion. [14:34] what should I do about lease time when I am using linux box as router ? [14:36] lordievader aren't there any sense function on ethernet interfaces ? [14:37] Blueking: As long as they are connected they don't care. [14:37] If that means you have a correct ip or not, they don't know. [14:38] lordievader got new access gateway today, the new ip wasn't replaced before I logged on linux and 'service networking restart' [14:39] mhh I need to override this pinning [14:40] lordievader linux box had been online for 16 months without any issues/problems, but last 2 weeks net hes dropped out 1-2 times a day [14:42] Blueking: that means you need to buy a new one :P [14:42] buy new what ? [14:43] YamakasY ? [14:43] server :) [14:44] naw supermicro mobo with xeon cpu and ecc ram shouldn't die that fast not even 2 year old [14:44] YamakasY [14:45] Blueking: Investigate why the network connection is dropping out ;) [14:47] I suspected it was old access gateway that was reason [14:48] I'll check if there are any bios updates to mobo [14:48] Randomly updating the bios isn't proper trouble shooting. [14:49] Investigate why the network connection is going down. Check logs. Run tests. Try the reproduce the problem. [15:15] hello! is someone interested in hearing about a strange phenomen? [15:16] i have an ubuntu server 14.04 and when i turn it off the next day it won't boot (only dots on the screen, grub doesn't open up) [15:17] then in the morning when i use a live cd with fsck and then reboot the server it works again until i reboot it again.... this happens now everytime [15:17] @SuperMX is it a virtual machine ? what fs type ? [15:19] no its an old acer computer from my uncle... windows xp used to run on there. so there is the bootable fat32 partition and then the linux fs with ext [15:20] ext 2 3 or 4 [15:20] i am quite sure its ext3 [15:21] maybe i am doing something wrong with the shutdown [15:21] well why don't try to check and repair ur partitions ? [15:21] u can use gparted for that [15:21] i always used fsck, should i try with gpart? [15:22] yeah definitely I prefer gparted [15:22] i thought that fsck repairs the fs as well [15:22] ah ok, i will try that [15:22] thank you :)) [15:22] gparted is easier and will show u useful info about what's wrong with ur partitions [15:22] then may be u need to reinstall grub [15:23] hmmm ok [15:23] so it will be definitly the fs [15:24] I think so [15:24] because when grub doesn't start it has nothing to do with ubuntu [15:24] true [15:24] do u have ur data backed up ? [15:25] a website is running on it, but this partition seems to be ok [15:25] but, yes i should back it up :) [15:27] thank you for your response and sorry if i wasn't technically exact [15:27] Gparted probably calls fsck, so using fsck is fine. [15:28] Perhaps it is a good idea to check smart data. [15:51] hallyn, zul, anyone having a pending upload to vivid qemu in the works? if not I would soonish push a small one [15:52] smb: i dont [15:52] me neither [15:53] ok. ta === markthomas|away is now known as markthomas === Guest39926 is now known as mgagne === bladernr` is now known as bladernr_ [18:51] Raid Question. Can you assemble a raid 60 with uneven amount of drives, for example 11? [18:52] No. [18:53] Erm on moment - uneven drive count, and RAID6? [18:53] Are you testing on how to horribly slow down a system? :) [18:54] I have a 12 disk system, need as much storage as possible with a good amount of stability. [18:54] Which drives? [18:54] 6tbSATA [18:55] Do you have a RAID controller? [18:55] freeNAS ;P [18:55] lucidguy: sounds like a problem tailor made for zfs, using one 11-disk raidz3 vdev [18:55] Yeah. [18:55] bekks, yes PERC H730P [18:56] Raid05 supports uneven amount of drives? [18:56] I'd start with workload. e.g. running RDBMS or other? make sure IO needs are met, and if they are, then go full raid6... or... PERC... then use built in PERC raid5 :) [18:56] raid50 [18:56] Uneven drive counts dont make sense for striping. [18:57] dont make sense? but you can raid 5 on three disks. [18:57] Yeah. So double that count. [18:58] Gotcha: So raid5 and raid6 can support uneven drive counts, but not 50 or 60 [18:58] Makes sense [18:59] So I could go a full 12 disk raid 60, gives me 4 disk fault insurance, over raid6 2 disks. [19:00] USe two hotspares and a single RAID6. [19:01] Why does this site/app allow me to setup 11 disk raid60.. should error/note.. https://www.icc-usa.com/raid-calculator/ [19:01] Ask that site? :) [19:02] Even if it is possible, it doesnt make any sense. [19:02] So its just silly, and misleading right? [19:02] Correct. [19:03] RAID calculators calculate things, they do not tell you "that looks good" or "thats quite stupid". [19:03] zfs would let you do that with two vdevs, 5disk raidz2 and 6 disk raidz2, but the tools would ask you for comfirmation before creating it, since it would be unbalanced; writes would prefer the six-disk raidz2 vdev first, since it has more free space [19:04] but if you could get twelve disks and do two raidz2 vdevs of six disks each you'd see more even writes and thus better available iops.. [19:04] sarnold, I have a decent raid controller, no zfs required. [19:04] ZFS is not a replacement for a RAID controller. [19:05] I know, but ideal for when you don't have one. [19:05] are you talking about formatting a hardware raid array with zfs? [19:05] lucidguy: most raid controllers won't do compression or checksumming.. [19:05] Or deduplication. [19:06] zfs is great when you have a whole lot of disks. [19:06] no? [19:06] Which you have. [19:06] 12 is not alot [19:06] Depends on the point of view. [19:07] zfs does like to control the entire stack.. so HW raid isnt a great choice unless its the only choice :) [19:08] This is a 12 disk NFS server.. raid controller over not with zfs makes more sense [19:09] some people with hardware raid cards (without jbod support) set each disk as a 1 disk raid 0 and then make the zraid with those [19:09] yeah if you were to go zfs you'd want to stick your raid controller in hba mode.. [19:10] Then why bother with a raid controller? [19:10] exactly [19:10] yes :) [19:11] With my setup/needs you should get better results with a raid controller over zfs [19:11] Because you believe in that or because you actually tested it? [19:13] I believe in that. ZFS has an overhead on the CPU. zfs performs well with fast flash/ssd memory for log/cache. [19:15] Belief is a good sign for religion, but absoluty useless when it comes to technology. [19:15] IS my statement not a fact? [19:16] You stated technical facts that do not imply that your RAID controller is faster, actually. [19:16] you may see better performance on some benchmarks using xfs on hardware raid than using zfs on an HBA.. but you'd lose the compression, checksumming, snapshots, etc. it's all tradeoffs. [19:17] sarnold, I agree with that. We will be using XFS also. [19:17] with a hardware raid card I would do hardware raid.. with a hba I would probably do md raid.. I had issues with the zfs module building reliably in the past [19:18] I have an x4500 thumper running ubuntu with ZFS. [19:18] lucidguy: ooooooo [19:18] those are impressive machines even today, hehe :) [19:19] It was quite the ZFS learning experience for me. [19:20] My 8 year old 60 disk jbod server lost all of its storage a few days ago due to a simultaneous dual disk failure. Hence my interest in going raid6 or 60. Im interested in 60 due to the increased speed and better performance. [19:22] So you lost more than two disks on one side of the RAID60? [19:22] not this was a raid50 system [19:22] maybe you should ask mdadm? === katherine is now known as katherine_temp [19:24] in the case you are still ok because the other raid 5 still has all the data right? [19:25] I dont have much exp with nested raid levels [19:25] rberg, nope because one of the single spans lost two disks. [19:26] ohh no.. I think I played out a 05 setup in my head.. [19:27] lucidguy: So you used 50 setup consisting of two 30-disk parts? [19:28] one raid 50 was 28 disks by 4 spans of 7 [19:28] all DAS or was it iscsi/fc? [19:29] direct === katherine_temp is now known as katherine [19:31] I think ive decided. 11 disks total, 10 for raid60 and the 11th a hot spare [19:33] why the switch from zfs to raid though? [19:33] There is no switch, just getting a box with raid. [19:34] what kind of usage will it be having on it? [19:35] one large xfs volume exported via NFS. About 30+ hpc workstations will be using it for storing large images etc. The build is restricted to a budget of around $15k [19:36] large images? [19:36] seq accesses? random accesses? [19:36] mainly reads? or writes? [19:36] a mix [19:37] mainly sequential [19:37] large images as in MRI scans of brains etc. [19:38] so really reads [19:38] as writes would be limited to the speed and quanity of scanning machines [19:38] you could probably go much larger than 11 disks per set [19:39] as long as you feel you have enough parity disks to handle failures [19:39] but you had two disk failures [19:39] if raid6 has two disk failures, you will no longer know if your data is accurate anymore, or got corrupted === markthomas is now known as markthomas|away [19:44] i'm lazy. I'd do 11 drives, one raid6, 1 hot spare. :p [19:45] make sure you do daily scrubs [19:46] and setup email alerts [20:03] any app armor experts around? [20:04] depends [20:04] !ask [20:04] bot be broken [20:04] hehe [20:04] 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 [20:04] i have a log with what appears to be an issue with app armor+lxc, but i'm afraid i don't know what i'm looking at: https://bugs.launchpad.net/juju-core/+bug/1441319 [20:05] log: https://bugs.launchpad.net/juju-core/+bug/1441319/+attachment/4369706/+files/container.log.gz [20:05] " lxc-start 1426805367.662 ERROR lxc_apparmor - lsm/apparmor.c:apparmor_process_label_set:183 - No such file or directory - failed to change apparmor profile to lxc-container-default" [20:05] it's not the first error, but it's the one that sticks out to me === CiPi is now known as cipi [20:36] katco: if you think apparmor is involved, please also attach a gren DEN /var/log/syslog or /var/log/audit/audit.log (If you're using auditd) [20:37] sarnold: this is a ci server, i'll have to see if i can get those. i was hoping for an opinion to point me in the right direction before spiking too far in one direction. === markthomas|away is now known as markthomas === rylinaux is now known as rcx === rcx is now known as rylinaux === bilde2910 is now known as bilde2910|away === cipi is now known as CiPi === kickinz1 is now known as kickinz1|afk === Lcawte is now known as Lcawte|Away