[00:01] linuxgeek_: qemu and kvm projects merged; kvm package is just virtual now [00:02] ah cool TJ- [00:03] linuxgeek_: kvm only maintained a separate userspace whilst qemu didn't support the Kernel Virtual Machine [00:03] sarnold: and what version of qemu are you running? [00:03] rharper: ^ dunno if you're keeping track [00:04] so on trusty, qrt gives no failures. build utopic's libvirt on trusty, 1 failure. built utopic's virtinst, 2 failures. [00:04] got a few more to track down [00:06] linuxgeek_: yikes, note 13.10 reached the end of its life in july [00:07] heh? stay on a security updated branch? what? :) [00:07] hallyn: 2.0.0+dfsg-2ubuntu1.2 [00:08] oh, was hoping you were using one of the bisect kernels [00:08] uh, qemus [00:08] hallyn: no, I hadn't done any testing yet :( [00:09] just being greedy [00:27] When trying to run "sudo apt-get install git-core" I get this error "E: Package 'git-core' has no installation candidate", help, please? [00:28] I'm on Ubuntu 12.04.5 LTS [00:28] pixels: juju-core is only available in a PPA for 12.04 LTS [00:28] juju-core? [00:28] sigh [00:29] pixels: don't mind me, it's obvious I shouldn't be ehind a keyboard any longer [00:29] also add-apt-repository gives me a command not found [00:30] pixels: try apt-get update && apt-get install git-core -- perhaps your lists just need refreshing [00:31] getting a ton of Hash Sum mismatch [00:32] and it still doesn't work [00:32] pixels: which mirror did you get? perhaps it is broken .. [00:33] us.archive.ubuntu.com [00:34] pixels: try sudo netstat -tnp | grep TIME_WAIT -- perhaps you can find which IP address was used [00:35] if none look right or if they've moved out, try the apt-get update again [00:35] pixels: .. and check dmesg to make sure you're not getting hard drive errors. that can happen. [00:35] pixels: Is your system behind a captive portal? [00:36] sarnold: doesn't show anything [00:36] and i'm using a VM [00:37] pixels: okay, re-run the apt-get update and while it's running, check netstat -tnp | grep http [00:38] how would i go about that? [00:38] i'm using a linux host so ctrl-alt-f2 won't work [00:39] heh? what does control-alt-f2 do? [00:39] switches to a different tty [00:39] no [00:39] alt-f2/alt-f3/... [00:39] pixels: screen or tmux are handy [00:40] Patrickdk: what [00:40] can you be a little more clear [00:40] control is never used to switch tty's [00:41] or really, vty's [00:41] it's what i use [00:41] the control is only necessary if you're running X, which is usually unlikely on servers :) [00:42] ah i see [00:44] 91.189.91.13 91.189.88.153 [00:46] thanks [00:46] also i figured out where i can get a zip file for the git repo but i'm still willing to troubleshoot [00:48] pixels: okay, probably the 'best' thing to do is add some new repositorys to your APT configuration, hopefully another mirror will work alright [00:50] * Patrickdk wonders [00:50] * Patrickdk just found fiberstore [00:50] pixels: try adding a new file to /etc/apt/sources.list.d/temporary.list http://paste.ubuntu.com/8254869/ [00:50] had the spf+ I needed, and they are a pain to track down [00:50] ordered some cables too, see how they work [00:52] Patrickdk: heh, cable weight, "kg/km" [00:53] :) === peter is now known as Guest77101 === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === daveops_ is now known as daveops [03:58] I've had a server thrashing and saw this when "last reboot" [03:58] reboot system boot 3.0.57-rescue-x8 Thu Sept 4 08:43 - 08:43 (00:00) [03:58] is that recovery mode? [05:32] zul: hey, so i pushed one more (trivial apparmor) fix onto the libvirt package and pushed it to ppa:serge-hallyn/virt for a test-run [05:48] sssss === liam_ is now known as Guest22638 === kickinz1|afk is now known as kickinz1 === kickinz1|afk is now known as kickinz1 === kickinz1 is now known as kickinz1|afk === kickinz1|afk is now known as kickinz1 [07:26] Good morning. [07:26] good morning to all [07:27] Hey salih-emin, how are you? [07:27] I should I be ? ITs FRIIIIIDDDAAAAAYYY !!! :P LOLOLO [07:28] How* === Tribaal_ is now known as Tribaal === nabamx_ is now known as nabamx === Lcawte|Away is now known as Lcawte === deegee is now known as drussell [09:18] chaps, I'm having an issue with Rancid, has anyone successfully suffered with it? [10:06] CAn someone help me with vsftpd please. I thought I'd change the umask so that all uploaded files have group rwx perms (since I can't seem to find any text that says how to change the owner of the file from the ftp user) but it hasn't changed [10:06] all uploaded files still have the 600 perms === trijntje_ is now known as trijntje [10:07] s/rw/rwx [10:33] i have 8 disks. sda to sdh. i want 200MB, , 20GB and these 3 partitions on *each* disk. then i want to create md0 for all eight 200mb partions. mount /boot on md0. then creaet md1 for all eight 20GB partitions. mount / on it. is this the correct recipe for this raid1 setup http://paste.fedoraproject.org/131210/99128401/ [10:45] hello! I have a deployable ubuntu-server 14.04 image for servers and it contains a fstab entry that doesn't exist on all systems (and I used "nofail" as mount option), so upstart prompts me to type S to mark all filesystems as mounted. how do I tell upstart to skip the fstab entry if the device does not exist? [10:49] psih0man: why not just get rid of the rogue entry on the affected server? [10:50] because I don't ant to [10:50] because I don't want to [10:50] I want it skipped [10:51] psih0man: Give it the option noauto? [10:51] cynicallemon: it's not just one server and I don't wat to customise the image for each and every case [10:52] lordievader: on systems that have the device, I want it mounted automatically [10:52] psih0man, sounds legit ... [10:53] the entry is "LABEL=ext-array1 /ext-array auto nofail 0 2 [10:53] " [10:53] the entry is "LABEL=ext-array1 /ext-array1 auto nofail 0 2" [10:53] psih0man: why not deploy the image then ssh into the server and delete entry - unless you have a thousand servers that is [10:54] cynicallemon: and if I have thousends of servers what do I do? [10:54] psih0man: do you? [10:54] cynicallemon: I do [10:55] that's why I administer servers not a desktop :) [10:55] then you should be looking at something like puppet maybe [10:56] cynicallemon: and what should I tell puppet? that I have some servers that need some custom setting? why doesn't upstart honor nofail? [10:57] it's my decision that it should not fail on error [10:57] I already use puppet [10:57] psih0man: you should be asking the devs, not me [10:57] but puppet-ing something looks like a hack around a bug [11:00] well upstart will be making way for systemd before long so you may have to find an alternative way anyway [11:01] psih0man: I think that upstart can't really tell the difference between the device not existing, and the device not having been hotplugged yet. [11:01] psih0man: an alternative might be to make it noauto, and to set up a separate upstart job to mount it on "hotplug". [11:01] This assumes that you don't need it to boot. [12:58] howdy everyone.. got some rkhunter questions... i just got a coupleof messages that are just oneliners that said the following: "Please inspect this machine, because it may be infected." [12:59] also, I have tripwire installed on my server.. I've gotten a number of different files added as of late.. I do remember update and upgrading, so that may have something to do with it.. I guess, where do i start when it comes to fixing up all of this? [13:01] then, inspect it :) [13:01] do you know where the rkhunter logs are by chance? [13:05] theToastIsDone, you seem to be reading this as "something is brokeded!" If rkhunter had detected a threat, that alert would have been quite explicit. [13:05] ah ok, sounds about right, cause there doesn't seem to be any problems [13:05] i appreciate it [13:15] psih0man, you might be able to use autofs. http://linuxconfig.org/automatically-mount-usb-external-drive-with-autofs. when the device is detected it's the same sort of hotplug event as inserting a usb drive [13:18] ppetraki: that might be a good idea -- thanks [13:19] psih0man, udev will devine the fslabel from the udev helpers, just run it in debug mode or whatever to profile the device and adjust the example accordingly [13:20] s/devine/divine : drinks turboshot [13:21] ppetraki: I was thinking of creating an upstart service (which is not a good idea since Ubuntu will be moving to systemd) to mount things at the end of the boot process. over fstab, a script has the advantage of being able to treat errors and conditional execution of mounts (a Turing complete fstab, if you like) [13:22] ppetraki: but autofs seems to be a ready to use solution [13:25] psih0man, yeah autofs is one of those "oh that's there" features that people forget about until you need something slick. Like I'm using it to automount an sshfs on a development node, so it forwards my build sandbox to the node in question automagicaly [13:25] psih0man, and it looks like systemd is integrating it, surprise [13:30] ppetraki: yeah... systemd is integrating everything :) I was reading a lot of flame wars on this topic lately - but I don't have yet a position about the issue. I tend to agree with Lennart. thus, a minimal Linux instalation will consist of only 2 projects: the kernel and systemd. it'll look more like FreeBSD's minimal system where all system essential components are closely bouded together and additional programs are not part of the base system [13:31] Does SSH always have an encrypted session no matter what? [13:32] !ssh [13:32] SSH is the Secure SHell protocol, see: https://help.ubuntu.com/community/SSH for client usage. PuTTY is an SSH client for Windows; see: http://www.chiark.greenend.org.uk/~sgtatham/putty/ for it's homepage. See also !scp (Secure CoPy) and !sshd (Secure SHell Daemon) [13:33] hi all. We're trying to install ubuntu 12.04 via netboot by using the latest HWE netinstall. We receive a kernel modules mismatch error [13:33] "No kernel modules were found. This is probably due to a mismatch between the kernel version used by this version of the installer and the kernel version available in the archive" [13:33] we have the modules in our repo... do we need to add some special line to the preseed? [13:35] Vladimir_, I would assume so [13:51] ppetraki: thanks man === JanC_ is now known as JanC [14:02] and you are 100% sure the kernel and initramfs that tftp is providing is the same version as the modules on the nfsroot? [14:02] you know NM I didnt see netboot there [14:08] Nothing more annoying than a box which ignores CPU power governers [14:08] and now I have two such boxes :| [14:08] Dell hardware is absolutely terrible [14:08] I will never buy anything Dell, as long as I live [14:09] weeb1e: you can select in its firmware setup "Active power management" or "OS power management" instead if "Maximum performance" which is the default [14:09] defaults are never good [14:09] psih0man: All of those have been tried, nothing can stop the CPU scaling down [14:10] No matter what, the CPU scales down to 1.6ghz and screws up my real-time services [14:10] ah: you mean you want your CPU to always be at full speed [14:10] Indeed [14:10] I was thinking it's the opposite [14:11] OS power management + any governor, including userspace with a set frequency for each core, is completely ignored [14:11] and the cores continue to scale down to 1.6ghz on both these boxes [14:11] Yet all my other boxes work perfectly with a simple performance governor set [14:11] what happens if you select "maximum performance" in firmware Setup [14:12] It continues to scale down [14:12] and the bios firmware is the latest [14:12] weird... [14:13] weeb1e: how's the CPU's temp? [14:13] Alright, what do you guys think has gone wrong here, http://paste.ubuntu.com/8259769/ [14:13] dasjoe: Fine, this is not CPU throttling [14:13] I can run cpuburn 8 times, and max out all 8 physical cores, with extra turbo boost speed without any thermal throttling [14:15] So the only conclusion is, Dell hard coded their bios firmware to ignore all options and software configuration, and simply scale down no matter what [14:15] and I was lucky enough to get two identical boxes which have this issue [14:17] weeb1e: try adding "processor.ignore_ppc=1" to your kernel command line, it should make the kernel ignore your BIOS's requests [14:17] weeb1e: "placebo configuration options": http://www.psmag.com/navigation/nature-and-technology/technology-deception-elevator-crosswalk-programming-robots-lie-89669/ [14:17] You can try it by "echo 1 > /sys/module/processor/parameters/ignore_ppc" [14:18] I'll try that now [14:21] dasjoe: Does not help [14:23] I've ever written a custom script to force the kernel to not allow the CPU to enter C states higher than C1 [14:23] But the scaling still causes a noticable performance hit [14:27] Without the ability to stop scaling, these boxes are essentially very expensive 1.6Ghz netbooks [14:32] weeb1e: have you tried setting the min speed to the max speed? or does it ignore that as well [14:32] rberg_: Of course, it ignores every possible option of configuration [14:33] have you tried complaining to dell [14:35] The hardware is out of warranty [14:36] weeb1e: Could be weird DSDT stuff, I'd try various settings for acpi_os_name or acpi_osi [14:37] I have no idea about DSDT, but I will do some research when I get a chance [14:39] weeb1e: the arch wiki has some information: https://wiki.archlinux.org/index.php/DSDT [14:41] Compiling a custom kernel for a box I don't have physical access to, would be quite a pain [14:42] Due to the infrastructure hosting these two boxes, I cannot get remote hardware access [14:42] and of course, this may not end up helping at all anyway === allegrem is now known as Guest72233 [14:48] This might be an unpopular question, but are there any decent, free, webpanels out there for Ubuntu server administration? That supports popular/large services etc. === allegrem is now known as Guest87877 === blkperl_ is now known as blkperl [15:29] Alright, what do you guys think has gone wrong here, http://paste.ubuntu.com/8259769/ [15:34] LucidGuy, looks like you're (xfs) getting pushed out by page cleanup [15:34] LucidGuy, you didn't do a bunch of fs stuff and then type 'sync' did you? [15:36] ppetraki, I did not. [15:40] LucidGuy, looks like memory pressure in one form or another, I'd have to look at what kswapd is doing to give a more informed answer. short answer is competition for free pages == lockup [15:40] LucidGuy, buggable, if you can reproduce it [15:42] ppetraki, I can't, this filesystem has been plagued by xfs/nfs instability for years. Its an NFS server exporting my users home dirs. I gave up awhile back and picked up a new server with a newer version of the OS, transferred the data, setup nfs etc .. and now this server is unstable. grrrrrrrr [15:44] LucidGuy, so you still have the old server around to reproduce the issue with? [15:45] ppetraki, the old is in production doing something else. I was never able to reproduce the issue [15:45] performed numerous xfs_repairs .. [15:45] so annoying [15:46] shit .. have to run. === apb_ is now known as apb1963 === iggi_ is now known as iggi === snuffeluffegus is now known as bob_dobbs === snuffeluffegus is now known as gurpgork [17:07] Anyone know how I can purge mysql from a system if it still says 'Unable to set password for the MySQL "root" user' when attempting to install it? [17:09] I have tried a few times now, and the following is not sufficient to fix this issue: apt-get -f install; apt-get remove --purge mysql-server mysql-client mysql-common libdbd-mysql-perl libmysqlclient18:amd64 mysql-client-5.5 mysql-client-core-5.5 mysql-common mysql-server-core-5.5; apt-get autoremove; apt-get autoclean; rm -rf /var/lib/mysql [17:10] ppetraki, any other ideas? [17:11] LucidGuy, play with your dirty pages ratio http://www.cyberciti.biz/faq/linux-kernel-tuning-virtual-memory-subsystem/ [17:12] LucidGuy, basically keep everything in ram as much as possible [17:12] LucidGuy, and I hope you have a RAID with a WB cache because individual drives are going to be stupid wrt caching [17:13] I finally found a solution, after so many "solutions" which did not help at all. If anyone ever has that issue, you need to use: echo "exit 0" >> /etc/init.d/mysql; dpkg --configure -a; dpkg --configure -a [17:19] Or not, it's still broken [17:19] Wow, if I had physical access I would have formatted the box by now [17:19] weeb1e: What is exactly the problem? [17:20] It's a brand new box, which I just started installing base dependencies on, the mysql install failed half way through due to a typo in the repeated password, combined with stdin being piped to /dev/null [17:21] Now everytime I try to install mysql, it says 'Unable to set password for the MySQL "root" user' [17:21] No matter what I do, I cannot purge the system of whatever is storing the bad state [17:22] weeb1e: sudo dpkg-reconfigure mysql-server [17:23] lordievader: dpkg-reconfigure: mysql-server is broken or not fully installed [17:24] weeb1e, sudo apt-get purge mysql-server; sudo apt-get install mysql-server [17:24] assuming it's never been cofnigured with data *maybe* purging the cnfigs with the remove would work [17:24] teward: Please read up slightly, to where I showed everything I tried to purge [17:25] weeb1e, there's nothing for me to read up on, client doesn't have scrollbac [17:25] apt-get -f install; apt-get remove --purge mysql-server mysql-client mysql-common libdbd-mysql-perl libmysqlclient18:amd64 mysql-client-5.5 mysql-client-core-5.5 mysql-common mysql-server-core-5.5; apt-get autoremove; apt-get autoclean; rm -rf /var/lib/mysql [17:25] That is not sufficient to fix this [17:25] zul: ok so hold back on that libvirt, bc it most definately breaks qrt. [17:26] okie dokie [17:26] could be just a mismatch of bindings to libvirt, still not sure [17:26] man am i gonna have to bisect? with libvirt? [17:27] weeb1e: Still, run what teward said. [17:27] weeb1e, or if you really want to mess around with it, use dpkg instead of apt to purge those packages (suggestion based on http://askubuntu.com/questions/253023/unable-to-set-password-for-the-mysql-root-user) [17:28] Sorry, I think I forgot "apt-get purge" from that list of commands I tried, I tried that too, and now tried it again, and it definitely does not help [17:28] teward: I really don't have a choice but to "mess around with it" [17:28] mysql is one of a ton of dependencies I need to install [17:29] teward: I have tried 10 different "solutions" to that error [17:29] None of which work [17:29] I've searched the hard drive for anything with mysql in its name too [17:29] I am completely out of ideas here [17:30] Like I said, at this point, it would be faster to format and reinstall ubuntu, but I don't have physical access [17:30] I can't believe, that of all things, installing mysql is causing an issue [17:31] weeb1e: Could you pastebin the full output of "sudo dpkg-reconfigure mysql-server-core-5.5"? [17:31] lordievader: I already pasted it to you [17:31] That was it [17:31] The full output, there is nothing else [17:33] weeb1e: Hmm, "sudo apt-get purge mysql-server mysql-server-5.5 mysql-server-core-5.5" [17:35] lordievader: I just tried running "apt-get purge" on that full list of mysql-related packages I mentioned above and still no change [17:36] weeb1e: Compare the list I gave with yours. Mysql-server-5.5 was missing from yours ;) [17:37] hm, maybe i'ts just this umask buglet [17:38] lordievader: Just tried again with that versioned package (which I assume mysql-server resolves to anyway), no change [17:39] weeb1e: Are they all purged? [17:40] `dpkg --get-selections | grep mysql` outputs nothing, so I assume so [17:42] weeb1e: Ok, navigate to /var/cache/apt/archives, locate mysql-server-core-5.5 and install that using dpkg. And please pastebin the full output. [17:43] lordievader: No errors from `dpkg -i mysql-server-core-5.5_5.5.38-0ubuntu0.14.04.1_amd64.deb` [17:43] lordievader: http://pastebin.com/AtccZNXL [17:44] weeb1e: Ok, next step do the same for mysql-server-5.5 [17:46] lordievader: mysql-server-5.5 pre-depends on mysql-common [17:46] weeb1e: Install that one first then. [17:48] lordievader: Same password error, console output: http://pastebin.com/3uLd4E9f [17:49] weeb1e: Where do you see a password error? I see that mysql-client-5.5 is not installed: sudo apt-get install -f [17:50] lordievader: On the ncurses mysql installer UI [17:50] weeb1e: You do get the option of setting a password? [17:51] Yes, after the second repeated entry of the password, it goes back to console for a split second and then returns to the mysql installer screen with the error message [17:51] Unable to set password for the MySQL "root" user [17:54] weeb1e: So you enter the password twice? [17:54] Yes [17:55] weeb1e: Does the mysql error log state anything? [17:59] hi [17:59] lordievader: http://pastebin.com/GmwmQPRg [18:00] hey guys, is the spamassassin corpus (https://spamassassin.apache.org/publiccorpus/) the best corpus to train my dspam? [18:01] I'm also looking for ham / spam corpus in Portuguese. Is there such a thing? [18:01] weeb1e: As I figured you have the same error "Can't create/write to file '/tmp/#sql_5e7d_0.MYI'" [18:02] weeb1e: See the answer of green7: http://askubuntu.com/questions/253023/unable-to-set-password-for-the-mysql-root-user [18:03] lordievader: I tried that hours ago, but let me give it another go, in case I missed something [18:04] Oh wait, green7, I missed that answer completely [18:06] lordievader: That does indeed seem to have solved it, even if I looked at the mysql log sooner, I would never have imagined that /tmp's permissions were messed up [18:06] I'd love to know what they did to this clean install, before handing the box over to me :( [18:47] this is the second time I setup a 14.04 server at hosting provider from self build image. [18:47] 20gb / 79GB /srv 1gb swap /usr mounted on /srv/usr and var also [18:49] there is a script while backing up that removes all tars zips and gzs === RiotBlasto_ is now known as RiotBlasto [18:56] streulma: is there a question in there? [18:56] no just to let you know [18:56] or [18:56] yes [18:56] how can avoid this that tars and zips are removed ? [18:56] all data is on srv [18:56] so backed up to home :) [18:56] ?? [18:57] what ikonia ? [18:57] I don't understand what your question is [19:02] Stop doing whatever you are doing that is removing the tars and zips. [21:49] How do I remove password of a user, and make sure they can't be used remotely? [22:10] miceiken: "man passwd" see "--delete" === Lcawte is now known as Lcawte|Away [23:24] I have a variable A that contains the value 'COUNT', I want to set the variable $COUNT to 5. I tried $$A=5, but it doesn't work. [23:24] Can I do this ? [23:25] bananapie: it depends upon the language you're using [23:25] bash* [23:25] bananapie: and that sounds like a terrifying thing to do :) [23:26] bananapie: try ${${A}}=5 [23:26] I am writing a link monitoring script, and I want to use bash functions. I can't pass variables by reference [23:26] it's telling me that it's bad substition [23:27] bananapie: try: B=${A} ; ${B}=5 ? [23:29] yea, I tried that. Trouble is, it tries to execute the value of ${B}. [23:30] If there is no obvious way to do it, I am probably doing this wrong, so I'll try something else. [23:30] tahnks [23:33] ok, I have to use eval :(