[00:00] I'm trying to set up courier but I get /usr/lib/courier/imapd.pem already exists. when I try to install courier-imap-ssl but if I delete it the output I receive is regarding a dead sym link. What should I do to properly deal with this? [00:07] Anyone set up openldap to authenticate Mac OS X Maverick clients against === keithzg_ is now known as keithzg [00:41] I've already run lvextend and lvdisplay shows the lv is much larger now [00:41] when I run resize2fs via sudo, I get a permission denied error. [00:41] http://paste.ubuntu.com/6962950/ [00:41] any help? [00:42] jrwren: umount the filesystem? [00:42] i want online resize [00:42] sarnold: shouldn't need to [00:42] jrwren: make yourself root [00:42] hey sheptard :) [00:42] sarnold: hoihoi [00:42] http://pastebin.com/jmWT98T0 Any takers? [00:42] same exact message from root shell started via sudo -i [00:43] jrwren: assuming the filesystem is ext3/4? [00:43] yes, ext4 [00:43] Bitwise: rm /usr/lib/courier/imapd.pem ? [00:43] Bitwise: ^ [00:44] oh, maybe one of those 'directories' is a symlink... [00:44] rm: cannot remove ‘/usr/lib/courier/imapd.pem’: No such file or directory [00:44] jrwren: I'd just sudo su; resize2fs /dev/derpderp [00:44] Bitwise: ls -ld /usr/lib/courier [00:44] same exact error via sudo su [00:44] drwxr-xr-x 4 root root 4096 Feb 19 19:37 /usr/lib/courier [00:44] jrwren: interesting [00:44] Bitwise: interesting [00:45] Bitwise: ls -ld /usr/lib/courier/imapd.pem ? [00:45] sheptard: indeed. googling gives me results with folks having issues with SD cards on raspberry pi :) I don't think those are LVM backed [00:45] ls: cannot access /usr/lib/courier/imapd.pem: No such file or directory [00:45] Bitwise: ok. confusing. try ls -l /usr /usr/lib ? [00:46] http://paste.ubuntu.com/6962964/ [00:47] really good info here actually: http://forums.gentoo.org/viewtopic-t-764297-start-0.html [00:49] sarnold, I think it will help to mention that I did rm -f /etc/courier/imapd.pem while trying to set up courier previously. [00:52] Bitwise: how about ls -l /etc/courier/imapd.pem ? [00:52] ls: cannot access /etc/courier/imapd.pem: No such file or directory [00:52] Bitwise: /etc/courier ? [00:53] http://paste.ubuntu.com/6962987/ [00:54] hello i am unable to install ubuntu server 12 lts x64 on a host-based raid1 array [00:54] the curses-based setups hangs when i say apply partitioning [00:54] Bitwise: that's all so confusing; the postinst script appears to run this program if the /etc/courier/imapd.pem doesn't exist: mkimapdcert [00:55] Bitwise: you could try running that as root and see how well that works, then re-try install? [00:55] cat /proc/mdstats goes into resync ... takes several hours until it finishes, and the setups is still frozen after successfull resync [00:56] sarnold, it seems I forgot how to Linux. [00:56] Bitwise: haha :) [00:56] I switched the parameters of ls -s [00:56] i had to reboot because the system hung and now it wants to resync again on applyng partitioing (tried it 2 times again, it always hangs, and starts to resync again) [00:56] Fixing that allows me to reinstall courier. [00:56] Bitwise: oh! ln -s ? [00:57] Bitwise: I -hate- ln. I can never keep the argument order in tact. [00:57] Yes, I tried to create the symlink so that it would work. [00:57] or intact either. :) [00:57] Bitwise: I'm just surprised the various ls -ling didn't show it..? [00:57] Is there anyone who knows what to do? [00:59] sarnold, I did. That's how I realized it. [00:59] Bitwise: sigh :) sorry I missed it. hehe. [00:59] It's no problem. I appreciate the help. :-) [01:07] i wonder if resize2fs is failing because teh system thinks the filesystem is dirty and needs unmounting and fsck [01:12] anyone aware of an online fsck for ext4 ? :) [01:13] jrwren: remount ro? [01:16] is that allowed? [01:16] I mean, I know I can remount ro, but can I remount ro, run fsck -y and then remount rw ? [01:16] I've never thought of doing that. I'm going to try it :) [01:16] yup [01:16] thanks hitsujiTMO === io is now known as IdleOne [01:17] i've been doing this too long to have never done that. where can I send the thanks hitsujiTMO ? :) [01:17] you can fsck a unmount or readonly filesystem. so remounting ro should be safe [01:19] It seems reinstalling courier doesn't replace /etc/courier with fresh files. Am I missing something here? [01:23] Bitwise: if you apt-get purge courier it'll clear all the configs, then a fresh apt-get install should put them back [01:25] Saved the day again! [01:33] hitsujiTMO: i was hoping I could force remount readonly like kernel does when there are disk errors. [01:33] I get mount: /mnt is busy [01:34] jrwren: with ? sudo mount /mnt -o remount,ro [01:34] yup [01:34] well, mount -o remount,ro /mnt [01:35] I'd think -o before device [01:37] jrwren: try adding the n flag [01:39] its not my root fs, shouldn't need it, but I'll try [01:39] jrwren: you may have to kill anything that has a file open with fuser [01:39] same thing. [01:40] yeah, the fuser/lsof thing is the same as unmounting. [01:40] altho i'm sure i've done this with / but on debian tho [01:40] i'm in ssh, but any way I can remote an alt+sysreq+U ? [01:40] or make that syscall or something? [01:40] did you use the sysrq "emergency readonly remount" thing? [01:40] * sarnold ^5s jrwren [01:41] I did not. [01:41] but I want to. [01:41] jrwren: hehe, I meant it as a question to hitsujiTMO -- I'm a bit happy you thought of the same thing [01:41] jrwren: echo "command key" > /proc/sysrq-trigger [01:41] jrwren: that'd be echo u > /proc/sysrq-trigger [01:42] 's' first if you want to sync [01:42] sarnold: hmm, not sure was a few years back [01:42] this is getting into crazy territory. I think I'll wait until I'm done reading from that FS. [01:43] I was terrified for my first offline filesystem resize :) trying to do it while online is just amazing to me.. [01:44] this isn't a first for me. [01:44] its a first for half online fsck [01:45] resizing an ext partition "should" be trivial enough to be able to do it even when its dirty. [01:45] afterall, i don't think it creates any new inodes [01:45] I'm here in this guide: http://flurdy.com/docs/postfix/#config-simple-imap how do I test to make sure everything is working properly? xD I previously set up roundcube but now when I go to log in it says "Connection to storage server failed." [02:02] hitsujiTMO: i'm getting the error for some reason. I wish I knew why [02:03] in this case I mean the resize error. [02:15] Nevermind, I didn't restart the services. === freeflying is now known as freeflying_away === mwhudson is now known as zz_mwhudson === freeflying_away is now known as freeflying [04:18] Hi. If I'm using Postfix and Courier (I installed postfixadmin to make it a bit easier) shouldn't I be able to log into RoundCube with the users I create in postfixadmin? [04:23] Bitwise: Can users log in via IMAP with desktop clients and just not Roundcube, or not at all? [04:25] Not at all as far as I know. I can't seem to figure out what the user/pass is supposed to be or which db it's in. === semiosis_ is now known as semiosis [04:40] Bitwise: Can you pastebin your courier config, minus any sensitive bits? [04:42] tonyyarusso, Which file exactly? [04:46] Bitwise: do you have an /etc/courier/authdaemonrc ? [04:48] tonyyarusso, Yes, http://pastebin.com/GWEJC7vi [04:49] I'm trying to follow this: http://flurdy.com/docs/postfix/#top [04:50] Okay, and then an /etc/courier/authmysqlrc I take it? [04:53] This tutorial puts them in a database called maildb [04:53] authmysqlrc: http://pastebin.com/9JC2cGjL [04:56] So yeah, see if you have anything in the users table of the maildb database [04:58] tonyyarusso, Last I checked there were no entries but at the same time I thought adding a user from postfixadmin would add a user in maildb. [04:58] Also, I now can't log into phpMyAdmin. My username was 'mail' for some reason. [04:59] If there are still not entries, then the next step would be to look at the postfixadmin config [05:02] No clue why but I suddenly can't connect to MySQL through socket ... (2) [05:12] Okay, I'm not getting the SQL error anymore. [05:14] tonyyarusso, I logged in as root on phpMyAdmin and I see the problem. I have like three different db's that contain user tables. === semiosis_ is now known as semiosis [08:45] if I am exporting something via /etc/exports is it possible to view all instances of that export being "imported" on other machines on the network!? [09:01] zetheroo: /var/lib/nfs/rmtab is a good start (but can contain stale entries if clients don't disconnect cleanly) [09:02] I am unable to connect to my DHCP home network. I can see my network card. I can ping local host, and I believe I have /etc/network/interfaces configured correctly. [09:07] shauno: great thanks! Is there any way to kinda clean up the file to show only current listings? [09:21] zetheroo: not easily, it's maintained by rpc.mountd, so it may just replace anything it still believes to be true [09:22] (man rpc.mountd being where this file & behaviour are documented) [09:24] ok ... and this is the easiest way to see where these exports are being mounted to .. [09:26] I actually just found 'showmount -a' via the same manpage, which I was unaware of. but it gets its data from the same file [12:56] so /etc/cron.d/hourly is like 0 * * * * ? [12:56] I mean /etc/cron.hourly is like 0 * * * * ? [12:57] I put a script in there, named it foo-bar and made it 0755, tested to run it manually and it worked. but crond never runs it. [12:57] it's 13:57 here now so I think it should run soon, if it's equal to 0 * * * * [12:58] I placed it in there 3 hours ago though [13:01] nope, nothing. I even did while sleep 1; do ps auxww|grep cron; done and I could see no hourly run. and now server time is 14:01 [13:02] made a more traditional cron in /etc/cron.d/foo-bar now 0 * * * * root /etc/cron.hourly/foo-bar [13:03] cron is usually not this hard ;) [13:19] Hello, I am trying to diagnose a memory usage problem with a non-repo app. pmap shows that 65404K is used by '[ anon ]', how can I find out what the 65404K is being used for ? === mjohnson15_2 is now known as mjohnson15 [13:54] i cant setup grub on my raid1 array (MBR) is this possible in general? or do i have to install it on e.g. MBR of /dev/sda ? [13:55] raid1ubuntu: sw raid? [13:55] raid1ubuntu: anyway - it should work well [13:55] just install grub on both disks (if it's sw raid) [13:55] RoyK: host based (intel embedded raid server technology ii) [13:55] raid1ubuntu: using that setup myself [13:55] raid1ubuntu: then it's probably fakeraid. better not use it [13:55] better use md raid [13:56] its md raid [13:56] RoyK: intel matrix raid, managed by mdadm is actually ok. [13:56] oh - didn't know that [13:56] RoyK: ditto ddf. all others are crap and thus only supported by dmraid =) [13:56] raid1ubuntu: I'm using mdraid mirrors for boot on serveral servers - works like a charm [13:56] RoyK: ubuntu setup automaticall detected it. and now wants to install grub on /dev/md126 [13:57] raid1ubuntu: installing grub onto /dev/md126 is correct. [13:57] that's wierd. I thought grub should go to the host drives [13:57] raid1ubuntu: i had success with UEFI booting off intel matrix raid with mdadm, but not with BIOS/mbr. [13:57] xnox: install on MBR (yes/no) in setup, i select yes [13:58] RoyK: no, the motherboard knows about the fakeraid array and thus has options to boot of the raid. [13:58] xnox: ok - never used fakeraid [13:58] raid1ubuntu: yeah, that's correct -> but my motherboard didn't like it and failed to boot =( [13:58] if i select NO it asks me for a device [13:58] raid1ubuntu: you should install into MBR. [13:58] raid1ubuntu: of /dev/md126 device. [13:59] xnox: ok, but that fails [13:59] raid1ubuntu: and in firmware configuration (ctrl+I on boot) make sure it's marked as bootable. [13:59] raid1ubuntu: yeah =/ works with UEFI just fine though.... [13:59] xnox: what would be the difference between using fakeraid and ditching that altogether for plain mdraid? [13:59] raid1ubuntu: in that case try installing onto a partition, instead of mbr e.g. /dev/md126p1 [14:00] xnox: why not the host drives? [14:00] ok i will try it [14:00] i have installed the whole OS on the raid1 array [14:00] RoyK: different ondisk format and different sets of events / controllers. [14:01] RoyK: if you install to the host drives, you can hose the intel-matrix-raid on disk format. [14:01] (cause our grub is big and uses extended area i believe) [14:01] RoyK: then again if all fails, one can try that.... [14:02] xnox: but what are the advantages of using fakeraid instead of just mdraid? [14:02] RoyK: plus motherboard bios/firmware sees and checks assembled array and tries to boot that.... not the underlying disks. [14:03] RoyK: in my experience fakeraid resysncs faster, and you can migrate it offline using firmware tools only - preboot. [14:03] (e.g. replace and rebuild a spare or failed drive) [14:04] RoyK: plus it works for dual boot installations as one can have windows on the same array with linux. [14:04] that's inded a plus [14:04] RoyK: whereas with mdadm one can't install windows on top of it. [14:05] RoyK: people say that "intel firmware sends better/faster notifications about failed drives" but i'm skeptical about that =) [14:05] guess I'll stick to old md for the moment [14:05] that or zfs [14:09] xnox: please have a look: http://picpaste.com/Photo_2014-02-20_15.07.12-1-f4s7bAFm.jpg i have no idea what to do [14:09] xnox: i tried /dev/md125 ..126 and md127 for grub but it always failed. i tried also /dev/sda [14:10] xnox: it says something about "... disk not found" in dmesg (see picture) [14:10] raid1ubuntu: have a non-raided bootloader elsewhere. [14:10] raid1ubuntu: or just use stock mdadm instead of intel-raid. [14:10] raid1ubuntu: do you need/must use intel-raid, or can you use any raid? [14:11] raid1ubuntu: https://help.ubuntu.com/12.04/serverguide/advanced-installation.html [14:11] i could use software-raid but i guess its slower than the intel one, right? [14:11] raid1ubuntu: no depends yes [14:12] raid1ubuntu: using raid, you are taking a penalty hit anyway. With raid1, your write is always the slowest of the two drives and reads are upto 2x faster. [14:12] raid1ubuntu: given processing / physical limitations it's a bit longer that slowest drive to write, and a bit less than 2x read gain. [14:12] raid1ubuntu: the difference between mdadm vs fakeraid is negligible. [14:13] xnox: yeah i know, i just mean raid1 on software is slower than raid1 host based, thats what i guess, otherwise the raid controller would make no sense^^ [14:13] raid1ubuntu: ..... intel matrix raid controller is software based, it's just software compiled on the rom on your motherboard. It's not actually server-grade hardware raid controller. [14:14] xnox: maybe i just only need to fix the "disk not found" problem [14:14] raid1ubuntu: you could use dmraid. [14:14] raid1ubuntu: why are you using mdadm with intel raid? === tjaalton_ is now known as tjaalton [14:14] mdadm with intel raid is not well supported yet. [14:14] i am not sure if it is intel matrix raid [14:15] raid1ubuntu: what does this say $ mdadm --detail-platform [14:15] raid1ubuntu: you can execute it from tty2 for example === railsraider_ is now known as railsraider [14:16] raid1ubuntu: if you are not sure what you are doing, stick with mdadm using linux software raid. It uses kernel md features and is very fast and reliable. [14:16] xnox: i have this board in this 1he server: http://ark.intel.com/products/29871/Intel-Server-Board-S5400SF [14:16] raid1ubuntu: i don't care what the board is, I care what the userspace detects it as =) [14:18] xnox: it says mdadm: imsm capabilities not found for controller: /sys/devices/.......... [14:18] raid1ubuntu: in that case..... i'm not so sure what you are setting up or how. [14:18] raid1ubuntu: cause you are not using fakeraid. [14:18] xnox: lspci says: Intel Corporation 631xESB/632xESB SATA RAID CONTROLLER [14:19] raid1ubuntu: typically, you'd boot into firmware and setup raid arrays as you wish. Have you done that? [14:19] xnox: yes i have created one big array with two disks [14:19] raid1ubuntu: then during installation, you should _not_ setup any raid. [14:20] raid1ubuntu: if it asks you to activate things say yes. [14:20] in the curses-based setup it tells me that it has detected sata-raid hardware, i select with yes and in manual partioning its there with the raid1 array [14:20] raid1ubuntu: during install select "use full disk, use lvm" (default option) and select only one drive -> that raid1 array. [14:20] raid1ubuntu: yeah so just install onto that. [14:21] xnox: if i select this, its only showing up the usb-stick i am installign from [14:21] raid1ubuntu: in manual partitioning do not enter "configure software raid menus" and don't change anything there. [14:21] xnox: yes, i know, i just select this [14:21] raid1ubuntu: can you try Trusty Tahr 14.04 installer? i think it may have a fix for that. [14:22] raid1ubuntu: http://cdimage.ubuntu.com/ubuntu-server/daily/current/ [14:22] xnox: i am on 14.04, 12.04 freezed when i said do partioning now [14:23] xnox: this is a screenshot of 12.02 but it is the same on 14.04 [14:23] http://picpaste.com/Photo_2014-02-20_02.32.07-1-CS2Wqwz5.jpg [14:24] xnox: the setup detects the (at boot time controller-setup-) configured array by itself [14:24] hi hallyn, i'm having a problem with lxc: http://paste.ubuntu.com/6965608/ just seems to hang [14:24] FWIW i can run the command manually and it works fine [14:25] so it is something about the environment, i suppose [14:25] tych0: which release are you on? [14:25] hallyn: trusty [14:26] raid1ubuntu: the device, at grub-installer installation time would be called /dev/mapper/isw__ (where name typically is Volume0) [14:26] xnox: i am now finishing the installation without boot, try to boot into the system from install, and then try to setup grub on the raid again, if that fails, i will remove the controller-raid1 disable the controller an go with software raid as you mentioned, i think this can be set up in the setup too? [14:26] raid1ubuntu: not md* [14:27] tych0: I'm guessing it' related to bug 1166841 === Malediction_ is now known as Malediction [14:27] raid1ubuntu: yeah, but to do software raid, you will need to change in firmware -> remove raid arrays, mark drives as non-raid. [14:27] (or something like that) [14:28] xnox: yes, i know thats what i meant with "remove the controller-raid1" hehe === aslaen is now known as aslaen_ [14:28] ah, hm [14:28] my $LANG is en_GB on this box [14:28] xnox: /dev/mapper/isd_letters_name ... ok i will look [14:28] well, en_GB.UTF-8 [14:29] tych0: when the lxc-create is hung,c an you confirm that locale-gen is running? [14:29] hallyn: yes, although it'll take me a few minutes to get back there, but i can do that [14:29] tych0: well, just setting yoru LANG to en_GB.UTF-8 in the script may fix it [14:30] ok, will do [14:30] there is some sort of hornet's nest lurking in the locale stuff, but I don't know what's going to come of it === roo is now known as Guest69319 === aslaen_ is now known as aslaen [14:34] :-) === rmk` is now known as rmk === DaIRCKing is now known as GTAXL [14:39] raid1ubuntu: mdadm in 14.04 will include fakeraid support. dunno if that is available and working yet [14:39] (dmraid no longer used) [14:43] pmatulis: well i should write the status report about htat. [14:44] pmatulis: let me mail it out. [14:44] xnox: thanks, mail destination? [14:44] pmatulis: ubuntu-devel probably. === freeflying is now known as freeflying_away [14:52] my server got 32gb ram how big should the SWAP partition be ? [14:53] raid1ubuntu: default - 32gb if you want suspend/hybernate to have a chance at working [14:53] raid1ubuntu: i use 32GB swap with 32GB of RAM on my desktop =) [14:53] * xnox ponders if can class myself as server.... [14:55] xnox: for a always-up server with no hibernate required ? [14:55] raid1ubuntu: you still want swap.... e.g. when i'm running big databases they use swap as disk-cache thus speeding up IO [14:55] xnox: the machien is going to be hosted in a data-center of my university (yeah i really managed to get them host my server :D) [14:56] raid1ubuntu: things go blazingly fast if you can store 10GB of data in ram and move it about. [14:56] raid1ubuntu: i presume you have like terabyte drives that 32GB of swap is peanuts and a big performance gain. [14:59] xnox: ok i will use 64gb [15:00] xnox: the setup does not offer any raid-menus on partioning screen (i have disabled the controller and want to go for soft-raid now) any idea how to enable them? [15:00] xnox: there is no fdisk in F2 console [15:02] raid1ubuntu: follow step-by-steps i've pasted to you earlier. [15:03] raid1ubuntu: go into something else / manual, enter configure raid, configure your raid1, back-out back to "Guided partitioning", select default install and then you'll be offered to install onto configured software raid device. [15:03] https://help.ubuntu.com/12.04/serverguide/advanced-installation.html [15:05] xnox: thats what i am actually doing (thx for the link) but there was no free disk.. i managed to fix the problem by selecting the sdba, select delete, cancel (this would take hours) and go back to the setup main menu, and reenter partioning, no i am able to create free disk space .... and NOW there is a raid menu.. *phew* === alex88_ is now known as alex88 === Guest69319 is now known as roo [15:26] Anyone here using cloudflare? Does it work like a full reverse proxy, so that people will see a cloudflare ip instead of the ip of your server when they ping a domain that uses cloudflare? [15:26] hallyn: http://paste.ubuntu.com/6965928 -- did 14.04 introduce a breaking change on QCOW2 compatability with older versions of qemu? [15:28] utlemming: yeah I saw something in the git changelog tha tmight be relevant [15:28] (breakfasting, biab) [15:29] !trusty|utlemming, [15:29] utlemming,: Ubuntu 14.04 (Trusty Tahr) will be the 20th release of Ubuntu. See the announcement at http://www.markshuttleworth.com/archives/1295 for more info. support in #ubuntu+1 [15:29] yay, grub on /dev/sda and /dev/sdb [15:35] hallyn: ping when you get back, if you would [15:38] is there a way to skip swap space wiping at the end of the setup? [15:52] utlemming: so hm. it looks like it is defaulting to compat=1.1 which i think should be ok... [15:53] (meaning it's a bug) [15:53] utlemming: presumably you can try compat=1.0... [15:53] no that isn't valid [15:53] hallyn: yeah, I tried compat=0.10 and it failed [15:54] hallyn: per what I found, you need 0.10 for pre-1.1 versions [15:54] d'oh. i missed htat you were on 12.04 for the mount [15:54] hallyn: yeah, the problem is going form trusty to precise :/ [15:55] utlemming: but 0.10 works everywhere? [15:55] hallyn: it seems that it works except on precise [15:56] I see - in your pastebin, that's the same file? [15:57] hallyn: it is...test.img created on latest trusty build, and then copied to the latest precise build [15:57] utlemming: please open a bug. i'll try to reproduce it with upstream qemu git, hand-built as 1.7 and 1.2 trees [15:58] (so i can file against upstream) [15:58] hallyn: sure, doing so now [15:58] thx [15:59] utlemming: fwiw there are some other qcow2 related bugs in trusty's git whic i need to nail down [15:59] hallyn: ack, good to know [15:59] hallyn: so locale-gen is not running when lxc-create is hung [16:00] and setting LANGUAGE, LANG, and LC_ALL to en_US.UTF-8 didn't fix anything :-( [16:00] tych0: please file a bug; i'll lose track otehrwise [16:00] hallyn: ok, on ubuntu/lxc or just lxc? [16:00] ubuntu/lxc [16:01] will do [16:02] hallyn: bug #1282655 [16:02] Launchpad bug 1282655 in qemu "qcow2 images are not backwards compatiable with 12.04's qcow2" [Undecided,New] https://launchpad.net/bugs/1282655 [16:02] smoser: do you have the rights to subscribe the server team to slof bugs? (I don't) [16:03] utlemming: thanks. (with lxc 1.0 release coming today it's good timing :) [16:03] hallyn: lol === ivoks is now known as ivoks_ [16:27] xnox: thx for your help! [16:28] cu [16:28] raid1ubuntu: are you all installed now? [16:28] =) [16:28] \o/ [16:36] hi using parted for the first time. I have a new attached disk 4GB. If I use the following command: ` parted -s -a optimal /dev/sdb mkpart primary 645 4294` Does that mean that the root partition will be 4GB and be aligned 32K? === yofel_ is now known as yofel === chuck_ is now known as zul [17:05] cross posting on #ubuntu [17:08] joshu__: please don't [17:08] joshu__: ask in the correct channel please, don't post across the channels [17:08] ikonia won't do again [17:09] no problem, thanks [17:32] Hello, is Linux's pppd daemon, which is used for PPPoE, userspace or kernel space? === tyhicks` is now known as tyhicks === MalcontentMatt is now known as mjohnson15 [17:50] I'm trying in vain but I can't get the right parted command to create the following root partition: ` 1 32.8kB 4295MB 4295MB primary` [17:50] the disk is 4gb (4294) === sarnold_ is now known as sarnold [18:58] how to activate the change in nsswitch.conf? changed from 'hosts: files dns' to 'hosts: dns [!UNAVAIL=return] files' === zz_mwhudson is now known as mwhudson [19:22] axisys: no change needed, the file is parsed on login [19:32] RoyK: hmm.. ok so I didn't need to reload nscd for all this tacplus sessions.. thanks [19:55] hi hallyn, when i do an lxc-start, i'm getting: lxc-start: command get_state failed to receive response [19:55] any ideas what's up? [19:56] tych0: again - is this on trusty? lxc from ppa, or from trusty archive? [19:57] trusty archive, sorry [19:57] tych0: how exactly did you create the container? [19:57] tych0: can you do 'lxc-start -n -l trace -o outout' and then pastebinit outout? [19:58] sudo lxc-create sudo lxc-create -n juju-bootstrap -t ubuntu-cloud -- -r precise [19:58] although when i install cgroup-lite it seems to work [19:58] hallyn: that's enough of a workaround for me; do you want me to trace it? [19:59] srgraber: ^ [19:59] tych0: lxc is supposed to require cgroup-lite [19:59] ah [19:59] :-) [19:59] so until cgmanager is in main, yes you must have cgroup-lite installed [19:59] ok, cool [20:00] that answers my question, thanks! [20:00] hi all [20:01] tych0 - np - ttyl === mwhudson is now known as zz_mwhudson [20:21] tych0: ok back to your locale issue :) can you reproduce the issue by just putting the lxc-create command inside a 2-line #!/bin/sh script? [20:22] tych0: and if so, can you do "env" right above the lxc-create, and pastebin the result? [20:22] hallyn: no i did try that [20:22] but i can do an env right before i run it in the script i'm doing now [20:22] and it fails that way? [20:29] hallyn: doesn't fail in a plain shell script [20:29] just in our crazy cloud installer script [20:29] so i'll have that run env and post that on the bug [20:31] tych0: actually i think locale isn't to blame here. i can reproduce your warnings, without a hang [20:31] i suspect actually that what you had [20:31] was a debconf menu waiting for input [20:31] yep [20:31] that's what i was thinking too [20:31] hung in timzeone [20:32] but why... you don't get the debconf when you do it by hand [20:32] stgraber: is there a way to tell apt-get that it should be non-interactive? [20:32] (regarding bug 1282657) [20:32] Launchpad bug 1282657 in lxc "lxc-create hangs in some cases on trusty" [Undecided,Incomplete] https://launchpad.net/bugs/1282657 [20:33] it is already doing dpkg -f noninteractive [20:33] bug in dpkg? [20:34] that would be really unfortunate :-( [20:34] i had some theory about this a while ago [20:34] i forget what it was === zz_mwhudson is now known as mwhudson === Den_Beiren is now known as DenBeiren [21:18] Hi, anyone willing to take a look at my configuration? I'm trying to set up a basic mail server but I think there is something wrong with my dns configuration. (I know asking to asking is bad but understanding me may take some time.) === tyhicks` is now known as tyhicks === hachre is now known as hachre_ === chmurifr- is now known as chmurifree === beisner- is now known as beisner [21:55] hm. qcow2 snapshots work fine locally on my trusty laptop... === beisner- is now known as beisner === genii_ is now known as genii === ciastek_ is now known as ciastek === stgraber_ is now known as stgraber === yofel_ is now known as yofel [22:22] hi [22:42] time for some git bisect goodness [22:43] i should come up with a cool script like arges does [22:43] except every compile requires manual intervention, so it probably wouldn't work too well [22:44] arges: actually... have you seen this bug? if you start kvm with a qcow backin store, it works fine on bare metal, but in nested grub fails to load the kernel [22:45] hallyn: nope. haven't seen that one yet : ) [22:45] (or, it could be weird environments involved - my nested case is trusty in precise's qemu on trusty kernel) [22:46] (but i'm discounting that since jdstrand has seen similar problems) === lazyPower_ is now known as lazyPower [22:51] Hi, how should roundcube be configured in relation to postfixadmin and the maildb? I'm following this guide: http://flurdy.com/docs/postfix/ [22:51] I'd imagine one login should get into all of those services. [22:56] first, what does any of what you said have to do at all with same random link you found in the internet? expecially that link [22:56] as roundcube is an imap client, it should be configured for your imap server, and has nothing to do with postfixadmin [22:57] I'm asking about making one login for all of them. [22:58] for all of what? [22:58] so far I only see postfixadmin and maildb [23:00] postfix, postfixadmin, roundcube, courier [23:00] in that case, you just have to make courier work [23:00] and make it work the way postfixadmin expects [23:01] though, I dunno why you want to make life so hard on yourself [23:04] Patrickdk, I think my courier configuration is correct. [23:06] then there you go then [23:06] everything uses the same password, time to move on [23:07] 123456 [23:07] well, apparently it's not the first level qemu. maybe it's actually the 1.0 qemu on the trusty kernel messing something up [23:08] I'm trying to add an mx record with noip.com. I see no reason why I can't use mail.mydomain.tld how do I make mail.mydomain.tld resolve to the same my same server? [23:08] Bitwise: do not use CNAMEs for MX records. [23:08] I'm not. [23:09] good good. [23:09] hmm, cnames are invalid :) === hazmat` is now known as hazmat === mikal_ is now known as mikal [23:10]