[00:38] [root@perseus ~]# showmount -e mynfsserver [00:38] mount clntudp_create: RPC: Port mapper failure - RPC: Unable to receive [00:39] i'm with a problem on ubuntu server 10.04 [00:39] in some clients the nfs server just works but in a specific client it doesn't. [00:40] can anyone help me? [00:50] problem solved. thanks.. [01:02] is there something lspci, but for getting info about what HDs are installed in my system? === jjohansen is now known as jj-afk [01:04] osmosis: lshw, lshal [01:04] osmosis: parted [01:04] astm, cool === ivoks is now known as ivoks-afk [01:33] I need to copy all folder except one, how may I do this? cp -R /folder .... [01:49] what is a good network usage monitor for Ubuntu? I would love if there was a good one that could keep daily/weekly/monthly usage stats as well as a live load count - even better if auto generates the report in HTML or something [01:52] demonspork: cactii mabye [01:54] !info cacti | demonspork [01:54] demonspork: cacti (source: cacti): Frontend to rrdtool for monitoring systems and services. In component universe, is extra. Version 0.8.7e-2 (lucid), package size 2033 kB, installed size 5488 kB [01:54] Andre_Gondim: perhaps one of the experts here may answer, but i'd suggest rsync with an --exclude switch [01:55] kinygos, thanks [01:56] Andre_Gondim: anytime [01:59] the Full Circle interview with Leann Ogasawara has been shared on facebook...in case she didn't know she's now famous :) [02:13] would the following command delete the partition table on my disk /dev/sdb... [02:13] dd if=/dev/zero of=/dev/sdb bs=512 count=1 [02:18] kinygos: that will blow away the boot sector, which contains the partition table [02:19] pmatulis: awesome, thanks :) [02:20] for some reason, running that command not only cleared the partition table on disk /dev/sdb, it also cleared a warning about an invalid flag 0x0000 against /dev/sda [02:22] but i've also lost the disk identifier... [02:25] kinygos, do you know the syntax? to copy all folder except one? [02:27] Andre_Gondim: i've been using the following (again, i'm a noob so the experts may jump in)... [02:28] Andre_Gondim: rsync -auHxv --exclude=/not_copied --exclude=/dest /* /dest [02:29] Andre_Gondim: check the switches though, for example, -H is to preserve hard links, you may not want that [02:29] where are the source? [02:30] in my command there, the source is /* [02:30] hmm ok [02:30] and the destination was /dest [02:31] You should always specify the trailing slash of directories in rsync calls [02:31] Andre_Gondim: sorry...the command should be: [02:31] rsync -auHxv --exclude=/not_copied/* --exclude=/dest/* /source/* /dest/ [02:32] twb: absolutely right...my mistake, my mistake [02:33] let me try [02:33] Andre_Gondim: and in that command, the source directory being copied is /source/ which cannot contain /not_copied/ so i will now go to sleep [02:33] kinygos, many thanks [02:34] Andre_Gondim: anytime, apologies for any confusion [02:34] no problem ;) [03:19] I am new to libvirt and struggling to get the network to work like I think it should. Is this a good place to ask about it? [03:40] Cacti. I have been over the manual for over an hour and I haven't even successfully gotten it to graph or even begin monitoring my network usage. Anyone have any pointers? my current break point is that I have no idea what I am doing with snmp, all of the guides just say "enter your snmp community" WHAT DOES THAT EVEN MEAN [03:42] a community is like a password. When you go into cacti, you will want to add devices, yes? So when you add the device, you should be able to turn on snmp, then choose the version and set your community and graphs [03:42] you can test to see if snmp is woking with snmpwalk from another box [03:43] n3kl: yes, or #virt (#libvirt?) [03:43] or locally, but iirc you need to change your /etc/default/snmpd to listen on another interface [03:43] twb: thanks [05:05] what package would i install to get ciphers for sshd working, i searched apt for blowfish and arcfour but it lists a lot of packages, not sure what i should install [05:06] i'm working off of a openvz minimal ubuntu server template, so i guess some of that stuff was left out [05:08] nm, got it working [05:11] DaveWM: sshd ciphers are compiled in [05:12] DaveWM: you can choose which are enabled (and their order of preference) by customizing the Ciphers value in sshd_config or ssh_config. [05:13] Note: the Cipher option (singular) only applies to legacy SSHv1 connections, and you really shouldn't use SSHv1. [05:17] twb : thanks , i'm trying to get x11forwarding to perform a little better over mobile connection [05:17] Anything telling you to set "Cipher blowfish" is obsolete; ignore it. [05:18] Oh, X11. You should look into NX. [05:18] yea i read plenty stating it as being faster [05:25] twb : what all would i have to do to set that up? i just did apt-get install nxproxy [05:26] I don't use NX, so I cna't help [05:26] ah, ok, well thanks for the tip anyway, looks like a good idea [05:26] crap. I can't check snmp from another machine, my other machine is down [05:56] twb : using nx running xterm remotely, its oddly faster than a normal ssh session to that server ?? heh [05:57] still not really great, i was hoping to run quassel irc client remotely, but i think its a bit too resource heavy for that sort of thing on this connection, i dunno [05:58] If you aren't encrypting NX you probably should be. [05:58] I prefer the irssi + screen method [05:58] I dunno why you wouldn't just learn irssi and run it in screen on a server [05:58] i like GUI, heh [05:58] GUIs are for pussies [05:58] i am what i eat === zz_sailerboy is now known as sailerboy [06:00] bare in mind i'm like 70% windows and only 30% linux... and that 30% has really mostly spawned recently [06:03] hi [06:22] so I get graphs that generate and update properly in cacti for localhost users, processes, load and memory [06:23] but the graph for my snmp traffic monitoring isn't generating [06:23] <_Techie_> has anyone had experience with ioncube loader? [06:23] I think it is because I have no idea what to put for Index Value in the Custom Data section of the data source [06:39] hi [06:40] can anyone let me know if I can upgrade an EC2 ubuntu server from 9.10 to 10.04 LTS with a plain dist-upgrade? [06:41] as in the same way as upgrading a local machine? [06:46] ok... how about just upgrading the kernel via a apt-get update && apt-get upgrade? is it safe? will the server automatically pick up the newest kernel on reboot? [06:59] nevermind... [07:16] Hi, I have a crontab in /etc/cron.d, but it doesn't appear to be running. Any ideas what might keep it from running? I have a simple test which does not seem to be working: * * * * * root /bin/echo "TEST" &> /var/log/test.log [07:17] oh nm, I keep forgetting crontab files can't have "." in the name [07:25] <_Techie_> has anyone had experience with ioncube loader? [07:28] twb : apparently i could have saved myself some trouble, i didnt realize that quassel has a client/core(server) component to it, so no need to forward the GUI, just gotta run the client and connect to the core remotely [07:28] Whatever man [07:29] kirkland, if I wanted to setup a UEC in VM's, should I use briged networking or with NAT will be enough? === Barre_ is now known as Barre [08:24] <_ruben> guess i should give up on trying to have dots in my non-fqdn hostname and go for a naming scheme based on dashes or whatever instead [08:30] hi [08:30] I'm planning to install Ubuntu Server in a VM. Does it matter whether I select 32-bit or 64-bit? Are there some packages only made for one but not the other? [08:33] is...uh...anyone around? [08:37] yes [08:37] I already got my answer in #ubuntu: the 64-bit version would not have run at all in Vmware. Thanks anyway though [08:38] <_ruben> 64bits works fine in vmware, *if* your host supports it [08:38] <_ruben> then again, vmware's a company, not a product [08:39] Yeah, sorry, I meant Vmware Workstation , which is a 32-bit program. [08:40] <_ruben> UweBollsFunder: doesnt matter either, it depends on if your cpu supports 64bits virtualization extensions (amd-v or intel-vt) [08:42] btw, give a medal to the guy who designed that "Get Ubuntu Server Edition download" page. 3 steps, collapsed by default, with a "Show me how" button that provides screenshots. Page is very minimal but extremely informative. [08:42] ruben: oh, I see. Yeah, mine doesn't, it's a lower-end C2D. [08:44] <_ruben> ah [08:44] I can always buy a fancier one if I ever need it [08:45] we'll see how this goes === sailerboy is now known as zz_sailerboy [08:46] how stable is mdadm between upgrades? Is it the sort of software I should never update because it's risker than sticking with the version that created the RAID? [08:49] <_ruben> never had any problems myself with mdadm upgrades, ymmv though :) [08:57] allright [08:57] thanks, and gn [09:17] good mornig everybody [09:17] somebody know a way to intall a deb package with dpkg that accept console clauses [09:17] i need install a software remotely and this is a problem [09:18] my doubt is how install an aplication that force to you to acept some clause in the middle of the instalation [09:18] and i want install remotely and can not accept this clausules [09:18] and i do with a Script, so no option like ssh dpkg or something like that [09:20] <_ruben> those things tend to be controlled through debconf, so with debconf-setselections you can pre-set the proper value, might need to use debconf-getselections or the logfiles to find out the proper debconf entry [09:24] _ruben where you from? [09:25] <_ruben> .nl [09:30] hi... what do i need to create a dyndns like service with bind9? (not update clinet for external service) [09:33] <_ruben> tom_w: what is it exactly what you are looking for ? [09:34] ruben what is the name of the package i can not installed [09:34] several routers on dsl should be able to change thir subdomain entries on a local DNS server [09:39] found a howto... http://idefix.net/~koos/dyndnshowto/dyndnshowto.html [09:39] ty [09:41] _ruben already have in my system debconf, that is the aplication? only have one page in the man and dont explain basicaly nothing [09:50] <_ruben> Osmosis: debconf-set-selections is installed by defaul, for debconf-get-selections you need to install debconf-utils [09:50] already done thanks [10:13] My computer need more time every boot to detect LVM drives. On fresh install it booted just fine, then next day i had to change rootdelay to 90 in order to boot. Today even 90s is not enough :/ What could be cause of that? im using 2VG, each on its own md device, using RAID1 [10:13] railing disks? [10:13] failing* [10:18] is it strange that /boot is not on the primary partition flagged as the boot partition? [10:24] kinygos: no [10:24] the bootable flag is ignored by grub [10:28] twb: thanks :) [10:31] how can i check if disk is ok? smartctl? anything else? [10:31] overlord_tm: smartctl can tell the disk to run self-tests [10:32] smartctl on its own is unlikely to find anything [10:32] what should i do then? i have booted with sysrescuecd and it took 3 minutes to detect logical volumes [10:33] yesterday 90s was enough, two days ago it booted in seconds [10:34] The command is something like "smartctl /dev/sda -T short" [10:34] The manpage should be clear [10:35] I'd also ask hdparm if it's in DMA5* mode or so, and check dmesg and logfiles for warnings [10:35] twb, i have just executed long test :) [10:35] Also, maybe you plugged the (external) disk into a USB-1 bus, or you added another device to the machine which slows down the probe cycle [10:35] overlord_tm: do a short test first [10:36] ok [10:37] yeah, i did, but it didnt get any faster when i unpluged and rebooted [10:37] There's a small chance that a full power-off will help [10:43] twb, another thing i noticed was that copy speed was dropping. Yesterday it was bout 40MB/s, today i barely reached 12MB/s [10:44] I don't really care [10:59] when i try to boot ubuntu 10.04 it gives message mount of filesystem failed CONTROL-D will terminate this shell and retry [10:59] snd gives prompt [11:21] can anyone spare 2 mins to look at my rsync command and tell me if i'm doing something silly (i've rebuilt my server so many times) ? i'm trying to copy the contents of one disk into a set of raid partitions that i've mounted in /raid/. what worries me is i'll be doing this on a running system: # rsync -auHxv --exclude=/raid/* /* /raid/ [11:22] (i spent yesterday trying all sorts of ways to do file copies on a running system and failed miserably) [11:23] remount read only the filesystem you read from? [11:24] it's a remote server...and i don't have access to external media :( [11:24] using LVM? [11:24] i wasn't planning to [11:25] i never copied files from live system, you never know in which state it was copied [11:26] i either shutdown or use lvm snapshot [11:27] when i try to boot ubuntu 10.04 it gives message mount of filesystem failed CONTROL-D will terminate this shell and retry [11:28] sid, have you checked? http://ubuntuforums.org/showthread.php?t=1305434 === oCean_ is now known as oCean === _Techie_ is now known as _TechAway_ [12:03] is there a way to completely disable fb on boot? [13:08] SpamapS: When you get this, can you update bug #644587 to state reasons for removal, and the future direction please? [13:08] Launchpad bug 644587 in drizzle "Please remove drizzle from maverick (was: fails to build from source on maverick)" [High,Confirmed] https://launchpad.net/bugs/644587 === ivoks-afk is now known as ivoks [13:54] Hey guys, could someone tell me why this might happen? http://pastebin.com/uhrrDg8Z === zz_sailerboy is now known as sailerboy [14:03] why when i install an ide drive to use as a back up I get errors on startup about the drives automatically mounted and it stating that the mount points are busy and it doesn't finish starting up? [14:03] haven't had a chance to format and add the new drive to fstab yet because I can't start up with it installed === sailerboy is now known as zz_sailerboy [14:05] How do i get ubuntu server to automatically get an IPv6 address? [14:05] I've got an IPv6 router running on my network now, and all the other machines manage it fine... [14:06] should just show up as hda after its recognized in the BIOS correct? [14:26] Daviey: about that nmbd bug i havent been able to reproduce it can you? === ivoks is now known as ivoks-afk [14:45] New bug: #652545 in apache2 (main) "package apache2.2-common (not installed) failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/652545 [14:55] Daviey: bug 644587 updated [14:55] Launchpad bug 644587 in drizzle "Please remove drizzle from maverick (was: fails to build from source on maverick)" [High,Fix released] https://launchpad.net/bugs/644587 [14:55] SpamapS: Awesome... the release team wanted a little more clarification.. it's "Fix Released" now \o/ [14:56] yep [14:56] cjwatson: thanks for that btw. :) [14:56] ok [14:56] * SpamapS wonders if cjwatson ever sleeps. [14:56] so does my wife [14:57] Want more clarification? [15:00] ScottK: on cjwatson's wife's perception of cjwatson's sleep schedule, or the bug? [15:01] I wasn't sure, so I was seeking clarification. [15:03] SpamapS: It didn't need to go in the bug, but don't forget once drizzle is in Natty, you can ask for a backport to get it in earlier releases. [15:05] ScottK: I often forget about backports, good point. [15:05] That reminds me that I need to backport the latest moinmoin to lucid [15:06] ScottK: whats the word on backporting something in main, that has new requirements not in main in the targetted release? [15:06] SpamapS: Backports doesn't care about Main/Universe. [15:06] sweeeeet === doko_ is now known as doko [15:17] New bug: #653020 in apache2 (main) "apache2 crashed with SIGSEGV in T.312()" [Medium,New] https://launchpad.net/bugs/653020 [15:24] join #android [15:26] New bug: #653027 in clamav (main) "package clamav-freshclam 0.96.1 dfsg-0ubuntu0.10.04.2 failed to install/upgrade: problemas de dependencias - se deja sin configurar" [Undecided,New] https://launchpad.net/bugs/653027 [16:02] hi people. Having a problem creating a directory, although doing it under root priv. and have +w priv for the folder [16:03] I get "mkdir: cannot create directory `scripts': No such file or directory" [16:03] the dir /srv is in place [16:06] Daviey: just a Q: what about euca-add-user? [16:30] i was wondering because i had read on a ubuntu server forum that updating regularly is hazardous is that true? [16:32] huh? updates are provided because the fix a serious problem or security issue - it's hazardous to NOT update regularly [16:32] zoopster: ok ty because that sounded wrong but its what it said [16:33] JamesPage: o/ [16:33] mathiaz: o/ [16:34] JamesPage: have you worked around the couchdb issue for the iso testing system? [16:34] mathiaz: yes but not much success - unable to get vm's to boot of a cloud hosted instance. [16:34] JamesPage: how does it fail? [16:35] mathiaz: gPXE randomly can't get files from the server - I can screenprint some of the errors if that would be helpful [16:36] JamesPage: yeah - if you can easily reproduce them [16:36] JamesPage: I'd have a look at the error messages [16:36] mathiaz: I'll spin one up now :-) [16:38] mathiaz: current hudson cloud instance is setup to configure the couchdb instance on a per slave basis so easy to change :-) [16:38] JamesPage: hm - what do you mean? [16:38] JamesPage: is there a couchdb running on each slave now? [16:39] mathiaz: pinged you and email a few minutes ago - take a look [16:39] JamesPage: ah - minutes... [16:39] * mathiaz gets new email [16:39] mathiaz: basically each slave can be configured to point at different couchdb instance so can be localised if need be [16:40] JamesPage: nice [16:40] JamesPage: I think we should move couchdb to each slave for now [16:40] mathiaz: also takes quite a long time to upload initrd and kernel on my adsl to the cloud! [16:40] * mathiaz nods [16:40] JamesPage: eventually we'll look at getting rid of couchdb [16:41] JamesPage: but for now everything is glued together [16:41] JamesPage: and it seems that running couchdb on the slave is the easiest workaround for now [16:41] mathias: agreed and it works; current successful tests where run in this way. [16:41] *mathiaz: [16:42] mathiaz: just tricky on desktop installs as conflicts with desktop-couch [16:43] JamesPage: I've just read your email [16:43] JamesPage: so it seems that local couchdb is the way to go for now [16:44] JamesPage: IMO it makes much more sense [16:44] JamesPage: that way (ie local couchdb) [16:44] mathiaz: I agree - keeps the test execution completely local with control in the master. [16:44] JamesPage: I'll try to setup a second slave on monday [16:45] JamesPage: it may be interesting to have an option to set whether to upload artifacts or not [16:45] JamesPage: I plan to setup my hudson slave on a server running in a DC [16:45] JamesPage: so uploading artifacts shouldn't take so much time [16:45] JamesPage: however for a DSL connection it may be better [16:45] mathiaz: suspect it gets more that 400Kbps upload! [16:45] JamesPage: or may be skip some artificats [16:46] JamesPage: IIUC initrd and kernel are the biggest files [16:46] JamesPage: how about not uploading these to the hudson master [16:46] mathiaz: makes sense - they don't add huge value [16:46] JamesPage: all the other files are smaller - it shouldn't take that much time [16:46] JamesPage: and we'd still have some debugging information in case [16:47] mathiaz: OK - I will work on this on Monday before you try your slave setup [16:47] JamesPage: great - is the trunk up-to-date? [16:47] I'll push at the end of today (so next 10 minutes) [16:47] quick question people keep saying that samba is best for sharing if you have windows computers in your network but what if i only have ubuntu linux computers and no windows computer is there a better file sharing software out there? [16:48] JamesPage: great - could you update the TODO file with the discussion point above? [16:48] JamesPage: to summarize: [16:48] JamesPage: 1. update documentation to outline couchdb setup on slave instead of master [16:48] JamesPage: 2. don't upload kernel and initrd as artifacts [16:49] JamesPage: that's all? [16:49] mathiaz: yep - no problem - I think that covers it. [16:49] tacomaster: i'm not an expert here (they're all kinda busy with maverick) so my noobish answer is NFS [16:49] JamesPage: excellent - keep up the good work! [16:50] kinygos: lol its fine if there busy i will probally figure it out sooner or later and ty for the advice on NFS :P [16:50] tacomaster: here's a page that may shed some light for you https://help.ubuntu.com/community/SettingUpNFSHowTo [16:51] kinygos: lol funny i had just found the page on google as you linked it [16:52] tacomaster: lol..sorry, i was a bit slow :) [16:52] kinygos: lol no reason to be sorry ty for the consideration [16:54] mathiaz: TODO updated and branch now up-to-date in LP. [16:54] hi all [16:54] JamesPage: awesome! [16:55] mathiaz: :-( it took a really long time but my test PXE boot from ec2 is actually working now. [16:55] mathiaz: they where failing before..... [16:55] JamesPage: oh well [16:55] JamesPage: given that you've got a working setup with couchdb on the slave === [1]iclebyte is now known as iclebyte [16:55] JamesPage: it doesn't really matter now [16:56] JamesPage: couchdb running on the slave makes more sense anwyay [16:56] mathiaz: will you be using -server for your slave install? [16:56] JamesPage: I'm not running the desktop edition on the system I plan to convert to a hudson slave [16:56] JamesPage: (if that's what you meant?) [16:57] mathiaz: thats great - allows you to avoid the desktop-couch conflict === _sjm_ is now known as sjm === luis__lopez is now known as luis_lopez [17:30] does grub want to mount the / or /boot partition? (grub legacy) === ivoks-afk is now known as ivoks [17:39] ScottK: Regarding bug #651138 , zul said he would look at it today. [17:39] Launchpad bug 651138 in likewise-open "likewise-open fails to build from source in maverick" [High,Confirmed] https://launchpad.net/bugs/651138 [17:39] Daviey: Thanks. [17:56] does anyone know how to overcome the automated install error during an automated install of a Node Controller?: "The file needed for preconfiguration could not be retrieved from http://#2 192.168.123.10:8774/preseed/preseed.conf" [17:56] my cc, clc, walurs, & ss are all on the same node [17:56] or machine rather [18:02] kentb: thats a very odd url its showing [18:02] http://#2 192.168.123.10:8774/preseed/preseed.conf [18:02] kentb: maybe a type-o in the configuration? [18:04] yeah, this about the only related bug I could find on it, but, I don't have more than one CC on my network: https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/524147 [18:04] Launchpad bug 524147 in eucalyptus "UEC NC failed to fetch preseed.conf from CC using lucid-server-amd64-20100218" [Medium,Invalid] [18:09] is there an easy way to completely wipe a physical disk? i'm still trying to get raid on my live box, and when i ask my isp to reset the box, they only reset the first disk [18:10] live box? i meant remote server [18:10] * RoyK can't wait to get his hands dirty on the new 120TB boxes arriving next week [18:16] kinygos: lol.. still? [18:18] kinygos: dd? === harrisonk is now known as harrisonk_away [18:19] SpamapS: you have no idea :( i've found articles that make it sound like it should be easy, like http://nixcraft.com/getting-started-tutorials/432-setting-up-raid-1-mirroring-running-remote-linux-system.html, but i fail at the first step [18:19] kinygos: how many drives? [18:19] RoyK: 2 physical disks [18:19] kinygos: sd[ab]? [18:20] kinygos: you were doing so well last time. Where did it fail this time? [18:20] RoyK: yep [18:20] SpamapS: installing grub... [18:20] kinygos: nobody could help you in here? [18:21] SpamapS: everyone seemed busy, so i've been searching online... [18:21] kinygos: boot on an install cd, switch to console two after the drivers have been loaded, for dev i a b; do dd if=/dev/zero of=/dev/sd$dev bs=1M count=1k; done [18:21] kinygos: boot on an install cd, switch to console two after the drivers have been loaded, for dev in a b; do dd if=/dev/zero of=/dev/sd$dev bs=1M count=1k; done [18:21] even [18:21] RoyK: unfortunately, that's my challenge...it's a remote server, and i don't have access to any external media :( [18:22] New bug: #652815 in eucalyptus "eucalyptus-common 2.0+bzr1241-0ubuntu4 not configured after package upgrade" [Undecided,New] https://launchpad.net/bugs/652815 [18:22] RoyK: but that command is useful to blat my sdb at least :) [18:25] shall i post to ubuntuforums, or is that a form of cross-posting given that i've been asking for help in here? [18:27] kinygos: explain your setup, please [18:27] RoyK: he's got a lame hosting provider that will only image his box, no install media possible. [18:28] RoyK: so we were trying to walk him through creating a broken raid1, copying everything to it, then booting off it [18:28] kinygos: so how eactly did it fail to install grub? [18:28] exactly rather [18:28] ah - right - I remember from last night [18:28] been like 3 days now [18:28] RoyK: i'm just having it rebuilt now. as SpamapS says, i have a dedicated server with 9.04 on one disk. a second disk is installed, but not mounted or partitioned. [18:29] SpamapS: here's the last set of steps i took: 1. copied partition layout from sda to sdb with sfdisk -d /dev/sda | sfdisk --force /dev/sdb [18:29] kinygos: then run dd to erase that drive and try to create a broken mirror there [18:30] SpamapS: 2. changed partition types to raid auto with # for partition in 1 5 6 7 8; do sfdisk --change-id /dev/sdb $partition fd; done [18:31] New bug: #653152 in dovecot (main) "package mail-stack-delivery 1:1.2.12-1ubuntu7 failed to install/upgrade: subprocess new pre-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/653152 [18:32] RoyK: here's where i've had a lot of confusion...sda has a swap partition and an extended partition that contains 4 logical partitions... [18:33] RoyK: so originally i was trying to mirror the disk, i'm now trying to mirror each partition [18:34] kinygos: I don't think you can create a mirror of an existing non-mirrored partition or disk [18:34] kinygos: just get the data over, install grub, try to boot on the new broken mirror [18:34] RoyK: that's what i thought...so that's what i've been trying to do, but i failed to install grub... [18:35] SpamapS: 3. i created corresponding raid partitions with # for partition in 1 5 6 7 8; do mdadm --create /dev/md$partition --level=1 --raid-disks=2 missing /dev/sdb$partition; done [18:35] kinygos: try over, please, first dd zeros over the first gigabyte or so on the secondary drive, then make new partitions and so on there, copy the data, try to install grub [18:35] kinygos: sounds great [18:35] RoyK: ok, that's what i'm doing just now [18:35] yeah, sounds good [18:36] apologies all in this channel, i should've put this in a pastebin [18:36] it's just a line [18:36] no problem [18:36] kinygos: All of that sounds right. So that should have created empty partitions. the next step would have been to mkfs/mkswap on them [18:36] multiline pastes should go to pastebin [18:36] at least > 3 [18:36] Spamaps: yep..that's what i did, # mkswap /dev/md1 [18:36] # for partition in 5 6 7 8; do mkfs.ext3 /dev/md$partition; done [18:37] kinygos: I think you should use ext2 for /boot [18:38] RoyK: /boot on sda is ext3, i was going like-for-like [18:38] just use ext2 - well - just in case [18:38] ok [18:38] some grub versions are a bit, well, grubby [18:39] lol [18:40] SpamapS: so next i created a directory /raid, mounted /dev/md1p6 in there, then created /raid/var, /raid/home, and /raid/boot, and mounted the corresponding /md1p partition [18:40] partitions [18:40] SpamapS: then attempted to copy all data with rsync -auHxv --exclude=/raid/* /* /raid/ [18:41] kinygos: just use /, not /* [18:41] in case there are hidden files around [18:41] RoyK: excellent :) [18:42] and -u is not needed in this [18:42] SpamapS: sorry, i missed 2 steps before i did the copy... [18:42] also, with -x, you'll only copy the root fs [18:42] so you need to rsync each fs [18:43] and then you won't need to exclude /raid, since -x does that [18:43] oh, i wonder if that's why it hasn't worked [18:43] kinygos: but so far it looks good, just man rsync first :þ [18:44] before the copy, i did 2 other things, one of which i don't understand (...) [18:45] kinygos: what was that? [18:45] RoyK: 1. mdadm -Es >>/etc/mdadm/mdadm.conf (which i think puts the uuid of the raid disk into the conf file) [18:45] pastebin mdadm.conf [18:45] RoyK: 2. update-initramfs -u (which i have no idea what it does) [18:45] normally mdadm.conf won't be needed [18:46] that's normally read from the drives in recent kernels (after y2k or so?) [18:46] RoyK: my server is currently being rebuilt, so i won't have any files for a bit [18:47] kinygos: you probably won't need to update-initramfs since you copy the bunch over [18:47] New bug: #653154 in dovecot (main) "package mail-stack-delivery (not installed) failed to install/upgrade: subprocess new pre-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/653154 [18:47] update-initramfs updates the drivers database, the drivers loaded before the root fs is mounted [18:47] RoyK: at last, a straightforward description of what that does :) thanks!! [18:47] or - well - it creates a cpio archive of needed drivers [18:48] but still the same, really [18:48] anyway - your initrd file resides in /boot and will be copied if you copy that filesystem with rsync [18:48] so once i'd copied all the data over, at least i thought i had, but it seems i hadn't because of the -x switch, i tried to install grub on the second harddisk [18:49] kinygos: chrooting into /raid might help a lot [18:49] chroot /raid [18:49] then check if things look familiar [18:49] from there, install grub [18:49] onto /dev/sdb [18:51] ok..that may have been the issue...when i launched grub, i called root (hd1), then setup (hd1) which failed [18:51] i then rebooted and escaped into the grub menu there... [18:52] i entered find /boot/grub/stage1 and got file not found...which baffled me because i can still boot into the first disk [18:52] well, try again with chroot [18:53] RoyK: but what you've suggested sounds right, so i'll do that and report back...it may be a while as i need to wait for the re-imaging to be complete, but thank you very much for your time and advice...i really appreciate it === Barre_ is now known as Barre [18:54] np :) [18:54] kinygos: can you pastebin 'mount' now? [18:58] RoyK: what do you mean? the mount commands i was using? [18:58] RoyK: or the output from running mount? [18:58] just which filesystems are mounted where - the output of 'mount' [18:59] RoyK: my server is being rebuilt, apologies :( [18:59] brief afk..real life issue [19:06] oh cool. i didnt know this was a channel lol [19:07] Dravekx: they can get kinda whiny over there ;) [19:07] just ask your question though - no need to ask to ask [19:07] I have lamp setup on ubuntu server and want to jail users in their home directories so they can create web pages in a certain directory. how can I do this and can someone help me set this up? [19:08] not sure I said that right. [19:09] I want to allow each user to have a folder ( say public_html ) that, when they go to the web and point to foo.com/~user it will display the contents of /home/user/public_html/ [19:10] Ive had this setup before, but my server crashed and bought a new one. [19:10] http://heriman.wordpress.com/2008/08/05/enabling-apache-user-home-public_html-directory-in-ubuntu/ [19:11] I guess that's the easiest way ;) [19:11] Ekips, ahhh!! that looks right!! thanks. :) [19:12] np :) [19:17] any idea how I can jail users over ssh so they cant login and cd around the server? [19:19] dravekx, ya, use ssh [19:19] and do you actually mean jail? not chroot [19:19] cause, ssh logs in users, that is all it does [19:19] or did you mean scp/sftp? [19:20] patdk-wk, scp/sftp :) [19:20] sorry lol [19:20] well, sftp is easy, scp not so much [19:20] scp is just normal ssh really [19:20] it's easy to chroot users, using sshd.config [19:21] patdk-wk, honestly, when I create a user, I want them to have access and view to nothing but their home directories [19:21] http://www.howtoforge.com/chrooted_ssh_howto_debian [19:21] that is old [19:21] I wonder if ubuntu has a new enough ssh for this in lucid, I haven't checked [19:22] I found this one, but not sure if it's what I'm talking about: http://www.marthijnvandenheuvel.com/2010/03/10/how-to-create-a-chroot-ssh-user-in-ubuntu/ [19:22] actually, that url I posted wasn't the one I thought it was :) [19:23] http://blogs.techrepublic.com.com/opensource/?p=229 [19:23] that was more what I was looking for [19:23]