[00:32] Greetings. I'm trying to set up a bittorrent tracker in Ubuntu Server 10.10. I ran "aptitude search bittorrent" and found a package called "bittorrent" that says it has a tracker. I installed it, but I have no clue how to set it up. I tried "man bittorrent" but there is no documentation and a google search yeilds everything under the sun related to bittorrent. [00:35] Is there any documentation I could look at? [00:46] New bug: #705691 in cobbler (universe) "cobbler-web should include a working configuration and a README file detailing the steps necessary" [Wishlist,New] https://launchpad.net/bugs/705691 [00:57] is there any special reason for why innodb-segment of my.cnf has been removed from mysql in 10.10? [00:57] They left the comments, but not the options. === ogra is now known as Guest96093 [02:41] I'm having some trouble seeing past the initrd load when installing via a serial console - does anyone here have experience w/ that or know a better forum to ask in? [02:49] So I'm having an odd issue. My server is loosing time. It is Lucid 64-bit. It looses about 6 minutes a day. I found a couple of threads that recommended syncing NTP. Is there a way I can ensure that it will stay reasonable close to on-time even if it is offlien? [03:03] jMyles: try changing your CMOS battery [03:21] these are the pxelinux appended options I'm using: append initrd=ubuntu-installer/i386/initrd.gz console=ttyS0,9600 -- [03:21] I can see the kernel load, then I can see initrd load, but once the kernel starts (I presume) my terminal screen blanks and all activity stops [03:22] I've tried numerous permutations on the console= parameter, and I've tried several permutations with earlyprintk - all to no avail [03:22] I've tried ttyS0-3 - no joy [03:22] so - I'm puzzled [03:25] pmatulis: It's not my CMOS battery - the time is accurate in the BIOS, just not in ubuntu [03:26] you're not running in a vm right jMyles? [03:32] svdasein: right [03:33] jMyles: do you have any cpu throttling enabled? [03:36] can someone help me with an upgrade issue? I just did an apt-get dist-upgrade, it installed the new kernel. When I did a reboot, I get GRUBs menu, but no matter which option I choose the server will not boot [03:37] can someone help me with an upgrade issue? I just did an apt-get dist-upgrade, it installed the new kernel. When I did a reboot, I get GRUBs menu, but no matter which option I choose the server will not boot [03:38] I have reloaded this server 4 times now over the past 2 days and I would like to figure out how to resolve this [03:40] I am running 10.10 32 bit on an older compaq proliant 1600 series [03:40] Please can someone help with this or point me to the group who could [03:43] no one can help me? is everyone asleep? I need to get this server operational. [03:45] reinstal from iso? [03:48] pickett: from a CD downloaded last night [03:49] here is the complete issue: [03:49] I am loading a server that will be a print server, I loaded 10.10 32 bit (it is an older Compaq Proliant 1600 Series) I do an apt-get update and then an apt-get upgrade. I then reboot the server and everything works fine. I then do an apt-get dist-upgrade and my kernal goes from 2.6.35.22-generic-pae to 2.6.35.24-generic-pae, it re-writes grub and now no matter which selection I choose either 22 or 24, or either re [03:49] grub now says msdos instead of ext4 is that the issue? If so how would I change that [03:50] sounds like grub's the prob [03:50] is it on an ext4 partition? [03:51] it said it was when I loaded it [03:52] when I edit the commands in grub for the failure this is what I see: [03:52] next time you install try making a seperate 300mb ext3 partition for /boot [03:56] redordfail, insmod part_msdos, insmod ext2, set root = '(hd0,msdos1)', search --nofloppy --fs-uuid --set c(realy long #), linux /boot/vmlinuz-2.6.35-24-generic-pae... [03:56] can I change the grub somehow to allow it to boot? [04:06] signs point to yes [04:06] Well, it's possible. Whether YOU can is another question. [04:11] twb: thanks [04:12] Personally I just set up a static extlinux bootloader and kick grub in the pants, but that's not the party line [04:13] well I guess it is reload # 5 and more then likely again no luck in getting it to work. [04:16] hroberts: would make sure the entire disk you are using is wiped clean. I usually do this gparted. Also search google for wiping the MBR. Then I would do what pickett said and create a /boot/ parition on the device so /dev/sda1 [04:17] hroberts: are you doing any raid stuff? [04:20] do most people use apt-get dist-upgrade? I usually just us apt-get update && apt-get upgrade [04:20] *use [04:43] how to get down sit0 ? [04:44] thesheff17: sorry on the server [04:44] I am doing a hardware raid 5 on this server, I do the dist-upgrade to upgrade the kernel, is there a better way to upgrade the kernel? [04:45] I wipe the server everytime, I have tried this with both LVM and without it [04:48] hroberts: I'm pretty sure apt-get upgrade does the kernel as well...I could be wrong... [04:50] hroberts: it says this: dist-upgrade will remove obsolete packages from your system, whereas upgrade will not. This function is out of necessity for upgrading from one distribution release to another. [04:50] so apt-get upgrade should do the kernel. [04:52] don't think it does [04:52] aptitude upgrade does [04:58] apt-get upgrade should only upgrade packages which are already installed. so linux-image-virtual Depends: linux-image-2.6.32-21-virtual, which is a new package (not a new version of linux-image-virtual), so it won't be pulled in automatically [04:59] I just use upgrade, and if there's a new kernel, it'll let me know by telling me linux-virtual & linux-image-virtual have been 'held back' [05:02] shauno: yea I only use apt-get update && apt-get upgrade and then I guess I get the new kernel when the new version comes out of ubuntu and wipe the server with the distro cd. [05:05] I tend to fetch them when I know I can schedule a reboot. especially running LTS, having a 2-year-old kernel isn't always a good plan :) [05:06] shauno: so what command do you run to get the new kernel? [05:08] http://paste.ubuntu.com/556437/ [05:09] 5-6 tell me there's a new kernel package but it won't be installed by upgrade. [05:10] (18-19 explain why. linux-image-2.6.32-27-virtual is a 'new' package, not an upgrade to an existing package) === IdleOne is now known as gpc [05:11] if they made them new versions of existing packages, dpkg would overwrite / remove the existing package, so you wouldn't have a known-good kernel as a fallback [05:11] which has the potential to really ruin your day [05:12] thx for the info..yea I guess I never update the kernel then. [05:19] 16:18 !!! ":Gnea!~gnea@unaffiliated/gnea JOIN :#ubuntu-server" (error Invalid search bound (wrong side of point)) [05:19] ...WTF was that? [05:19] Hey everyone, I've got a weird problem - I've put ubuntu server 10.04 amd64 onto a 4gb usb stick with unetbootin and booted it - when it boots, it goes through the normal stuff until it tries to mount the device as a cdrom, but it fails [05:20] twb: probably a problem with your irc client [05:20] Somehow the JOIN event borked my channel buffer [05:20] Yeah, I just don't see what [05:20] It's not like you have a unicode combining digraph in your nick [05:20] but anyway [05:21] it's failing to mount the device, how can I fix this? [05:21] I don't support unetbootin, but you might like http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/current/images/netboot/ [05:21] network installation is not an option at this point [05:21] Then I dunno [05:22] Maybe you can build a working image with jigdo [05:22] well, I tried using usb-creator-gtk and it kept getting an i/o error [05:22] I don't approve of GUI-based ISO-to-USB converter things. [05:22] I used usb-creator-gtk last night w/ 10.04 server and it worked fine [05:22] Gnea: ooh.. what version of Ubuntu is your install CD? [05:22] Gnea: the ISO you passed to unetbootin, I mean [05:23] maybe 4GB is too big? [05:23] ... [05:23] 10.04 [05:23] Darn, too old. [05:23] I was using 10.04.1 [05:23] why would 4gb be too big? [05:23] 10.04 is hardly too old [05:23] With maverick onwards, you can convert an ordinary .iso into an image that works DIRECTLY as both a CD and a USB key [05:23] I see it's trying to use /dev/sr0 [05:23] Gnea: too old for syslinux 4 [05:24] twb: aaah [05:24] I'm surprised Ubuntu doesn't do it officially [05:24] well this is geared toward a production server, so maverick is out of the question. [05:24] Yeah, no worries [05:25] try putting the mini.iso on the usb stick [05:25] then net install [05:25] pickett: apparently he has no network [05:25] I wonder if it's due to a udev rule... [05:25] yeah, it's off the grid for now [05:25] it has to be done without the network for now, no choice [05:27] try using the maverick iso creator to create the usb [05:27] I don't run maverick on any system. [05:27] 10.04 only [05:27] is there a backport? [05:28] can't use a live mav cd to do it? [05:28] don't have a cd. [05:28] Which, I suspect, is the crux of the problem [05:28] it's part of it. [05:28] try it in vmware or vbox [05:28] it should *just work* with a usb stick [05:28] not an option. [05:29] I put backtrack on a 4gb usb stick and it works fine, no problems [05:29] same with elive [05:29] so why not ubuntu? [05:29] hrm, going to try something [05:30] eh? [05:31] * Gnea uses vim [05:31] i have installed several servers from 4 gig stick [05:32] gobbe: the udev rules appear to be forcing the actual cdrom drive to be used in lieu of the usb stick [05:32] !usb [05:32] For information about installing Ubuntu from USB flash drives, see https://help.ubuntu.com/community/Installation/FromUSBStick - For a persistent live USB install, see: https://wiki.ubuntu.com/LiveUsbPendrivePersistent [05:33] hmmh? [05:33] so you cannot install it from stick? [05:34] have you not read a single thing I've typed since I entered the channel? [05:34] well i don't have so long backlog, i'm sorry [05:34] ah [05:34] np [05:34] here... [05:34] Hey everyone, I've got a weird problem - I've put ubuntu server 10.04 amd64 onto a 4gb usb stick with unetbootin and booted it - when it boots, it goes through the normal stuff until it tries to mount the device as a cdrom, but it fails [05:34] ^^ original question [05:34] Gnea: Error: "^" is not a valid command. [05:35] uvirtbot: shut up [05:35] Gnea: Error: "shut" is not a valid command. [05:35] I'm hoping that's an official bot [05:35] ^help [05:35] Gnea: (help [] []) -- This command gives a useful description of what does. is only necessary if the command is in more than one plugin. [05:36] ^owner [05:36] Gnea: Error: "owner" is not a valid command. [05:36] gobbe: make sense now? [05:36] Gnea: yes, so you cannot install it because it trys to mount your stick as cdrom [05:37] what kind of stick you are using? [05:37] gobbe: it tries to mount the cdrom as cdrom, not the stick as cdrom [05:37] sandisk cruzer blade [05:38] tough little thing - kick it, drop it, roll over it with a car, it won't die [05:38] yea [05:38] yet it can't install ubuntu [05:38] :) [05:38] but it's tough ;-( [05:38] lol [05:38] ;-) i mean [05:39] heh [05:39] yeah [05:39] hmmh, i sounds like somekind of udev-thing tho. I have used few kingston models without problem [05:39] hmm [05:40] I do have a 2g cruzer, that could work [05:40] it's sandisk, but different model where the connector slides in/out [05:40] and it lights up [05:48] where does munin store its data? [05:48] okay, this is messed up [05:48] i renamed a node and all its previous data went away :( [05:48] /dev/sdb1 vfat 1.9G -8.0K 1.9G - /media/D3A3-0FFC [05:48] -8.0k?! [06:07] well, almost made some progress, but that's only actually made my faith in the installation process become less than what it used to be [06:08] apparently, if I disconnect the cdrom from the system entirely, the installation will ask me if I want to do it manually - if I do that, it will scan everything on the system EXCEPT for the /dev/sdb1 itself. It actually considers /dev/sdb, but not /dev/sdb1. [06:08] This is epic fail. [06:08] bad ubuntu, BAD! [06:08] * Gnea considers kicking to the curb [06:28] can you try getting grub to boot from the iso? [06:29] pickett: how would that help? [06:31] wouldn't need the usb then [07:53] Hi. I get "System information disabled due to load higher than 1" on some of my multi processor/core systems. [07:54] If anything, shouldn't that message be "System information disabled due to load higher than ${NUMCORES}" [07:54] As a load of 1 is nothing to my 8 core system. [07:58] psteyn: you get it... from what? [08:02] twb: as I log in, so probably from motd [08:02] log in via ssh [08:02] Ergh [08:02] Try grovelling through /etc/update-motd.d [08:04] found it: 50-landscape-sysinfo will submit a patch [08:12] I've seen that mentioned a few times in here, but bug #675245 (the only one I can find relevant) is untouched [08:12] Launchpad bug 675245 in update-motd "evaluation from the systemload on multiprocessor machines in 50-landscape-sysinfo" [Undecided,New] https://launchpad.net/bugs/675245 [08:24] shauno: ah ok, then im not going to bother :) [08:30] not even sure if update-motd is the right package to be filed against, it's not on my system (and that file doesn't appear to exist on packages.ubuntu.com) [08:33] does /etc/update-motd.d/50-landscape-sysinfo exist on your system? [08:34] nope [09:28] Ey guys. On my ubuntu 10.04.1 server, i have 5 users, 3 of which cannot ssh and 2 where it works (where root is one of em). I havent been able to figure out why, and ssh with the -vv argument doesnt give me anything that could indicate a problem. But i just noticed that in my /etc/passwd, all the users where ssh works have the same numbers in user and group id, whereas the users where it doesnt work have a different number in the user and gro [09:30] ChrisBuchholz: do you have anything showing up in /var/log/auth.log or /var/log/daemon.log ? [09:31] let me check [09:33] greppy: nope, nothing that seems out of order [09:44] greppy: have to go to the store, will be back [09:57] hey [09:57] can i get some help? [09:57] ask your question dont ask to ask [09:58] just checkin see if someone was here [10:00] i want to remote desktop connection using my htc evo 4g, specs :ubutu10.04 amd64, 2gb drr2 ram 2.1 amd anthon2x [10:02] hello? [10:18] New bug: #658864 in samba (main) "rhythmbox crashes with Game/Audio CD in drive (dup-of: 529714)" [Medium,New] https://launchpad.net/bugs/658864 [10:18] New bug: #659334 in samba (main) "rhythmbox crashed with SIGSEGV in _nss_wins_gethostbyname_r() (dup-of: 529714)" [Medium,New] https://launchpad.net/bugs/659334 === ogra is now known as ogra_ === ogra_ is now known as ogra [10:22] so anyone else seeing "System information disabled due to load higher than 1" on login despite the load being nothing? [10:25] jmkgreen, I don't know why it does that either [10:25] perhaps the motd takes a while to refresh, and when it was reloaded last the cpu load was higher than one? [10:26] nah this has been ongoing for months on the same boxes [10:26] blooming annoying since it no longer shows any available security updates [10:27] blooming? [10:27] flipping [10:32] jmkgreen, I think I found it [10:32] /etc/update-motd.d/50-landscape-sysinfo [10:32] that's the script that updates the sysinfo [10:36] however, that doesn't explain why your system has a high load. [10:36] my system does not have a high load [10:36] hence the problem [10:36] since it's getting its info from the kernel and it says its greater than one. [10:36] 10:36:43 up 8 days, 23:27, 2 users, load average: 0.05, 0.12, 0.16 [10:36] not high :-) [10:37] yeah right now [10:37] does it happen every time you log in? [10:37] those scripts in update-motd.d are run by pam [10:38] yes, and after login 'w' still shows no more than 0.4 (as a result of the login) [10:38] it's as if there's a stale file not being overwritten [10:39] if I run ./50-landscape-sysinfo it displays things fine [10:39] I am now running watch on it and I'll login again in another terminal. right now it shows system load of 0.39 [10:40] 0.48 it rose to and now it's coming back down [10:40] Daviey: Do you have any reservations my uploading a fresh Nova to Natty? [10:42] wtf [10:42] why is "Cornish (kw)" an english language choice in launchpad? [10:42] someone taking the piss? [10:44] lol [10:44] jgreenso6bkpdSB1Dcav5ztaRORa [10:44] oops ww [10:44] that your password? [10:47] that was my launchpad password (now reset) [10:48] :P [10:48] keypass auto-type. first time I'd used it [10:48] going to ask a question about this load display issue [10:48] you should use ssh keys for your box [10:49] well [10:49] do you want it to display or do you want to know why it's having this issue? [10:49] because you could probably change the value in that script to accept a higher system load [10:53] IrishWristwatch: I just adjusted it to read if [ $(echo "`cut -f1 -d ' ' /proc/loadavg` < 2.0" | bc) -eq 1 ]; then - but it still happens [10:53] check your /etc/motd [10:54] * soren sighs at that check. [10:54] wait [10:54] soren: something I did? [10:54] nvm >_> [10:54] I forgot [10:54] jmkgreen: No, something whoever wrote that did. [10:54] soren: not brilliantly written then? [10:54] can't say I'm a shell scripting guy [10:55] * soren isn't impressed, no [10:55] heh [10:55] fair enough [10:55] if grep -q ^0 /proc/loadavg; then foo; else bar; fi [10:55] yeah this whole motd thing seems very fickle [10:55] How hard is that? [10:55] and the ubuntu wiki isn't helping much at all [10:56] IrishWristwatch: indeed [10:56] soren: there was a bug somewhere to make it check if the load was less than the number of cpu cores found [10:56] someone pointing out that on an 8-core box a load of 2 wasn't necessarily a problem :-) [10:57] alright goodnight [11:01] jmkgreen: That's simple to adjust for. I'm still saving 4 clone(2) calls. [11:09] jmkgreen: maxload=1; currentload=$(cut -f1 -d. < /proc/loadavg); if [ $currentload -lt $maxload]; then win; else lose; fi [11:11] while that might be desirable, that doesn't help my issue [11:12] still says the load is higher than 1, which it isn't [11:14] Can you just make it output what's in /proc/loadavg? [11:16] I've modified the text that read 'System information disabled due toload higher than 1' to read 'System information disabled due toload higher than foo' [11:16] the original text was still sent to displat [11:16] so that script is just not being executed [11:16] Yeah, I forget when update-motd gets run. [11:17] Allegedly, "Executable scripts in /etc/update-motd.d/* are executed by pam_motd(8) as the root user at each login" [11:18] er [11:18] jamesg@blofeld:/etc/update-motd.d$ ./20-cpu-checker [11:18] exec: 3: /usr/lib/update-notifier/update-motd-cpu-checker: not found [11:18] that probably causes the rest to fail and for nothing to be written to the motd file [11:19] That's not unlikely. [11:20] ah ha! [11:20] sudo apt-get install update-notifier-common is the magic to fix all things [11:22] jmkgreen: Cool. Filed bug 705828. [11:22] Launchpad bug 705828 in update-notifier "update-motd hook should check whether update-motd-cpu-checker exists" [Undecided,New] https://launchpad.net/bugs/705828 [11:24] soren: ta, I'll leave you lot alone now. Thanks for the help. [12:10] Hi there. Am working at a ISP that has IPv6 enabled everywhere. We use Router Advertisements to bring out information about two routers (gateways) to the client. [12:11] In linux world, RA makes the machine (client) create it's own address based on the mac number. But what if I want to create something myself, but still listen auto set gateways? [12:12] Let's say I want 2a02:470:0:1::19 instead of 2a02:470:0:1:5652:ff:fe17:dc7b. How can I make this happen? [12:12] Announced net is 2a02:470:0:1::/64 [12:12] heh [12:12] Anyone here using IPv6 to start with? [12:12] that would be the difference between ra and dhcp [12:13] patdk-lap, yeah, but RA makes the client choose it's own address. I guess I can control this some way= [12:14] Control this on the client of course. Which is a ubuntu system [12:16] <_ruben> there's probably some sysctl to disable slaac, or just don't advertize the prefix on your routers [12:17] _ruben, The advertisement contains important information about gateways in the network. So I still need to listen to RA. "slaac" you say. A keyword! Let's try that in google :) [12:20] <_ruben> Blinkiz: i didnt say "disable the advertisements", i said "disable the *prefix* advertisements", which would result in a default gateway being configured on the clients, but not a autoconfig (slaac) address [12:20] _ruben, Hmm. interesting [12:21] _ruben, anyway, I can not disable the prefix announcement. :) [12:22] Doing a minimal install of 11.04 still seems to yield a high amount of installed packages, is the initative to create a "true" minimal install still happening? [12:22] I will look more into how slaac behaves and how I can control it on the client. But I need to handle a support ticket at the moment. Need to get back to this in a couple of minutes or so [12:26] <_ruben> might wanna look into /proc/sys/net/ipv6/conf/all/accept_ra_defrtr and /proc/sys/net/ipv6/conf/all/accept_ra [12:27] <_ruben> not sure if enabling the first works when disabling the second [12:28] anyone know what is needed to setup a DNS Server? like how I register it or what I need to do to enable people to acutaly find the DNS to resolve the Domain? [12:30] shaggy2: you want to setup bind9 and use the server as your own personal dns server? [12:31] https://help.ubuntu.com/10.04/serverguide/C/dns.html [12:31] well sort of, I aint sure what u mean by personal dns server, basicly I am setting up a hosting company, and I want to be able to use the dns server installed with my hosting package (ispCP [12:31] and I am using 10.10 [12:31] https://help.ubuntu.com/10.10/serverguide/C/dns-installation.html [12:32] <_ruben> shaggy2: when you buy/register a domainname, you have to provided the registrar with a list of dns servers, those dns servers will be running bind/unbound/pdns/whatever, and ppl will be directed to them automagically through the glue records [12:32] ahhh [12:32] thank you, that is what I wasn't sure on [12:32] * _ruben takes note not to buy hosting services from shaggy2 [12:32] hey I am only setting up [12:33] <_ruben> "learn first, do later" tends to work best, tho it's not commonly acted out [12:33] basicly before I SELL hosting I will have set this up about 20 times atleast and tested with mine and my mates [12:33] I am learning [12:33] the way I learn is by doing [12:33] it's not actualy public yet [12:33] <_ruben> ok :) [12:33] I have my Domain and my mates [12:34] best of luck to you shaggy2! _ruben, wouldn't setting up a primary and secondary master be enough for what he's looking to do? [12:35] but no actual public connections coming in, I never even thought of using my own DNS but then I thought it would be easier if customers would be able to use the hosting panel to setup the dns records on there own domain [12:35] brb [12:37] <_ruben> what's the domain? [12:38] shaggyweb.net and cjg-designs.net shaggyweb aint working I was getting DOS attacks so my ISP decided to shove me back on a dynamic IP while they take my static and find the fella, got work back that he is located in NSW Australia [12:39] he was routing it through about 5 sep proxies, and I don't know why he was trying to go at me, bloody little shit [12:39] he was able to use 5 proxies and DoS you from one source? IPTables couldn't kill it? [12:40] ok back [12:40] well that was silly [12:40] lol [12:40] <_ruben> depending on the type of dos, iptables wont be of use [12:41] but ok so my understanding is, I change the dns server to my own with the register, and then my DNS Server would work [12:41] <_ruben> if it's setup properly, yes [12:41] ok thanks ruben [12:41] will test it once I get my static back [12:42] I know it's stupid to do it, but can the primary and secondary be hosted on the same server? [12:42] atm I only have 1 server and 1 IP [12:42] <_ruben> then again, doing *any* "serious" hosting on a consumer internet connection is asking for trouble [12:43] I will have fiber this time next year, and I am planning on using that for the REAL hosting, just wanted to get it setup and learn what I need to now [12:43] <_ruben> having only 1 dns server is far from being recommended either, then again, if the dns server is only used for domains that are hosted on the same server, then it doesnt matter (if dns is down, the site is as well, so dns doesnt matter anymore) [12:44] <_ruben> and there's several free dns services out there, including ones that allow you to run your own master and have them slave for you [12:44] I currently have bind installed [12:51] Hi again. In my quest of knowledge about how to decided my IPv6 address, I have found a great topic to start my search from. All the people that do not like that the MAC address is the second part of a IPv6 address in linux. If I can find out how to control this, I will find my solution [12:52] So, the net is announced with RA messages and each client creates is own address based on the MAC number. How can I control this? For example, in the Windows world, they have some kind of generator to set a unique MAC number without using the MAC number. [12:52] A few keyword within the topic would be great. [12:54] <_ruben> Blinkiz: /proc/sys/net/ipv6/conf/all/use_tempaddr .. temporary addresses and privacy extensions are the keywords here [12:54] _ruben, nice :) [12:55] <_ruben> Blinkiz: another hint: there's also #ipv6 ;) [12:56] Ohh ;) [12:56] what can I say.. "ubuntu geek" [12:56] Need to include... eehh.. "linux" in my world.. hehe === oubiwann is now known as oubiwann_ === oubiwann is now known as oubiwann_ [13:54] hi [13:54] hello === doko_ is now known as doko [13:57] how can i check what is blocking my os and making my proc stay in average 98% wait-io state? [13:58] a harddrive :) [13:58] iostat [13:58] top? [13:58] hi patdk-wk [13:59] mianosm: top along with htop show lot of process with state D, but these processes work most of the time normally, this wait-io spike happens once every few minutes and stays up to 2 minutes, so it's hard for me to find a cause [14:00] iotop show no more activity during wait-io spikes than in normal work state [14:00] most likely a drive had bad sectors [14:00] dmesg? [14:00] run smart selftests on your drives? [14:02] patdk-wk: dmesg shows something like this: http://pastebin.com/5NzmMw74 [14:02] Morning all!! [14:03] nice, something is crashing :) [14:03] patdk-wk: is there a way to tell what exactly, or at least is it software or hardware? [14:03] pickup, you run postfix? ;) [14:04] Here might be some interesting info for you as well: http://nixcraft.com/linux-software/431-what-i-o-wait-under-ps-command.html [14:04] binBASH: yup [14:04] spajderix: so that one is crashing [14:05] spajderix: maybe some big mail in queue? [14:05] ps -wauxf [14:05] mianosm: thanks for link [14:06] Seemingly the databases usually cause the issue, if you're pulling addresses from a db to process through postfix maybe? [14:06] Those D statuses suck, but ps -wauxf should be a good visual for the issue possibly [14:07] let me describe the specs to give a better overview: it's 4x core processor (Intel or something), 4GB RAM running apache,postfix,courier,mysql and configured via ispconfig [14:07] mianosm: and yes, postfix is taking auth data from mysql [14:08] lshw -html > specs.html [14:08] <\sh> spajderix: dstat is also a good indicator of problems....most likely you have a disk problem or cpu 0/core 0 is under heavy load and let interrupts sequentially through or you have a real cpu problem or memory [14:08] moin \sh :) [14:09] <\sh> hey binBASH [14:12] \sh: I just received mail 5 of my system will be delivered today ;) [14:13] <\sh> binBASH: congrats :) send some pics when they arrive :) [14:13] in end of february ;) [14:13] The provider will setup all machines and place them in rack you know [14:14] I will go to datacenter in mid./end of February [14:14] and burn them down? [14:15] no [14:16] I will move some machines from old datacenter and place them in the rack as well [14:17] mianosm: lshw -html http://dl.dropbox.com/u/140067/specs.html [14:17] patdk-wk: http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/ [14:17] In some months I will need some of those ;) [14:17] heh [14:19] binbash, don't want to just go with something more stable: http://www.nexsan.com/satabeast.php? [14:19] can someone who is good with grub help me with this wierd issue. WHen booting my server ubuntu 10.10 with 2.6.35.24 kernel, the server will not boot. If I rebbot and go to the grub menu, press c for the command and then manually linux the vmlinuz-2.6.35-24-generic-pae file, and then initrd the same file, I can press esc, and select the kernel and it boots every time. Is there something I can add to grub to resolve [14:20] patdk-wk: And it costs 3 or 4 times more? ;) [14:20] probably, but is redundant, and has fc :) [14:20] so it's not limited to a single computer using it [14:21] can someone who is good with grub help me with this wierd issue. WHen booting my server ubuntu 10.10 with 2.6.35.24 kernel, the server will not boot. If I rebbot and go to the grub menu, press c for the command and then manually linux the vmlinuz-2.6.35-24-generic-pae file, and then initrd the same file, I can press esc, and select the kernel and it boots every time. Is there something I can add to grub to resolve [14:21] patdk-wk: What you mean with being redundant, the power supply? ;) [14:21] no, psu, controllers, network, fc, drives, ... [14:21] ahh ;) [14:22] patdk-wk: well, here I will take 2 or 3 of the boxes for being redundant. Wanna use ceph as software with radosgw [14:22] the issue I have with backblaze units, you would have to have drbd between two of them, incase one died [14:23] patdk-wk: http://ceph.newdream.net/ [14:23] no drbd needed [14:25] can someone who is good with grub help me with this wierd issue. WHen booting my server ubuntu 10.10 with 2.6.35.24 kernel, the server will not boot. If I rebbot and go to the grub menu, press c for the command and then manually linux the vmlinuz-2.6.35-24-generic-pae file, and then initrd the same file, I can press esc, and select the kernel and it boots every time. Is there something I can add to grub to resolve [14:26] hroberts: provide grub.conf in a paste [14:27] so glad you keep pasting that question before the last paste scrolls off the window - otherwise Id never know you asked it [14:27] lol compdoc [14:28] binBASH: please remind me where grub.conf is located at [14:28] locate grub.conf [14:28] hroberts: /boot/grub/menu.lst or grub.conf, depending on age. [14:29] what twb said [14:29] what's the difference between the ubuntu enterprise cloud and server? [14:30] Yale: a few letters [14:30] basically no difference? [14:30] the cloud is still evolving - seems to be many things [14:30] Yale: if you chose ubuntu server you can still build a uec with it ;) [14:31] # [14:31]