[00:18] is there a channel for KVM on server? or even just KVM? [00:19] #kvm [00:19] i know, sounds crazy right? [00:20] wow, yeah last time a guessed a channel name it wasnt pretty :'/ [00:20] but thanks :') [00:20] haha no worries === MinicMalkav is now known as MalkavianManiac [02:01] if CTRL-ALT-DEL isnt responding, are there any other hotkeys I can try? [02:09] osmosis, for what? [02:10] ctrl-alt-insert i think for certain vms [02:10] ctrl-alt-end i think is a rdp thing [02:10] MalkavianManiac, i found the sysreq reboot key === aarcane__ is now known as aarcane [03:06] Evening all.. [03:08] Could anyone assist me with an Apache2 / SSL question? [03:30] hi guys, I'm looking for some debugging pointers. I've got a server that is hanging in the reboot process. It stops _right_ before halt is called and the kernel complains about it on the console. The tough part is that I don't have any ttys at that point and the logging daemons have all stopped [03:31] does anyone have any idea how I can figure out what's holding up the server? [03:52] Hello [03:53] I have a question. If I change my user password on my server, will everything still work correctly end-user side? === zz_sailerboy is now known as sailerboy === sailerboy is now known as zz_sailerboy [05:17] any ideas on why a win2k3 kvm guest on ubuntu server works great on hardy, but runs super slow on lucid ? [06:27] date === trapmax is now known as xampart === twister004_ is now known as twister004 [07:17] ScottK: thanks, tag added (tag is server-mrs, stands for "maverickrelease status" and serves to generate the our Release status page) [09:11] Holy crap, glusterfs 3.1 looks awesome. [09:12] <_ruben> is that appropriate language in here ? ;) [09:13] For this sort of thing, I'm sure it is. :) [09:13] <_ruben> ok :) [09:13] <_ruben> so, what's so awesome about it? haven't looked at it in ages [09:14] <_ruben> or better yet: what do *you think* is awesome about it, in order to prevent a rtf website response ;) [09:15] Trying to find the summary... hang on. [09:15] GlusterFS has veen teh awsome since forever. Now it's being even more awesome by letting you do everything on-line. [09:17] http://www.gluster.com/community/documentation/index.php/GlusterFS_3.1beta# [09:17] Look at section 3. [09:20] <_ruben> looks pretty sweet indeed [09:31] New bug: #648604 in openldap (main) "package libldap-2.4-2 2.4.21-0ubuntu5.3 failed to install/upgrade: cannot access archive: No such file or directory" [Undecided,New] https://launchpad.net/bugs/648604 [09:31] New bug: #648605 in dovecot (main) "package dovecot-common 1:1.2.12-1ubuntu5 failed to install/upgrade: Package is in a very bad inconsistent state - you should reinstall it before attempting a removal." [Undecided,New] https://launchpad.net/bugs/648605 [09:36] New bug: #648609 in samba (main) "unable to unmount broken cifs mount" [Undecided,New] https://launchpad.net/bugs/648609 === MalkavianManiac is now known as _TechAway_ [10:01] New bug: #648619 in tomcat6 (main) "tomcat6 failed to purge" [Undecided,New] https://launchpad.net/bugs/648619 === _TechAway_ is now known as _Techie_ [10:05] hi, anyone got any advice on page allocation failure errors [10:05] as per http://pastie.org/1183944 [10:09] crb: Maybe you're out of swap. [10:09] crb: This is on EC2, right? [10:10] Yep [10:10] A micro instance? [10:11] nope, c1.medium [10:11] Mem: 1781976k total, 1463264k used, 318712k free, 193556k buffers [10:11] Swap: 917496k total, 60760k used, 856736k free, 439176k cached [10:11] for example [10:14] best I've found in other bug reports is bad drivers [10:28] soren: any ideas? [10:28] crb: Are you getting these errors on multiple instances? Have you tried moving to a new instance in case it's a problem with the underlying host? [10:29] erichammond: Yes, I'm getting them on four instances across two AZs [10:29] crb: :( [10:29] they're under quite a bit of load, they're our production web servers [10:29] indications from other bugs suggest it's when the kernel can't allocate RAM for connection tracking/netfilter [10:30] it /might/ be fixed in 2.6.35, but those error reports generally relate to a specific NIC driover [10:31] also reported here: http://developer.amazonwebservices.com/connect/thread.jspa?threadID=43939 [10:32] crb: You can expand the amount of space the kernel can use for nf contracking. [10:33] crb: What Ubuntu release? [10:34] esh: 10.04 [10:34] crb: cat /proc/sys/net/netfilter/nf_conntrack_count [10:34] You really should be getting a better error message if that's the limit you're hitting, though. [10:35] (That's the current count, not the limit) [10:35] not there [10:35] *blink* [10:35] The file is missing? [10:35] I don't think thats my specific error, thats what I found in reading a similar backtrace related to madwifi [10:36] I doubt I'm running conntrack [10:36] these are just web server nodes [10:36] Ah, right, if iptables aren't loaded at all, that makes sense. [10:36] my pastie doesn't refer to conntrack, though it does refer to nf_hook_slow, nf_iterate [10:36] i'm assuming nf == netfilter [10:37] It is. === th0mz_ is now known as th0mz === twister004_ is now known as twister004 === _Techie_ is now known as _TechAway_ [11:50] <_chris_> hej all [11:51] <_chris_> we have a small linux server here with a small hdd. we want to get a bigger hdd and move the whole installation on it, can this be done with clonezilla for example without problems ? === boki is now known as brontosaurusrex [11:57] _chris_, not sure if clonezilla is aimed at that sort of thing, but yes, you can simply do a disk image from one to the other [11:57] using dd, if nothing else [11:57] and then resizing the filesystem [11:57] whats the easy way to install apache h264mod? [11:57] streaming module that is [11:59] brontosaurusrex, doesn't look like it's in the repositories [11:59] I'd look for a PPA [12:00] qman__, ok [12:02] failing that, if the creator of the software doesn't have a .deb download, compiling from source is the last resort [12:02] if you have to compile though, use checkinstall or similar to create your own package, rather than just 'make install' [12:02] to make it easy to upgrade or remove later [12:04] qman__, nah, i will not compile myself and there seems to be no ppa on https://launchpad.net, only for lighttpd [12:39] New bug: #648527 in clamav (main) "continuance from password changer .Bug 647912" [Undecided,New] https://launchpad.net/bugs/648527 [12:50] What is the prefered tool to encode WAV to MP3? [13:01] a_ok: lame [13:28] morning [13:35] how can i view what is in zip file without unziping it ? [13:37] classical__: `unzip -l /path/to/file.zip` [13:39] joschi, txh [13:39] joschi, thx* [13:58] Heya - you might see this question alot but I've got to ask it anyway. Are there any known issues in with Ubuntu when using ISPConfig 3 and Jailkit (version 2.12, most recent version) - as I can't for the life of me get new users to be created jailed! :( [13:59] I followed the Perfect Server guide for Ubuntu 10.04 guide on HowtoForge [14:06] The main issue I seem to be having is the shell is always set to /bin/false, not /usr/sbin/jk_chrootsh when a new user is set as Jailed [14:34] hello all [14:35] Hi yellabs [14:35] what would be the option to install ubuntu server , to run on one computer and , then multi virtual machines on that pc === Seidr2 is now known as Seidr [14:38] I can't say I've had a whole load of experience in running VMs on a server, but I have used VMWare in a limited capacity before [14:38] or in other words, the best way to use ubuntu for virtualisation [14:38] esx server ? [14:38] Someone else might be able to give you a better answer though - if anyone else is awake that is. I'm waiting for someone to answer myself [14:39] Seidr, and what was your question ? i might know that [14:40] It's regarding ISPConfing, Jailkit(2.12) and Ubuntu 10.04. I followed the Perfect Server guide on HowtoForge to the letter (including installing Jailkit BEFORE ISPConfig) yet any user I create jailed from ISPConfig has their shell set as /bin/false [14:40] there doesn't seem to be any interaction between ISPConfig and Jailkit at all going on in the logs [14:43] yellabs: vmware esx is another operating system. you probably mean vmware server [14:43] yellabs: what kind of virtualization do you want? os-level virtualization (e.g. openvz), paravirtualization (e.g. xen) or full virtualization (e.g. virtualbox, kvm, vmware server)? [14:45] i thought kvm would be closer to ubuntu [14:46] ISPConfing, Jailkit(2.12) and Ubuntu 10.04. , thats a hard one [14:47] I was afraid of that - the guide calls for 2.11, however the changes between 2.11 and 2.12 are very minor so I doubt it's a compatibility issue [14:47] especially as I'm seeing no errors during ISPConfig setup and no errors in the auth log [14:48] i think you followed this tutorial ? http://howtoforge.net/perfect-server-ubuntu-10.04-lucid-lynx-ispconfig-3-p4 [14:48] that's correct [14:49] and you are not the only one with the error , http://serverfault.com/questions/185012/ubuntu-10-04-pefect-server-jailkit-2-12-not-working [14:49] but there are few who have attempted to do it [14:49] Hah - uhm, that would be me :) [14:49] hehe [14:49] I'm just about to go through the process on a VM, to be sure I've not made a stupid mistake somewhere [14:49] (it was late at night when I setup the server, so it's entirely possible) [14:50] SpamapS: ping [14:50] but I've even gone back, removed ISPConfig and made doubly sure Jailkit is installed before proceeding again, so I'm stumped [14:50] maybe the best thing, although it takes time, is to find out who wrote the tutorial and e-mail him / her [14:50] Mm - agreed - once I've gone through this VM setup and confirmed the problem I think I'll do just that. [14:51] Thanks for your time yellabs [14:52] or install it again , but using the old 2.11, as said in tutorial [14:53] god luck [14:53] good luck * [14:53] bye bye [15:17] hggdh: ping [15:18] zul: at your service, sir [15:18] hggdh: did you reproduce that dovecot bug? [15:19] zul: I had it happening to me while going thru a dist-upgrade on Saturday [15:19] hggdh: how did you reproduce it [15:19] did you have like proposed enable as well? [15:20] zul: I did not _reproduce_, I just found a bypass. I am still at loss on what might have caused it... I was going, now, to reinstall dovecot to see if it triggers the issue [15:21] (apt-get --reinstall install dovecot-...) [15:21] hggdh: gotcha === ivoks-afk is now known as ivoks-zombie [15:21] but *both* 'stop dovecot' and 'start dovecot' would hang [15:22] zul: proposed on Maverick? [15:29] hggdh: proposed on lucid [15:29] zul: I am running Maverick [15:29] k [15:34] hggdh: k im able to reproduce it [15:37] hi, i'm trying to pick an iso image when creating a new vm with virt-manager, but the file dialog doesn't show any directories or files [15:42] and if i specify the path manually, it doesn't find the media :( [15:51] * zul shakes his fist at upstart [15:53] zul: having fun? [15:53] JamesPage: %$#!$%...no [15:54] lol [16:12] zul: I have a small issue with upstart: my /tmp fs does not get cleaned on restart. So I run upstart with --debug and... it works... [16:12] so: welcome to the group :-) [16:13] hggdh: ?!?! [16:13] zul: the group of people surprised by it [16:14] SpamapS: reping [16:22] smoser: i'm away Thursday, will you be around to drive the cloud images RC release ? [16:22] yes. [16:22] ttx, do you know of any reason not to use today's images for expected RC ? [16:22] ie, that i can't start testing today ? [16:23] the version of lighttpd available (1.4.19-0ubuntu3.1) on Hardy has a bug ( http://redmine.lighttpd.net/issues/897 ) that's causing me problems... aside from doing a dist-upgrade or manually installing a new version of lighttpd, are there any other solutions? I don't think there's a backport of a newer version available... [16:24] smoser: I'd use the first one generated after the freeze (00:00 UTC tomorrow) [16:24] ah. ok. [16:24] smoser: for consistency reasons (RC in cloud images ~= RC in the rest) [16:24] yeah. definitely. [16:24] smoser: unless you *know* something introduced today breaks everything :) [16:24] well, i know, i'm just not telling [16:25] smoser: good good. [16:36] Any of you know of a good heartbeat system for monitoring server and website uptime? [16:49] ttx: ppppooonnnggggg [16:49] SpamapS: heeyyy [16:49] SpamapS: so everything freezes tonight, 00:00 UTC on Tuesday [16:50] ttx: Haven't seen feedback on the apache bug [16:50] SpamapS: I think your apache proposed fix is a bit too scary at that point, do you concur ? [16:50] SpamapS: I commented on the BMP [16:50] ttx: drizzle is nearly done [16:51] ttx: http://people.canonical.com/~chucks/dovecot-debdiff.patch [16:51] ttx: yes, I think it needs a good solid review by the security team, though its straight forward enough I think it should be SRU'able if they're comfortable with it. [16:51] SpamapS: full ack. I'll remove it from release radar [16:52] I need to increase the number of file open handlers for ubuntu 10.04...I have done this for 8.04 but I have a feeling that the /etc/pam.d/common-session session required pam_limits.so isn't taking...anyone do this for 10.04 yet? [16:54] hi...is this the ubuntu server development channel? [16:55] i'm looking for help with a support(ish) question on ubuntu server [16:56] i heard that ubuntu server 9 is coming to the end of its support lifetime...will it be challenging to upgrade a server running ubuntu 9.04 that's hosting a web application? [16:58] i'm about to go live with a new web application, but the hosting company only provides ubuntu 9.04 on their dedicated servers...i'm concerned that if something comes up that forces me to upgrade ubuntu, i'll be stuck as it will be in production and every minute of downtime will be costly [17:00] kinygos: best bet would be to install 9.04 on another server, configure your software, do a dist upgrade and see what happens [17:01] kinygos: don't do it directly on production. [17:01] kinygos: upgrading the os is the easy part..making sure your application still works is the hard part. [17:16] Ubuntu JeOS, I'm unable to get programs that segfaults to produce core dumps. I've tried to "ulimit -c unlimited", but this does not come into effect as coredump is stuck at 0. Is there a way to fix this? [17:20] jiboumans: how are we looking for starting RC testing tomorrow? [17:22] Hi, I was wondering what the lastest upgrades were; linux-headers-2.6.32-25{a} from linux-headers-2.6.32-24{u}? [17:26] I don't beleive there is a changelog entry === aliverius_ is now known as aliverius === ivoks-zombie is now known as ivoks [17:45] ttx: o/ - I'll take the RC release management over [17:50] smoser: hey. this may be a dumb question, but a) what is the most up to date ec2 image for maverick and b) can you point me to the latest/most up to date/relevant wiki page for ec2 and ubuntu? [17:51] https://help.ubuntu.com/community/EC2StartersGuide is probably the best place to start. [17:51] it has pointers to maverick images [17:51] New bug: #645773 in chkrootkit (main) "chkutmp crashed with SIGSEGV in _Unwind_Backtrace()" [Low,Incomplete] https://launchpad.net/bugs/645773 [17:51] which is [17:51] http://uec-images.ubuntu.com/server/maverick/current/ [17:51] smoser: perfection. thanks [17:52] and available programattically from stuff under http://uec-images.ubuntu.com/query/ [17:52] ie: http://uec-images.ubuntu.com/query/maverick/server/daily.current.txt [18:09] mathiaz: ping did you have a look at ttx's email? [18:09] zul: yes [18:09] zul: what's the status of bug 646858 [18:09] Launchpad bug 646858 in dovecot "dovecot-postfix and upstart incompatibility" [High,Confirmed] https://launchpad.net/bugs/646858 [18:09] zul: ? [18:09] mathiaz: that debdiff fixes the bug [18:10] and we got confirmation from someone else that it fixes it [18:11] mathiaz: i just want someone to review the debdiff and get an ack [18:11] * mathiaz looks at the bug [18:17] zul: were you able to reproduce the bug? [18:17] mathiaz: yep [18:18] zul: what is the problem? [18:18] mathiaz,: hggdh did as well [18:18] yes [18:18] mathiaz: upstart looses track of the pid so it doesnt shutdown [18:19] zul: I'm trying to understand what the problem is in order to check that the patch addresses the correct problem [18:20] zul: is there any debugging traces that the problem with upstart is losing pids? [18:20] mathiaz: thats fine because it does [18:20] mathiaz: no there isnt [18:20] zul: so how do you know that upstart looses track of the pid? [18:21] mathiaz: because when i do a status dovecot it reports the wrong pid [18:21] thesheff17: thank you very much for the advice...apologies for the delay, i had a real life issue to deal with :) [18:22] New bug: #649114 in samba (main) "package winbind 2:3.4.7~dfsg-1ubuntu3.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/649114 [18:22] kinygos: np [18:23] zul: do you have any idea why ttx is not able to reproduce the problem? [18:23] zul: hggdh: what is different in your environement? [18:24] mathiaz: no..but to reproduce it i had to set the postfix to local configuration [18:25] mathiaz: its possible he didnt have postfix configured [18:26] zul: I think the debdiff is correct. Given that where we are in the release cycle (about to enter RC freeze) I'd like to make *sure* the patch is actually correct [18:26] zul: and not a shot in the dark like... hm it seems to work [18:27] zul: to do so I'd like to fully understand what the problem is [18:27] zul: and prove that the patch is addressing the issue [18:27] mathiaz: go ahead ill be waiting :) [18:27] mathiaz: I do not run postfix [18:28] zul: hggdh: could you outline a specific configuration that shows the problem? [18:28] and then start upstart in debug mode to figure out why it gets stuck? [18:28] mathiaz: sure configure postfix to be a local in lucid and upgrade to maverick [18:28] hggdh: is your setup easier? [18:29] hggdh: is it an upgrade from lucid as well? [18:29] mathiaz: no, I was already in maverick, running a dist-upgrade that included dovecot [18:29] hggdh: do you have postfix installed in the system that show cased the bug? [18:31] mathiaz: just confirmed, I do have it, for local delivery [18:31] hggdh: is dovecot-postfix installed? [18:31] mathiaz: no [18:32] hggdh: is dovecot configured to use SASL? [18:33] mathiaz: no. Active dovecot config: http://pastebin.com/TXRnHh1L [19:02] As .img files are not produced any more, what is the best way to install latest server edition via USB? [19:03] ziarkaen: you can create bootable usb drive with usb-creator [19:03] ziarkaen: and a -server iso [19:04] Sorry, I'm not running Ubuntu; I'm using another distro. [19:04] Is that utility released as source tarball? [19:06] zul: nope [19:06] zul: I haven't reproduce the issue yet [19:07] zul: could you start up a test system that is not working and turn upstart debug on? [19:07] zul: the goal is to check why dovecot fails to restart [19:07] mathiaz: sure but you should able to do a sudo start dovecot and then sudo stop dovecot as well to reproduce it [19:08] zul: well - I tried - and it works as expected [19:08] mathiaz: did you have postfix-dovecot installed? [19:08] zul: yes - I've install dovecot-postfix [19:08] from lucid? [19:08] zul: note that this is a brand new maverick install [19:09] mathiaz: heh - you're playing with fire [19:09] mathiaz: please try the upgrade from lucid with dovecot-postfix installed and postfix configured [19:10] hggdh: is your system also an upgrade from lucid? [19:10] hggdh: is your *dovecot* system also an upgrade from lucid? [19:10] zul: could you compare your lucid dovecot configuration to maverick configuration? [19:10] zul: in /etc/dovecot/ ? [19:10] zul: there may be a change during the upgrade? [19:11] mathiaz: yes it prompted for a chance [19:11] change even [19:11] zul: so the old configuration files should be laying around on the upgraded system [19:11] zul: as .dpkg-{old, new} [19:11] zul: what has changed? [19:12] mathiaz: just a sec [19:12] mathiaz: yes, I upgraded it from Lucid some 3 months ago [19:12] hggdh: ok [19:13] mathiaz: and I had to hand-stitch the Lucid config for Maverick [19:14] mathiaz: http://paste.ubuntu.com/501671/ [19:14] zul: what's the content of include_try /etc/dovecot/auth.d/*.auth [19:15] zul: ? [19:15] http://paste.ubuntu.com/501675/ [19:16] zul: and include_try /etc/dovecot/conf.d/*.conf [19:16] zul: ? [19:17] http://paste.ubuntu.com/501677/ [19:19] ziarkaen: http://www.revouser.com/forum/viewtopic.php?f=7&t=794 [19:19] ^^ is that helpful... [19:19] holstein: Error: "^" is not a valid command. [19:21] holstein, That seems like a specific installation problem; I'm more after an official method to install ubuntu-server from another distro with usb. [19:22] ziarkaen: can you use unetbootin on that distro? [19:23] holstein, Yes. [19:23] i havent tried it personally, but i have suggested it with the ubuntustudio iso [19:23] another alternate installer [19:23] and i was told it worked [19:25] holstein, Thanks. [19:31] smoser: last question. looking at http://uec-images.ubuntu.com/query/lucid/desktop/daily.current.txt I see 'ebs' and 'instance-store'. what is the difference? (I know ebs is elastic block storage, but don't know what that means in the report) [19:32] https://help.ubuntu.com/community/UEC/Images has a *little* more info on what you can find there. [19:32] but basically that is the 'root store' . there are two types of instances, ebs root and instance-store. [19:32] ebs root has root filesystem on a ebs volume [19:32] that is what I was thinking [19:33] instance-store has root filesystem on the instance store (ie, when system goes away, so does the root fs) [19:33] smoser: so ebs is going to be more expensive? [19:33] you pay 10 cents per GB/Month of provisioned storage [19:33] and you pay per IO ops also onto ebs [19:33] k [19:34] so, yes, it will be $1.50 / month + IO ops [19:34] smoser: thanks again [19:49] smoser: hrm: [19:49] ec2-run-instances ami-4ef9a80b -k ec2-keypair -t c1.xlarge [19:49] Client.InvalidAMIID.NotFound: The AMI ID 'ami-4ef9a80b' does not exist [19:50] that is in us-west-1 [19:50] ec2-run-instances -region us-west-1 <....> [19:50] ah [19:50] I didn't see that in https://help.ubuntu.com/community/EC2StartersGuide [19:50] --region [19:50] (i missed a -) [19:51] http://uec-images.ubuntu.com/server/maverick/current/ has cut and paste lines [19:51] i think that SpamapS has some "launch latest" scripts [19:51] thanks [19:51] but he hides them from me [19:52] smoser: because you are not apart of the secret society...besides you have to show the dance [19:55] zul: ok - I'm able to reproduce bug 646858 [19:55] Launchpad bug 646858 in dovecot "dovecot-postfix and upstart incompatibility" [High,Confirmed] https://launchpad.net/bugs/646858 [19:55] zul: the problem seems triggered by managesieve [19:55] zul: without managesieve dovecot doesn't spawn an extra process and upstart is able to keep track of it [19:55] New bug: #649189 in samba (main) "Samba fails to bind IPv6 on boot" [Undecided,New] https://launchpad.net/bugs/649189 [19:56] mathiaz: acked [19:56] zul: with managesieve enabled dovecot spawns another process and upstart looses track of it [19:57] mathiaz: have you tried the debdiff? [19:57] Is it possible to kickstart lucid install with lvm on sw raid, or i have to use preseed? [20:12] zul: the debdiff looks good - please go ahead with the upload [20:12] mathiaz: thanks [20:12] After installing LAMP, I would like apache to use a www directory on and ntfs partition as the webroot directory. Is that possible? [20:13] is bug 648202 something we should consider for Maverick? I personally do not think so, but... [20:13] Launchpad bug 648202 in vsftpd "vsftpd started even if not in standalone mode" [Low,Confirmed] https://launchpad.net/bugs/648202 [20:13] mathiaz: thanks [20:15] help with apache configuration anyone, or should I be in another channel? [20:19] heyblondie, your config is possible [20:19] (s)he left [20:39] SpamapS: hi - are you working on the drizzle FTBS? [20:41] Daviey: hey! [20:41] Daviey: any news on bug 644733? [20:41] Launchpad bug 644733 in eucalyptus "On upgrade from Lucid to Maverick CC entries in /etc/eucalyptus/eucalyptus.local.conf are commented out" [High,Confirmed] https://launchpad.net/bugs/644733 [20:43] mathiaz, working on it atm [20:44] Daviey: great - will you need sponsoring? [20:44] mathiaz, aye [20:44] mathiaz, Well actually, right now... i'm trying to sort out a db schema issue [20:44] * mathiaz guesses it means yes? [20:44] heh [20:48] Daviey: I'll be offline for a few hours [20:48] mathiaz, ok [20:48] Daviey: I should be back a couple of hours before the freeze is in effect [20:48] Daviey: which is at 00:00 UTC [20:48] mathiaz, cool [20:48] yeah [20:57] smoser: ok, so I finally got an image started. I can't ssh, so I looked at the console and see: http://paste.ubuntu.com/501737/ [20:57] smoser: is this known? [20:57] smoser: this is ami-f8638991 [20:57] bug 613273 [20:58] Launchpad bug 613273 in linux "kernel panic on ec2 in system_call_fastpath" [Medium,Confirmed] https://launchpad.net/bugs/613273 [20:58] restart it [20:58] ec2 reboot-instances [20:59] smoser: should I just do that until it comes up? [21:00] it should come up next time [21:03] hey all i'm having an issue with raid45 - i keep getting : "raid45" is not in the kernel - how can i fix this so my raid5 works? thanks - im installing ubuntu server 10 amd64 [21:04] smoser: yes, it did. I got confused for a second forgetting that the console is not updated immediately [21:05] max2, do you want to use mdmonitor and software raid? or do you have an onboard fakeraid chip? [21:05] I have an intel board with SATA Raid - and 3x 1GB drives [21:05] jdstrand, yeah, thats a pita [21:06] smoser: sorry about me being a pita-- it's been a little while since I cranked up an instance :) [21:06] in the raid bio config i setup a RAID5 for the 3 drives - but i keep getting "is not in the kernel" error [21:06] max2, then lucid-server might be missing the dmraid [21:06] http://forums.citrix.com/message.jspa?messageID=1016922 [21:06] jdstrand, no problem. [21:06] its a bug that has been an issue for over 2 years [21:06] Similar question -- can I put together a software raid (via mdadm) and use multiple hard drives + a partition that is slightly smaller without reducing the usage of each disk to the smaller size? [21:06] raid45 is the dmraid stuff [21:07] max2, try the alternate CD [21:07] why? i just got this iso off the site [21:07] <_ruben> JavaAtom: only when using raid0 (which isnt really raid) [21:07] its the only x64 iso avail [21:07] _ruben: So I can't do a raid-5 of sorts like that? [21:08] <_ruben> JavaAtom: raid1/5/10/etc all require equal sized members [21:08] _ruben: ahhhh shazbot. [21:08] <_ruben> JavaAtom: and thus will use the largest *common* size available [21:08] well, dropping 30G off of each 1TB drive isn't *too* bad, you think? [21:09] (I've got six drives) [21:09] no one knows how to get raid5 working? [21:09] <_ruben> JavaAtom: why would you chop off a part of only 1 disk? [21:09] <_ruben> max2: unless you plan to dual boot with windows or something similar, forget about the raid bios and go with plain software raid (mdadm) instead [21:10] _ruben: because one of the discs has a normal partition scheme (within about 30GB) and then the rest was used for an LVM partition, which got lumped together with five other 1TB drives to give nearly 6TB of usable space. [21:10] <_ruben> so how should my STAT drives be setup then? [21:10] _ruben: I'd like to use the remaining 900GB on that first disk without installing the whole OS to the raid. [21:11] isnt hardware raid better then software raid? [21:11] <_ruben> max2: fakeraid isnt hardware raid [21:11] max2: It's supposed to be, but software raid can do interesting things. [21:11] why will raid5 not work? [21:11] <_ruben> hardware raid >> software raid >> fakeraid [21:11] ? [21:11] <_ruben> dunno, never cared for fakeraid myself [21:12] <_ruben> software raid tends to outperform it easily [21:12] so i should just set my raid sata drives as notmal drives and use LVM? [21:13] hah. I just had a drive fail on that setup [21:13] <_ruben> JavaAtom: personally i'd go for creating 2 raid volumes, one for os, one for storage, or just one big one, and carve up using lvm for both os and data [21:13] i just want to use raid5 with linux [21:13] why is that so hard [21:13] this sucks [21:13] <_ruben> max2: lvm has nothing to do with this [21:13] Hi, I have an md RAID 10, with one disk that drops because of bad blocks. is there a way to use badblocks or something to get it to just map out the bad ones until we can replace the disk? [21:13] <_ruben> software raid can do raid5 just fine [21:13] <_ruben> my home fileserver has software raid5 over 6 1tb disks [21:14] max2: See, the complaints don't help. "This sucks" is counterproductive and generally makes people like myself unhelpful. [21:14] sorry [21:14] im just frustrated because this has been a bug for years [21:14] max2: No worries. I feel your frustration. [21:14] <_ruben> max2: configure bios as normal sata disks, and in ubuntu server installer configure the software raid, and then lvm on top of the software raid [21:14] _ruben: Maybe I can do what you just suggested to max2... [21:14] <_ruben> max2: the real bug is hardware vendors still shipping fakeraid chipsets [21:14] so i should set all my disks to "non-raid" [21:14] <_ruben> max2: yes [21:15] <_ruben> max2: and then let the os (ubuntu) do the raid part [21:15] okay ill do that [21:15] <_ruben> easier to maintain, and most likely better performance as well [21:16] <_ruben> Norkakn: not that im aware of, i'd just pull the disk and get a replacement asap [21:16] _ruben: thanks. I remember them thinking about adding support for it, but I'm not seeing anything about it going live [21:17] <_ruben> Norkakn: ic, i must admit i dont follow the software raid (mdadm) development very closely [21:20] <_ruben> ugh .. writing ip addressing schemes is a pita .. ipv4: too little ip addresses .. ipv6: too much ip addresses :p [21:22] _ruben: heh -- I'm in the middle of a migration from 192.168.0.0/24 to a 10.72.0.0/16 [21:23] <_ruben> JavaAtom: smart move :) [21:23] <_ruben> cant wait to get rid of our 10.0.0.0/24 .. it's almost as bad as using 192.168.{0.1}.0/24 ;) [21:23] _ruben: Yeah, well, when you have 8 ip addresses free and you hire three people... that should throw up some flags. :P [21:24] <_ruben> haha [21:24]