=== _Techie_ is now known as _TechAway_ === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates === _TechAway_ is now known as _Techie_ [00:46] hi, anyone on? [00:49] does anyone chat in this chat room? [00:51] yes [00:51] hi spam [00:51] hi all [00:51] hi tan [00:51] i have a question about 64-bit 10.04 install.... [00:52] install won't mount the cd/dvd drive. tried everything on the Ubuntu site and forums [00:53] some help is cryptic, like "there should be an entry for /dev/cdrom...", but doesn't say what to do if not! [00:54] HP does not have any drivers except win7...there are no IDE ports to try an older cd.... [00:55] drivers for a cd drive? [00:56] Thats pretty standard/generic hardware.. why would it need special drivers? [00:56] no idea...blueray/dvd/cd combo... [00:57] boots to graphical install menu, gets past "kb identification", won't recognize or mount the cd/dvd [00:57] it is a sata device...apparently is seeing the sata hd [00:58] tried changing controller to "ide" in bios, same effect [00:58] have you asked in #ubuntu? [00:59] yes, they said "go to #ubuntu-server"...it is 64-bit server edition on cd from Ubuntu [00:59] fedora 13 32 bit loaded fine [01:00] win2k3 server needed bios set to "ide" for some reason [01:00] attempting to overwrite those installs with 64-bit ubuntu server 10.04 [01:01] timmy: probably worth posting it as a bug report [01:01] timmy: if somebody else has your same identical hardware, they can confirm or help out. [01:02] someone has...nothing done yet... [01:02] everything says "try install from usb..." that seems silly tho [01:03] may have to try it, or get a pci-e ide controller to run a different cd rom. [01:03] honestly, its pretty silly to install from spinning discs when you can just install from a solid state drive that is light weight and reusable. === _Techie_ is now known as _TechAway_ [01:06] New bug: #632791 in whois (main) "upgrade fails: trying to overwrite '/usr/bin/mkpasswd', which is also in package mkpasswd 5.0.6ubuntu1" [Undecided,New] https://launchpad.net/bugs/632791 [01:08] but spam...that is the entire problem not only with linux, but with it in general: [01:09] a cd should just run the prog...period. like turning a channel on a tv. [01:09] if SSD is way to go, they should offer that as media [01:09] the BIOS runs the program [01:10] the BIOS is written by the HW manufacturer, and has its own set of drivers to read from the CD [01:10] it is as if it would be acceptable to get a car, then be told "it won't start? oh, well just put a space shuttle engine in it...it's not hard" [01:10] now, the CD SHOULD be playing nice as a normal SATA disc drive [01:10] so its kind of weird that the kernel can't see it as such [01:10] but the bios did its job [01:11] i just don't get why there isn't a generic cd driver that would work....like vga is to vid, you know? [01:11] THANKS FOR YOU TIME ANYWAY. [01:11] in this case, somebody paid the car manufacturer to design and engineer it to work in a given set of parameters [01:12] your frustration is understandable completely [01:12] Hi everyone. Can anyone give me a hand with Ipfw? [01:12] I'm in total agreement that it doesn't make sense that this particular CD drive wouldn't work [01:13] timmy: for your car analogy, Linux is like putting an after market chip in the car's computer systems. Sure everything *should* work as expected.. but the people who made the car probably didn't test it on your custom chip. ;) [01:13] mathiaz, did you open a bug ? [01:14] timmy: I have to run. Good luck with your issue (and I really do suggest trying a USB key. you should be able to make one from the server cd iso.. ) [01:43] Hey. RackSpace uses lame mirrors. Does anyone know how to change my mirror easily? [01:45] hello. why does samba appear to be missing ldapsam built in by default on ubuntu samba 3.4.7? [02:35] Hi everyone. Can anyone give me a hand with Ipfw? [03:45] does anybody know how to update root to user on mysql [03:45] What does "update root to user" mean? [03:46] well change root to user [03:46] for mysql [03:49] nvm I just created a new user === _TechAway_ is now known as _Techie_ [04:26] You know how 2TB drives changed the block size from 512b to (IIRC) 2kB? [04:27] Will my hardy server Just Work with such drives? [04:28] Make that 4kB blocks. [04:30] twb: I know there has been some work done during karmic or lucid cycle to improve support for such disks. I'm not sure if it has been backported to hardy. [04:31] New bug: #632863 in bacula (main) "package bacula-director-mysql 5.0.1-1ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/632863 [04:37] * twb reads http://public.dhe.ibm.com/software/dw/linux/l-4kb-sector-disks/l-4kb-sector-disks-pdf.pdf [04:40] IIUC the bottom line is that 1) if you buy something labelled "Advanced Format", it has a firmware layer that presents 512b sectors to the OS; and 2) such a drive will Just Work, but will perform significantly better if you take the time to align on the invisible 4k boundaries. [05:29] hi all [05:31] Don't mind everyone rushing in to say hello mate... :) [05:32] lol. [05:32] i see [05:32] ...might be overwhelming... [05:33] no doubt..well, here is the thing..i have just gotten the love of taking over a ubuntu server and i've little experience in this domain..i think i have a networking issue as the server is slower than molasses [05:33] when i do a netstat -s i see a lot of errors and such..so i'm not sure if the box is comprimsed or if there is a network setting error [05:33] vanirahto, Slow in what respect - network response, or SMB or...?? [05:33] network response.. [05:34] like it takes forever to connect via ssh [05:34] ping times are good [05:34] can't use sftpd now either as it times out [05:34] vanirahto, Well, you can check the /etc/resolv.conf and see what you're dealing with as far as DNS servers... => and as well, is this machine setup for a static ip? [05:35] it has 2 ip addresses..one is dhcp and the second is static [05:35] the ip i'm connecting to is dhcp [05:35] Roger that...here's a question for ya - is both the static and dhcp setup for the same dns and all that wonderful jazz? [05:36] yes indeed..that is all given by the hosting company [05:36] i can send you the output of the netstat -s command [05:36] if you like [05:37] vanirahto, Ok...roger that...have you checked to see - and this may impact overall performance, what the contents of the /etc/sysctl.conf are like? Are they "generic" or have they been tweaked? Cuz you can eek out some performance on your network from some settings there... [05:37] vanirahto, You can use the pastebin for that mate [05:38] pastebin? sorry but i'm really a newbie to this whole thing..first time w/irc et all [05:38] am checking on the etc/sysctl.conf now [05:38] Roger that...hang on... [05:39] http://pastebin.ca [05:40] vanirahto, http://www.ubuntugeek.com/performance-tuning-with-system-control-sysctl-in-ubuntu.html [05:42] here is the post: http://pastebin.ca/Yi7UHja4 password is ubuntu [05:45] vanirahto, Far out - ok - firstly, you may want to change the box's root password. Then you may want to tweak the /etc/sysctl.conf => and apply the settings. Then you may want to check/reconfigure the firewall on this box. As well, you might want to double check/tweak the network settings and then restart the network. Check, also, if you've got enough memory/swap for giggles and grins. Turn off any services that are j [05:45] ust, well, not used or not wanted. [05:45] vanirahto, What version of Ubu is on this thing? [05:45] right on..am checking on the version [05:47] 2.6.16.53-070731a [05:48] vanirahto, Do this in your console: cat /etc/lsb-release [05:49] k [05:49] DISTRIB_ID=Ubuntu [05:49] DISTRIB_RELEASE=6.06 [05:49] DISTRIB_CODENAME=dapper [05:49] DISTRIB_DESCRIPTION="Ubuntu 6.06.2 LTS" [05:50] vanirahto, Egads - um...kinda a bit outdated that...however, you'd like to have this box stay up as long as possible with the least amount of work/time/effort. [05:51] yeah..its an old box..never had any major problems w/it till recently when it started crawling [05:51] That being the case, follow through with what I had told ya - tweak the /etc/sysctl.conf, tweak the firewall, tweak the network settings. [05:52] vanirahto, I'll assume that first and foremost, especially on seeing the netstat output, that you're going to see an increase in performance just by tweaking the /etc/sysctl.conf and applying the settings - then restarting the network on that box [05:52] right on..can you point me in the right direction to tweak the firewall and network settings? i'm looking into the sysctl.conf now [05:52] ok..i'll work on those points now..very odd though that just 2 days ago all was well on the box.. [05:52] Do the sysctl.conf firstly mate. [05:52] right on! [05:53] Could be that it's a target mate. [05:53] i see a lot of entries for the sendmail mta when i do a ps -ef [05:54] vanirahto, Have you audited the users? Check to make sure they're for-real - and check the mailboxes and forwarding and all that wonderful jazz? [05:54] in the process [05:54] have not done a full audit [05:54] but i have a feeling that is what i'll be doing ALL night :( [05:55] vanirahto, Paid by the hour, or salary? :) [05:55] me? man, thats a long long story there [05:55] vanirahto, Hehehehe...been there done that mate. [05:55] lol..me too..was hoping this would not be the second time around though [05:56] short answer: salary but DEF not enough [05:56] lol [05:56] For us, it's never ending. [05:56] must be done i guess.. [05:56] ok..let me check on those things and i'll be back in a bit..are you online longer? [05:57] its ur daytime now, eh? [05:57] Yes - well, it's arvo here - might duck out for a bit, but mostly here (in my lab) all day... [05:57] right on! :) thx! === rmk_ is now known as rmk [06:50] Does vsftpd have internal per-user quotas, or is my only option normal filesystem quota(8) quotas? [06:52] twb: the latter [06:52] Thanks [07:01] Anybody remember if you can leave the filesystem mounted (either -oro or -orw) while generating the initial aquota file? [07:02] *aquota.{usr,grp} === crazygir is now known as Guest54812 [09:33] morning === ivoks is now known as ivoks_away [09:37] mornin' [09:37] Are there any cloud weenies that have actually looked into the issues of data sovereignty and legal jursidictions? [09:38] ups, twb, this is not a easy one [09:39] OK, here's an easier one: [09:39] tmpfs defaults to a 50%-of-RAM cap [09:39] If I mount and fill two tmpfs, will RAM be 100% used or 50% used? === jjohansen is now known as jj-afk [09:55] twb: No. [09:55] :) [09:56] twb: There's a difference between tmpfs and ramfs. [09:56] New bug: #632997 in mysql-dfsg-5.0 (universe) "package mysql-server-5.0 5.1.30really5.0.75-0ubuntu10.5 failed to install/upgrade: el subproceso post-installation script devolvió el código de salida de error 1" [Undecided,New] https://launchpad.net/bugs/632997 [09:56] twb: tmpfs can be swapped out. [09:56] soren: *if* there's swap [09:57] twb: Sure. [09:58] twb: Well, if there isn't, then of course you're going to be 100% full. [09:58] New bug: #626723 in apache2 (main) "init script resets isig flag in an incorrect manner" [High,In progress] https://launchpad.net/bugs/626723 [09:58] twb: Why wouldn't you? [09:58] I wondered idly if maybe tmpfs was cleverer than that [09:58] How could it be? [09:58] I mean.... What could it possibly do? [09:59] Some kind of shared pool with an overall cap as well as per-mountpoint caps [10:00] I doubt it. [10:00] ramfs might. [10:00] twb: Can you elaborate on your "data sovereignty and legal jursidictions" concerns? [10:01] OK, suppose I'm a multinational with a head office in, say, Zurich. [10:01] In my Zurich offices I have a machine room and a bog standard rack full of gear in it. [10:02] For most intents and purposes, I own the hardware, software, and data on it, and legal issues take place in a Zurich jurisdiction. === ivoks_away is now known as ivoks [10:03] Now suppose I replace most of that with images running in a cloud provider's (e.g. amazon?) cloud [10:03] Mm hm. [10:03] Who does the data (for various values of "data") belong to? [10:04] Depends on the terms of service, I suppose. [10:04] If some flunky in .my decides his bribe wasn't big enough this week, and siezes a cloud node in his city which happens to be running your image or hosting part of your database, what happens? [10:05] It.. goes away? [10:05] I'm not sure what you mean. [10:05] soren: but is he allowed to look at it? Is he allowed to sell all your secrets to another multinational? [10:06] twb: According to what? In most (all?) countries, stealing is illegal. [10:06] personally, I think there's a time and a place for 'cloud'. and if such concerns actually feature as part of your disaster recovery process, it may not be the place for you [10:06] Well, I'm not a multinational, but I know at least one that decided not to do cloud-y things because of such concerns [10:07] twb: Of course you need to trust the provider before you give them all your data. [10:07] soren: so know I need to know if such a scenario *is theft* according to... which jursidictions? Mine? My provider's head office? Every country in which the cloud nodes live? [10:08] twb: Alternatively, you can set up an instancen at 10 different providers, run tahoe lafs on them, and scatter your data all about the place. [10:08] twb: Very likely the place where the nodes live. [10:08] twb: Everything else is an agreement between you and the provider. Or so I think. [10:08] IANAL. [10:09] So my original question is: has anyone actually thought that stuff through, maybe asked their lawyers, and written up a summary [10:09] twb: I think the perfect question is what jurisdiction does the data come under etc, so w/ Amazon you'd likely have to obey w/ US laws, plus there is a chance that US Govt could order access to your data etc w/o you knowing (going from what I've read) [10:11] Hey, I just noticed that a vm I created with ubuntu-vm-builder from lucid creating a lucid vm was missing update-notifier-common which sets the APT:: confs for periodic updates; it's in the server seed though, is this normal? [10:11] twb: I don't know. I've never really cared about this stuff. Stuff can be illegal all it wants, that doesn't stop it from happening, so if I care about my data, I don't put it all in one place that I don't control. [10:12] (also, I can't join #ubuntu-virt, apparently it's on invite only?!) [10:12] lool: It's been discontinued. #ubuntu-virt, that is. [10:12] lool: It's been folded into this channel. [10:12] One could liken it to owning a house... so say you own a house where you live, plus you own a house in US, and you have your US house stock piled w/ a heap of 'stuff', if the 'stuff' in your US house is illegal under US law, you'd be in trouble, but as far as people breaking in, you'd be protected under US law [10:13] soren: Thanks; and any idea about the vm-builder question? :) [10:13] lool: vmbuilder doesn't install the server seed. [10:13] I would like to put RSA/DSA public keys into my ldap server, is there an easy way to do this with ubuntu? [10:13] twb: thats my take on it anyway [10:13] Perhaps where I said legality, I should've said trust [10:13] soren: BTW the other day on #openstack, I asked whether there is a new tool to create VMs for openstack; someone from rackspace at debconf told me that was the case which quite surprized me [10:14] lool: That's the first I've heard of it. [10:14] lool: I know of no such tool. [10:14] lool: eevans? [10:15] I'm not sure; sysadmin smoking a lot :-) [10:15] I mean cigarettes [10:15] twb: Can you restate your question with that modification? I can't work it out. [10:15] lool: I figured :) [10:15] Maybe "When migrating from a conventional server room to a cloud, how do you continue to avoid being taken for a ride by arbitrary governments and/or organizations around the world?" [10:15] lool, we all smoke a lot [10:15] very unhappy about Ubuntu lucid breaking under his feet because it was before the release [10:15] :)) [10:16] twb: answer, you can't :) [10:16] twb: The same way you do so for anything you outsource. [10:16] twb: you've just can to research and take your cloud where the laws are acceptable [10:17] (to you that is) [10:17] Again: I don't get this. [10:17] Being illegal doesn't actually stop stuff from happening. [10:17] G: so can I say to e.g. Amazon "please only host my data and images in the following countries" ? [10:18] twb: well w/ say Amazon you can choose which buckets you want your data hosted in [10:18] twb: The locations of their hosting facilities are well-known. [10:18] twb: Somewhat, at least. It's known which state they're in. [10:18] twb: but remember they are a US company, so they may have to obey US Search Warrants for international buckets anyway [10:18] twb: Same for Rackspace, btw. [10:19] twb: well not have to, but they may do it anyway, to prevent pressure in other places [10:20] I seriously doubt that if the FBI wanders into Amazon's data centre saying "we've got a warrant to search your facilities" that Amazon can say [10:20] "oh, no, you can't do that. there's stuff in there that belongs to people from other countries." [10:20] that's not how search warrants work, I think. [10:20] soren: sure, but what I'm talking about is buckets located in their EU zone etc [10:20] G: Ah. [10:21] soren: US Patriot Act apparently has some weird super-secret-warrant provisions [10:21] soren: as mentioned in: http://www.rsync.net/resources/notices/canary.txt [10:22] It's all pretty simple, really. Deal with the fact that the provider has access to whatever you give them access to. If that's a problem in itself, being illegal or their being accountable doesn't change anything in terms of being sure your data is safe. [10:24] An attacker just have to be slightly more cunning. There's cunning people everywhere. [10:24] yeah [10:24] Some people are happy if they can just encrypt their block devices. [10:24] That way, someone stealing the physical disk doesn't get anything. [10:25] ...but someone with access the hypervisor can do whatever they want. [10:51] hi, i'm having trouble updating php [10:51] i think i have the wrong sources in apt-get === _Techie_ is now known as _TechAway_ [11:41] New bug: #633054 in ipmitool (universe) "ipmitool TSOL buffer overflow" [Undecided,New] https://launchpad.net/bugs/633054 === dendrobates is now known as dendro-afk === freeflyi1g is now known as freeflying [13:15] morning [13:30] zul: Morning [13:46] New bug: #633129 in mysql-5.1 (main) "mysql-client-5.1 no longer uses readline :(" [Undecided,New] https://launchpad.net/bugs/633129 [13:51] New bug: #633138 in php5 (main) "package libapache2-mod-php5 5.3.2-1ubuntu4.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/633138 === rgreening_ is now known as ghost === ghost is now known as rgreening [15:02] New bug: #565238 in drbd8 (main) "drbdadm segfaults as non root user " [Wishlist,Opinion] https://launchpad.net/bugs/565238 [15:14] hello [15:14] i would like to export the X from a server tu use wireshark [15:14] i've installed xserver-utils [15:15] but it says me xhost unable to open display ":0.0" [15:15] which packages do I need on the server side to export X? [15:20] export X? [15:20] ttx: man bug 584048 is just a royal mess. I only just realized that two of the commenters were not the same people bc one is using libvirt and one is not. [15:20] Launchpad bug 584048 in qemu-kvm "kvm images losing connectivity w/bridged network" [High,Fix committed] https://launchpad.net/bugs/584048 [15:21] I'm going to have to force myself to be MUCH stricter about that [15:21] arh [15:21] (in my defense it dates back long before me, but that's no excuse) [15:21] hallyn: handling bug noise takes a bit of experience [15:21] ttx: so, i want to mark it as only valid for lucid. hints? [15:22] (actually, as 'incomplete' for lucid) [15:22] hallyn: you need to have maverick and lucid tasks open. You probably can't do that yourself. i'll fix it [15:22] then you can close the maverick task and set the lucid one to incomplete [15:23] * ttx approves lucid nomination [15:23] hallyn: now you have two tasks [15:24] hallyn: just mark the lucid one incomplete and the maverick one invalid, if that's what makes sense [15:26] ttx: thanks! [15:28] hallyn: no pb [15:28] I am using ubuntu 10.04 - server with / and /boot on an lvm2(on raid1( on /dev/sda & /dev/sdb) ). After I did a dist-upgrade from the fresh installation the kernel does not recognize my /dev/sdb (and thus the follow-up devices). Grub tells me "waiting for root device". Any ideas? [15:29] Afternoon all [15:30] Any one have any suggestions for mirrioring files in realtime ( or close ) ? [15:31] *inside openvz containers running ubuntu10.04 [15:41] I am trying to setup my system to run on NTP but the NTP is about 10+ minutes fast [15:41] run on ntp? [15:41] sorry? [15:41] ntp isn't fast, what ever server you connected to might be [15:41] what is your ntp config? [15:41] the ntpdate should auto sync it though [15:42] stock [15:42] is ntp running? [15:42] but I ran ntpdate ntp.ubuntu.org [15:42] yes [15:43] ntpdate doesn't do crap if ntp is running [15:43] elb0w: don't use ntpdate. that'll only synchronise once (or in certain intervals if run by cron). use ntpd or openntpd instead [15:43] looks like it just synced [15:43] odd [15:43] elb0w: ntpd will also take care of clock drift, which ntpdate can't [15:43] mathiaz: so AIUI it's still not possible to have personal blog posts auto-list under ubuntu server blog, right? [15:44] yeah its working now, thats weird [15:44] daxroc: local replication or over network? [15:44] ttx: have you got 10 to discuss bug 452900 and bug 452910? [15:44] Launchpad bug 452900 in asm2 "Please run testsuite during build" [Medium,Triaged] https://launchpad.net/bugs/452900 [15:44] Launchpad bug 452910 in asm2 "Encoding javac and javadoc warnings during build" [Low,In progress] https://launchpad.net/bugs/452910 [15:46] joschi: It's over local network ( gigE ) [15:47] daxroc: glusterfs would come to my mind. or drbd, but I'm not sure that'll work with openvz kernels [15:48] joschi: I've been looking at glusterfs but it seems to barf when run inside the containers. [15:48] shringo: sure [15:48] ttx: great - I've spent some time looking at the testsuite that ships with the asm2 source package. [15:50] Its not in great shape; it executes with a number of failed tests; it slow.. bit time - it uses the rt.jar as part of testing which has over 17k classes in it [15:50] and includes a number of performance comparison tests with bcel, serp and javassist; which would mean increasing the number of build dependencies where poss or removing this aspect of the testing from the source package. [15:51] All-in-all looks like quite a bit of work to get running it running well - and I don't think that it will come from the upstream project - last release to asm2 was in 2006! [15:51] from my /dev/sda, /dev/sdb, /dev/md0, /dev/mapper/lvm_foo* only the /dev/sda is listed. /dev/sdb is missing, thus the other layers cant start. After booting the kernel I always stick to busybox. What may be the issue. The revocery system from the cd does get the /dev/md0 after an additional check and the /dev/mapper/lvm* after a third one. Any ideas? [15:52] Advice on where to go next with this appreciated :-) [15:52] shringo: hmm [15:53] I've enountered quite a few of those half-broken testsuites [15:53] it depends how much are left when you disable the undesired ones [15:54] if you end up with 10% of the test suite, obviously it's not worth the trouble [15:54] if you end up with 80%, it's worth it... [15:54] how realistic is it to try to converge to using the same version of asm ? [15:55] we currently have in main 1.5.3, 2.2.3 and 3.2 [15:55] I had a look at the reverse dependencies (apt-rdepends generated some nice pictures for me) but I would need to assess the compatibility between asm3 -> 2. [15:56] I've not looked at 1.5.3 yet!!!! [15:56] the problem with asm is... usually it's embedded in code rather than just compiled against it [15:56] (upstream) [15:56] My personal take it that we could drop the performance tests - they are more of a 'look how great we perform compared to' type test. [15:57] so they jarjar to include it in a subpackage [15:57] in a different namespace [15:57] debian undid that [15:57] they usually patch so that it uses the system lib... and compile against it rather than include it [15:57] if I make sense [15:58] I currently am thinking about reinstalling if noone has a clue. [15:58] so you end up with a strong dep on a specific version of asm, and they seem to be widely incompatible with each other [15:58] It would be nice to converge on one version for platform; and then maybe take the jarjar approach for deviation from the platform version - I know this goes against policy but its got to be worth considering for this type of dependency [16:01] Anyway - that feels like a conversation for another day - how about this for an approach [16:01] a) assess how much of the test suite needs to be turned off to make it function [16:01] b) If acceptable > 80% then proceed [16:02] lets say 75% :) [16:02] c) Identify an appropriate build dependency (something simple with just a few hundred classes) to run the tests against - rather than rt.jar with its 17k classes [16:02] about the added build-deps: it's ok as long as they are in main too [16:03] Ok - i'll pick one from main - should not be to tough! [16:03] i'd not do a MIR (main inclusion request) just to enable 5% more of a testsuite [16:03] espcially a broken one [16:04] MIR's are presumably required for inclusion into Maverick from now until release candidate? [16:05] shringo: there are two theories. For one, MIRs fall under FeatureFreeze [16:13] ttx: which was back in early August? [16:15] https://wiki.ubuntu.com/MaverickReleaseSchedule [16:15] Aug 12th [16:15] Got that one; so it feels like this is not one for Maverick... [16:16] shringo: like I said, I'd not ask for an exception to enable part of a testsuite [16:17] ttx: fine - I guess the same applies for bug 452910 - it just general housekeeping [16:17] Launchpad bug 452910 in asm2 "Encoding javac and javadoc warnings during build" [Low,In progress] https://launchpad.net/bugs/452910 [16:17] yes, if it requires adding new packages to main, it probably should be postponed [16:20] ttx: OK - thanks for the advice - I will tidyup the work I have done so far and park it for the time being. === crazygir_ is now known as crazygir [16:23] shringo: you might consider joining #ubuntu-java. It's quite dead, theoretical topic is Java packaging in Ubuntu, but sometimes the right discussion ofr question wil happen there [16:28] I have a file in a directory with an 'unspeakable' filename. It appears to be a character encoding problem and I can't bring it up in bash completion. How can I delete such a file on the command line? [16:30] To answer my own question - probably delete the directory with the file in it. [16:46] Daviey, zul: can I use the UEC rig? [16:47] hggdh: i think its ok with me [16:47] thanks [16:47] might want to check with daviey as well though [16:50] I did, got no response [16:52] if I have a bunch of servers, and I want to run some command on them as root with parallel ssh, how can I accomplish this ? [16:53] hggdh, hey [16:53] hggdh: lemme know when you are done [16:54] something along the lines: I upload one bash script which does stuff to all of them with parallel scp, then I run that script (somehow), with sudo or how? I can't really enter the password when sudo asks me [16:54] Daviey, zul: it will be about 4 hours -- testing the euca 1.6.2 lucid proposed [16:54] anyone knows a smart way to trace use of memory per process during life time of kernel? [16:54] hggdh, that is awesome.... however [16:54] do you think you'd be able to test without cemedek? [16:54] cemc: example at http://www.linux.com/archive/feature/151340 [16:54] oh [16:54] cempedak ? [16:54] hggdh:ok [16:54] cempedak is already dead... [16:55] but I can free up one of the servers [16:55] hggdh, Oh.. no worries then :) [16:55] Daviey: sorry [16:55] Unless you don't *need* it? [16:55] no, I do not need *all* for volume testing [16:55] certainly not [16:56] hggdh: can you re-install them when you are done for me so I can work on what I was doing [16:56] hggdh, Ok.. which ones of these bad boys can we use? [16:56] zul, technically, we can re-provision them - but perhaps it would make sense for us not all to edit configs :) [16:57] so... you can use sapodilla & soncoya, I will restrcit myself to cemepdak, mabolo, marula, and santol [16:57] dammit, we are allowed the only ones i can't spell :) [16:57] heh...i need to eat [16:58] o/ [16:59] Daviey: I will unset the PXE boot, then [16:59] hggdh, are sapodilla & soncoya on? [16:59] Daviey: no, not right now. Before rebooting, I have to unset PXE [17:00] hggdh, mv $tftroot/$macaddress .tmp ? [17:00] yeah [17:02] cool [17:02] hggdh, Can you let me know when it's booting pleasE? [17:04] Daviey: actually, the way it is set up there: rm sapodilla && ln -s default sapodilla [17:04] hggdh, Hmm.. wouldn't the PXE client work down the list - reaching default eventually? [17:04] nijaba: I don't see any answers to my question there. I know the methods (pssh, clusterssh), I'm also using clusterssh at times, but I would like to be able to just run a command as root with pssh on the remote hosts. is there an easy way for that? [17:05] Daviey: it would, but this would mean that -- right now -- sapodilla and soncoya would end up with Lucid [17:05] cemc, That doesn't make much sense to me [17:05] ahhhh [17:05] hggdh, i see - i bow to your wisdom :) [17:05] :-) [17:06] Daviey: say I have a 100 hosts, and I want to remove package X from all of them (without any interactive stuff like clusterssh) [17:07] cemc: I'm afraid you will need to allow root access from pssh or allow passwordless sudo. none of them seem perfect. I would recommend using an ssh key in any case. [17:07] cemc, 'expect' or 'fab' (fabric) is what i would use :) [17:08] nijaba: mhm, I thought of these, but I hoped there was a better way ;) [17:08] Daviey: I'll take a look at those [17:08] cemc: not that I know of... [17:10] Daviey: sapodilla and soncoya are up. They should be in the same state you left them [17:12] hggdh, confirmed ; thanks [17:16] I installed the linux-virtual package on LTS 10.04 and now my server does not find its discs anymore. Can you help? [17:18] saulus: Why didn't you install linux-server ? [17:18] jpds: I did [17:19] Is it vserver compatible? I thought I get the vserver-compatibility with linux-virtual [17:25] saulus: did you uninstall -server? [17:25] never mind [17:32] how do i check for package that are installed on ubuntu server [17:33] dpkg --get-selections [17:35] patdk-wk: its not working i got error message [17:35] then dpkg is screwed on your system [17:35] it should list everything, with installed, deinstalled, ... [17:39] ivoks: I just tried. Now my initrd is missing. How can I get it back? [17:39] saulus: you tried what? [17:40] ivoks: i tried to remove linux-server [17:40] no! why?! [17:40] aeh, forget it. I ment I tried to remove linux-virtual. [17:41] I misread your suggestion above: 18:25 < ivoks> saulus: did you uninstall -server? [17:41] if you still have -server, reboot and boot -server kernel [17:42] I removed linux-virtual. Thus my kernel was gone. In /var/cache/apt/archieves I found the generic one and installed it. dpkg -i worked fine. But now my initrd.img-2.6.32... is missing. [17:42] sommer: hey are you around? I'm wondering what the best way to tell users of MySQL that they should always use InnoDB is.. I'm thinking maybe we should add it to the server guide. [17:43] they should always use innodb? [17:43] I almost never use innodb, too slow [17:43] You give up a lot of data security for that speed. [17:43] How do I get initrd.img, if it is missing? [17:43] ya, lucky I don't care about data security :) [17:44] saulus, update-initramfs -u [17:44] Also InnoDB can actually be a lot faster than MyISAM when tuned and given the proper resources. [17:44] spamapS, well, most of my use cases are logging applications [17:44] patdk-wk: if you never update your tables, then MyISAM is great. ;) [17:45] ran the logger into innodb, horrible, myisam, no issues [17:45] patdk-wk: did not do anything [17:45] patdk-wk: yeah, if you never update/delete .. myisam is awesome, and actually is far less likely to lose data on crash. [17:45] and yes, I do know how to tune myisam and innodb [17:45] ya, never update/delete [17:45] the applications that do, those are innodb [17:45] patdk-wk: did you try batching up the log entries into bigger transactions? [17:46] patdk-wk: the issue with innodb is it has to double-write on little transactions.. [17:46] SpamapS, actually, yes [17:46] the syslog logger, always did batching of like 50k rows [17:46] or 1second [17:46] the apache logger I modified to batch later on [17:46] patdk-wk: if you never delete/update a from a myisam table, it has a special structure that is basically like a text file.. so you're just appending data with no computation/seek time at all.. [17:47] yep [17:47] patdk-wk: so yeah, for logging, its pretty good. But if you crash during a write.. you just have to wait for myisamchk to finish. ;) [17:47] na, rotate it :) [17:48] patdk-wk: have you tried the archive engine? I've heard some people have better success with it because it does even less disk I/O since the data is compressed. [17:48] I rotate the logging db's daily, and compress/reindex them [17:48] no, I need indexs, and no index with archive [17:48] ah [17:49] patdk-wk: I think its worth it to suggest the *safer* option as the recommended option, and let people tune their way to MyISAM. ;) [17:50] would that include changing the mysql default engine to innodb? [17:51] just innodb tuning is so much harder than myisam [17:51] patdk-wk: opening a bug report to suggest that right now. :) [17:51] patdk-wk: Its not really.. increase buffer pool.. done. ;) [17:51] I would say, use innodb by default would be a good thing [17:51] I just wonder about user abilities for that case :( [17:52] * patdk-wk doesn't think, more ram == optimization :) [17:52] in that case, why use a harddrive? :) [17:52] saulus: update-initramfs -u -k all [17:52] patdk-wk: Drizzle, which is designed to be a *lightweight* database, has removed MyISAM completely. [17:53] patdk-wk: MyISAM only achieves its high insert speed by doing things in a dangerous way. ;) [17:53] ya, I know :) [17:54] patdk-wk: all the user tuning in the world is meaningless if their datacenter goes dark and they come back up without any data. [17:54] ty, ivoks. Finally I got it working. [17:54] hmm, I use federated tables a lot, love them :) guess not a drizzle thing :) [17:54] 4 hours of downtime restoring backups means all your tuning failed. ;) [17:54] patdk-wk: IIRC, federated was ported to drizzle [17:55] what I really want, and I just can't figure out, and just haven't bothered with a bug report yet on is [17:55] oh the suspense [17:55] I resolved the following reproduceable bug: If I install linux-vserver then my harddiscs are not recognized at boot. [17:55] I perfer to use percona's mysql build, instead of ubuntu's [17:55] Since I need the virtual package for linux-vserver I need to solve the bug! [17:56] hi guys i ahve install ubuntu server 10 version but i got porblem my application dont run on php 5.3 it only run on php 5.2 any help how to downgrade [17:56] but if I use anything in ubuntu that uses mysql, it wants to wipe them out, and reinstall ubuntu mysql server [17:56] (like perl dbi) [17:56] SpamapS, sorry, really really hungry :) [17:57] ruben23: that may be tricky! [17:58] patdk-wk: is that because the percona packages don't provide the proper package names? [17:58] SpamapS: what could be my options for this..? [17:59] ruben23: you could try forward-porting the php packages from 8.04... [17:59] ruben23: but thats not going to be easy... [18:00] ruben23: you could also compile a local copy of php just to get you by while you update your application [18:00] http://ubuntuforums.org/showthread.php?t=1459163 <----------------------i see a script here would this work..? [18:01] ruben23: you might also be able to do it with "apt pinning".. but I'm not really very familiar with that concept. [18:01] does ubuntu 9.04 still install php5.3..? or php 5.2..? === dendro-afk is now known as dendrobates [18:06] hallyn: nope - it's not possible to have personal blog auto-list under the ubuntu server blog [18:06] hallyn: the ubuntu server blog is a blog on wordpress.com - not a planet style [18:06] spamaps, I think [18:06] hallyn: an rss widget can be added to the blog to include your blog in the side bar [18:06] sorry, it's been a few months, so it's not fresh in my head the exact issue at the moment :( [18:07] maybe I need to file the bug with percona [18:07] hallyn: but full articles won't be published on the blog (and thus not show in the rss feed) [18:07] why I haven't filed with anyone yet [18:07] and haven't had time to dig into it [18:07] hallyn: what could be done though is to write up a copy-post script [18:08] ruben23: https://launchpad.net/ubuntu/+source/php5 shows the versions for each release [18:08] mathiaz: ok, thanks - just wanted to make sure that i remebered right [18:09] patdk-wk: the unholy marriage of libmysqlclient16 and mysqld is the problem [18:10] ya [18:10] I am not sure how to break it on ubuntu :) [18:10] I do on rhel all the time :) [18:10] New bug: #633364 in mysql-5.1 (main) "InnoDB should be the default table type" [Undecided,New] https://launchpad.net/bugs/633364 === dendrobates is now known as dendro-afk [18:47] mathiaz: can I add my blog to the side bars on ubuntu server too? [18:47] SpamapS: yes [18:47] SpamapS: go ahead [18:47] mathiaz: btw, was that overstepping bounds posting the gluster write up to it? [18:48] mathiaz: I figured since it was part of a blueprint it would be cool. :) [18:51] mathiaz: oh, and I don't think I have rights to add or configure widgets [18:57] SpamapS: what's your rss feed? [18:58] http://fewbar.com/feed/ [19:07] SpamapS: done [19:08] hmm, kernel test request [19:09] patdk-wk: kernel test? [19:09] kirkland: howdy!! I was wondering if you've used SystemTAp before to examine KVM related stuff in the kernel? [19:09] !bug 608429 [19:09] Launchpad bug 608429 in grub2 "black screen after grub menu" [Undecided,Fix released] https://launchpad.net/bugs/608429 [19:20] mathiaz: cool thanks! [19:21] mathiaz: when you have a moment, I want to talk to you about javascript access to the launchpad API. :) [19:26] smoser, Are you around? [19:26] smooooooooooooseeeeeeeeeeeerrrrrrr [19:26] well, of course i am. [19:26] whats up mr Daviey [19:34] * smoser twiddles thumbs [19:35] smoser, mumble? [19:35] * smoser mumbles bad things about Daviey under his breath [19:35] pah [19:43] hmm, horible launchpad caching :( === jsalisbury is now known as jsalisbury_afk [19:44] Everything after launchpad was redundant. [19:45] :( [19:46] is there a way to run some route up commands when network manager brings up an interfaces? [19:46] interface* === oubiwann is now known as ouibiwann === ouibiwann is now known as oubiwann-away === oubiwann-away is now known as oubiwann [20:00] cemc: /etc/network/if-up.d scripts will do what you need [20:01] cemc: man interfaces for more info [20:03] SpamapS: thx [20:12] mathiaz: just committed a new puppet section... won't be on doc.u.c until tomorrow though. === jsalisbury_afk is now known as jsalisbury [20:19] anyone have thoughts about Dell RAID controllers? i find myself in the unfortunate position of having to make a server purchase today, and I'm finding some reasonably priced refurbished Dells. Having dealt with FakeRAID issues in the past (with other vendors), I'm hesitant to pull the trigger too fast. [20:20] At minimum I need a RAID 1 mirror on real hardware RAID. [20:20] The cards I'm seeing are PERC 4, 5, and 6 [20:21] pittstains1: I prefer HP's servers [20:21] SpamapS: why? [20:21] pittstains1: but DELL is my 2nd choice. They always have good linux drivers. [20:21] pittstains1: Because HP's haven't ever failed me. :) (I should say compaq's .. proliant good, netserver, bead [20:22] s/bead/bad/ [20:22] i won't be buying directly from Dell, though, so I might just be getting the machine and no disks [20:22] Dell has all their drivers and such online. [20:22] SpamapS: Dell is much better than HP [20:23] ha, didn't mean to start a religious war, just want to make sure if i get a PERC 3 that I'll be able to do a true hardware RAID [20:23] pittstains1: as for the dell PowerEdge Raid Controllers. they are great [20:23] 3 dc? [20:23] or... [20:24] infact....waht model dell is that? [20:24] Dell PERC 3/Di Hardware RAID with 128MB Cache [20:24] mmmmm [20:24] PowerEdge 2650 [20:24] aetaric: I don't think I said HP is better [20:24] aetaric: I would never. They're way too complex to pass an ultimate judgement like that. [20:24] pittstains1: great server [20:25] aetaric: I just prefer HP. :) [20:25] SpamapS: never said you did pass judgement. i was just informing you that dells are better :p [20:25] I'm also looking at a Dell PowerEdge 840, which sports a Dell PERC 5I RAID Controller [20:25] but hp's are good too [20:26] i have a 6650 with a PERC 3/DC it's pretty awesome [20:26] aetaric: you're going to have to back that up with facts. ;) [20:26] i might just have to do it for the price, then [20:26] SpamapS: like the fact that every company i have ever worked for almost never had a bad server from dell [20:27] i just don't want to deal with fakeRAID again -- you know what i mean? [20:27] yeah.. fake raid is fail [20:27] aetaric: that is entirely subjective. Like my preference, its not really an objective case where you can say that they're better. You've just had more subjective experience, so you prefer them. [20:28] go raid 10 or raid 50..... if you get the perc 5 card and raid 5 if ou get the 3 [20:28] SpamapS: maybe. [20:29] i also prefer to have a dell employee replace parts on servers for me [20:29] i think i like RAID 1 because it's cheaper and faster [20:29] RAID 5 gets slow [20:29] and requires more than 2 disks :-) [20:30] thats sort of a false assumption [20:30] Anyone have any experience wrapping up local ubuntu server images to send to ec2? [20:30] the slow part. ;) [20:30] ah [20:30] i was going to ask [20:31] pittstains1: yeah you really can't notice a speed diff. [20:31] i think it's fairly well documented that RAID 1 is faster than RAID 5, but I should add that I'm really a programmer masquerading as a sysadmin [20:31] veenenen: many in here have, but you may want to check out #ubuntu-cloud too [20:31] and that i try to stay away from the hardware as much as possible [20:32] pittstains1: there are dedicated RAID5 controllers w/ battery backed write cache that will *destroy* RAID1 for multi-threaded I/O [20:32] the perc cards are really better for raid 5 [20:33] SpamapS: perhaps, but that doesn't sound cheap :-) [20:33] all 19 of my server's drives are in raid 5 arrays [20:33] Yeah, RAID50 with 14 disks kills RAID10 with 14 disks in HP's SmartARray cards, as long as you have battery backed write cache.;) [20:33] pittstains1: the 2650 is battery backed [20:34] heh that is the 2U sister system to my model === dendro-afk is now known as dendrobates [20:56] well, i did it. i bought the 2650. [20:56] thanks for your help, ladies and gents [20:56] * RoyK hands out Fedora CDs for the masochists === ivoks is now known as ivoks-afk [21:03] can anybody tell me how can I remove this link: http://mail.nwohiobb.com/squirrelmail/src/login.php just go back to mail.nwohiobb.com [21:05] Tim_R: are you running Apache? [21:05] yes I am SpampS [21:06] Tim_R: Redirect ... http://httpd.apache.org/docs/2.2/mod/mod_alias.html#redirect [21:07] probably something like Redirect permanent /squirrelmail/src/login.php / [21:08] You can probably just do Redirect permanent /squirrelmail/ / [21:08] if you're shutting down squirrel [21:08] yea I want to shut down squirrel [21:08] Tim_R: that goes in whatever config file you've used to configure mail.nwohiobb.com [21:09] * SpamapS must be going now [21:09] Tim_R: good luck === _TechAway_ is now known as _Techie_ === shennyg_ is now known as shennyg [21:19] mathiaz: any idea if LP: #423252 is going to be fixed soon? [21:20] zul: ^ [21:20] well that didnt work [21:20] any other ideas [21:21] mdeslaur: umm... [21:24] mathiaz: ^^^ [21:26] I'm not seeing a ton of response from the ubuntu-cloud chat room. Anyone here know how to get a kvm guest machine transfered onto ec2? === _Techie_ is now known as _TechAway_ [21:41] New bug: #633516 in unixodbc (main) "package odbcinst 2.2.11-21 failed to install/upgrade: " [Undecided,New] https://launchpad.net/bugs/633516 [21:47] New bug: #633521 in unixodbc (main) "package odbcinst1debian1 2.2.11-21 failed to install/upgrade: " [Undecided,New] https://launchpad.net/bugs/633521 [21:47] New bug: #633523 in unixodbc (main) "package unixodbc 2.2.11-21 failed to install/upgrade: " [Undecided,New] https://launchpad.net/bugs/633523 [21:50] i am trying to download vmware server and was wondering if there is a way to download it dirrectly from the command line? I am using ubuntu server 10.04 and cant seen to figure out how to make wget work. [21:53] clayd, the last time I tried, you had to sign up for an account and apply for keys first [21:53] then they gave you a download dlink [21:54] free to use but strings attached [21:54] thats what i thought but the link you get from them doesnt work. the bad part is that someone a while ago told me about a command line tool that allowed me to see a text version of the vmware site. yep i forgot the name of the tool [21:55] probably links or some variation thereof [21:55] such as elinks, links2, or lynx [21:56] i think it was links. that sounds right. [21:56] i will try it . thanks [21:57] yep that was it. thank a ton qman === Overand_ is now known as Overand [22:01] suggest elinks === SuperLag is now known as Guest72672 === _CaTeYe is now known as CaTeYe === _TechAway_ is now known as _Techie_ === funkyHat_ is now known as funkyHat === dendroba` is now known as dendro-afk [22:41] New bug: #633568 in dhcp3 (main) "dhclient3 not recognize -4 option sent by network manager" [Undecided,New] https://launchpad.net/bugs/633568 [22:47] <_Neytiri_> i am haveing issues connectiong to my mysql server remotely and mysql is correctly setup and so is the user i am trying to connect with === dendro-afk is now known as dendrobates [22:51] when you say mysql is correctly setup, i assume you mean you have added a user for the host you wish to connect from? [22:51] <_Neytiri_> yes [22:51] <_Neytiri_> i set it as a % host [22:51] did you do a FLUSH PRIVILEGES;? [22:52] <_Neytiri_> yes i believe so [22:52] <_Neytiri_> but i cant even connecto to mysql from the local machine usieng the addres of the interface card [22:52] check your my.conf to ensure the server is actually listening on the right IP/interface? [22:52] <_Neytiri_> that i didnt [22:52] by default it listens on the loopback interface only I believe [22:53] it's a security thing [22:53] <_Neytiri_> ok where do i find that? [22:53] /etc/mysql/my.conf? [22:54] <_Neytiri_> can i set mysql to listen on any address the machine gets? [22:55] you can set it to listen to a certain interface i believe yes [22:55] <_Neytiri_> i cant eneter a 0.0.0.0 or 255.255.255.255 [22:55] think it might say so in the comments for the config file... I thought you can just enter eth0 or whatever [22:57] <_Neytiri_> got it ty [22:57] hello all, I'm new to Ubuntu (and Linux for that matter). I'm trying to get a server running. I have (somewhat) successfully gotten Ubuntu 10.04 up and running, but the mouse isn't working. The PC has one PS/2 port that I have the keyboard hooked up to. I have the mouse hooked up to a USB port. Any suggestions (including RTFM is ok, but where). Thanks. [22:58] i assume you have a GUI installed then? [23:01] DrNick: yes. [23:02] I opened a terminal session, and executed "lsusb", the settings do not show a USB mouse is loaded. Only the default display. [23:03] do you have the desktop version of ubuntu installed? [23:04] DrNick: I downloaded the server edition, not the workstation. [23:05] DrNick: I was on the Ubuntu channel, and they thought that Ubuntu Server didn't have a GUI version. [23:08] so if you downloaded the server version was that what you installed? [23:08] DrNick: yes [23:11] soo... you don't have a GUI then. [23:12] DrNick: No, there is a GUI window. (I have issues with negative logic). [23:21] dollarbang1: you installed a GUI afterwards? [23:22] JanC: No, that was the default installation. [23:22] Ubuntu Server has no GUI by default [23:23] maybe you have Edubuntu? [23:23] Edubuntu terminal server has a GUI [23:23] anyway, lsusb should show your mouse if the _hardware_ detects it, maybe it's broken? [23:26] JanC: right now, I'm even questioning if I'm awake. I downloaded the 10.04 ISO from Ubuntu, used the "install from a Memory Stick" instead of creating a CD, the ran the installation from the USB Stick. Everything looked like it was working, I just never looked to see the mouse was having any problems, nor did I look at the installation log to see if it had problems. [23:28] JanC: Once it finished installing the software, I removed the memory stick, and rebooted the PC, the GUI loaded after I entered my password. === _Techie_ is now known as _TechAway_ === _TechAway_ is now known as _Techie_ [23:58] mdeslaur: re bug 423252 - I don't know [23:58] Launchpad bug 423252 in sudo "NSS using LDAP+SSL breaks setuid applications like su and sudo" [High,Confirmed] https://launchpad.net/bugs/423252 [23:58] mdeslaur: I haven't looked at the bug for some time now [23:58] SpamapS: what's up with javascript + LP API? [23:59] mathiaz: :(