[00:22] zoran119, no those patches got skipped [00:22] zoran119, safe, arguably [00:24] bogeyd6: it's ubuntu 8.04 lts... if i do dist-upgrade these packages should get upgraded... but will i still have 8.04 lts? or will be it someting else? [00:37] HI I am trying to join my PDC and i got this error: rpccli_netlogon_set_trust_password: unable to setup creds (NT_STATUS_ACCESS_DENIED)! any idea? [01:15] how i can install winbind? [01:16] i got 2 winbind on apt-get winbind4 and winbind [01:16] New bug: #521973 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/521973 [01:19] zoran119, still 8.04 [01:21] hi i just delete samba folder and unistall the application, when i try to create again it doesnt create the samba folder. any idea? [01:27] which is the difference between winbind and winbind4? [01:28] hi please i dont know what id wrong but samba doesn't install properly. [01:28] doesn't create the samba file [01:28] and tesparm doesn work === liva is now known as livret_a [02:19] Sigh. Did Canonical *really* need to change Ubuntu's etckeeper to default to bzr instead of git? [02:38] twb: makes sense [02:39] Only in the sense that canonical is in bed with bzr and everyone else got over it five years ago [02:41] twb: canonical in bed with bzr? the bzr developers are canonical employees [02:44] Bedding is a reflexive relation. [02:45] Er, not reflexive. [02:45] Symmetric? [02:45] Yeah. [02:54] twb: you're thinking too much [02:56] pmatulis: are you aware of any way to run UEC on a hardy server? [02:57] pmatulis: I have a hardy server that I depend on too much to upgrade to lucid at this point in the cycle, but I wants UEC :) [02:57] lifeless: sorry, i have almost no experience with UEC [02:58] kk, no worries [03:00] lifeless: depending on the dependencies, you might be able to just backport it [03:00] lifeless: I wouldn't do that with a production machine, of course [03:01] twb: UEC is eucalyptus - so it wants kvm :) [03:01] as well as a tonne of java [03:01] I was thinking about running it up in a lucid chroot, kindof thing. [03:02] It shouldn't need kvm specifically -- surely qemu or any other full virtualization solution would do [03:02] Java, otoh, can foad [03:03] I wouldn't want to try running KVM in a chroot, but if you have hardware VT, I can't think of any obvious reason why it wouldn't work. [03:08] hi somebody can help me i am not dure what i did wrong or what i am missing in my PDC config i am unable to get information with wbinfo -u my config is : http://pastebin.com/d4c68efa. Also i can join my own PDC [03:36] eucalyptus on 8.04? [03:38] http://open.eucalyptus.com/forum/eucalyptus-ubuntu-hardy-backport [03:38] you need at least two servers for uec [03:39] oh cool [03:39] bogeyd6: no, you don't. [03:40] oh, the link says 'we failed' :) [03:40] bogeyd6: recommended topology is multiple machines, but I have a test machine here running MANAGED_NOVLAN sc, cc and nc just fine. [03:40] I found some bugs in the setup scripts etc which I documented and filed, but it works just fine and reliablu. [03:57] Is it easy - or possible - to get dovecot to work like gmail, meaning by saving each email to a safe folder apon being downloaded by a client on pop? [03:59] IE. archive the servers copy of the email, while sending it to the client === dendro-afk is now known as dendrobates [06:20] CRAP! [06:21] I need my wiki; my wiki on my dev server; my server is being entirely rebuilt [06:31] And now I can't get my VM to grab an IP [06:32] I wonder if it's an issue with the bridged network... [06:34] hi after take off kerberos winbind doesn't work any more...wbinfo -u doen give me information. any idea about this prblem? [07:25] hi i cant find the package libnss_winbind.so in my server. ho i can install it? [07:39] Roxyhart0: about the worst time of week to find support... [07:40] yes, and just this weeks every thinks happen on the servers? this is my work week in the job history! [07:40] things [07:46] Roxyhart0: I wish I could help/// [07:47] If there's a S17mysql-ndb-mgm, S18mysql-ndb and S19mysql in /etc/rc3.d/, does this mean they are all started in this order? I'm wondering what the first two are, never heard of 'em. [07:47] thanks me too...anyway without problems it could be boring isn't? [07:48] Roxyhart0: true; issues keep your mind sharp :) [07:48] just like the issues I've been having [08:30] Roxyhart0: looks like it's in package winbind (/lib/libnss_winbind.so.2) [08:35] thanks ttx [08:35] i just recover it [08:38] i am trying to configure a PDC samba server , try to goin to the domain (which is in the same machine) but i cant see the Domain. say something that i dont have privileges, i am wondering if is something with winbind configuration / pam [08:38] ls [09:25] Oh, puppet. Why must thee spite me so? [09:29] hi all.. I have ubunut server 9.04 on a VPS, I installed the desktop pack on it, but now when I type STARTX I get this error log: http://paste.ubuntu.com/376753/ [09:37] KamaL: You won't be able to run standard X on a VPS. Standard X expects a monitor to be connected to the machine. Besides, this is off-topic here. Try in #ubuntu. [09:37] so i won't be able to connect through remote desktop to a ubuntu GUI this way? [09:38] hi, how i can reload my pam configuration without reboot the server? [09:38] KamaL: Try in #ubuntu. [09:38] no one is answering there :/ [09:39] KamaL: That does not magically make it on-topic for this channel :) [09:45] hi guys, i changed the /etc/network/interfaces file and lost ssh connection to my server [09:45] can someone advise why this happened? [09:45] i changed the IP and now, pings are not going through [09:58] soren: got a mystery task question on #ubuntu-devel... if you know what happens, let me know [09:59] someone know how i can reload my pam.d configuration without reboot the server? [10:06] Roxyhart0: for what? [10:06] ssh I think is parsed each time [10:06] for the gettys, you'll need to restart them [10:07] s/each time/on login/ [10:07] perhaps no restart at all [10:07] guys, any advise or reason why this happened? [10:08] i mean, i only changed the interfaces file [10:08] swift: and restarted networking_ [10:08] ? [10:08] no [10:08] i didnt do that [10:08] if you changed interfaces and restarted networking, your box is on a separate ip [10:09] if you didn't start networking, I don't know. [10:09] that file should only be read by /etc/init.d/networking (re)start [10:09] i didnt restart networking? [10:09] i didnt restart actually [10:10] yes [10:11] i just see that and my pam crash? somebody have any example of configure authentication with, samba, unic and ldap (pam.d) [10:19] swift: changing that file doesn't restart anything [10:19] swift: but rebooting after changing it, obviously does [10:19] i just changed the file and lost ssh access to the server [10:19] also, i can't ping to the new as well as onld IP [10:20] this is wierd [10:23] changing that file doesn't do anything [10:23] might have been something else [10:31] so, how often is ubu's keyserver reachable? [10:36] hi somebody have a example how to configura pam.d to access with samba and ldap? [11:18] hi do i need pam_unix2.so if i am susing pam_smbpass? or just pam_unix.so is ok? [11:43] the date is wrong on my system, how to reset it ? [11:44] i mean system time [11:46] someone have some example how to configurate pam.d with samba, ldap, unix sincronization? [11:46] :~# date [11:46] Sun Jun 26 00:02:52 CDT 1988 [11:51] maxagaz : date --help [11:51] check for ntpdate to keep the system synchronized [12:07] Hello, someone has an idea why my mouse in karic is moving from alone to the top of my screen? [12:07] liva, ntpdate ntp.ubuntu.com [12:11] You can have various servers in fact [12:11] It depends where you live [12:11] editing /etc/cron.daily/ntpdate [12:44] Since the new tftpd-hpa upload, where it now runs as a daemon as opposed to running from inetd.conf, it no longer autostarts when my system boots. [12:44] I need to manually run it. Any clues on how to make it autostart? [12:45] when i send (external) emails postfix does not at all respect the hostname set in main.cf .. this leads to the sender being like .. how can i set these mappings? [12:53] Does anyone know why libaprutil1-dev is uninstallable on amd64? [12:53] the anjuta package FTBFS because of that lib http://launchpadlibrarian.net/39209795/buildlog_ubuntu-lucid-amd64.anjuta_2:2.29.90.0-0ubuntu2_FAILEDTOBUILD.txt.gz [13:05] Nggghghhh!! [13:05] Is anyone using puppet in Lucid? [13:14] hello [13:14] I have trouble using kickstart to configure a NIS client [13:16] I update /etc/passwd, group & shadow during post. [13:16] After the install, I am able to log as a NIS user [13:16] but not with root [13:21] if I execute the modification after the install, by myself. It worksss [13:32] fhermeni: what is kickstart? [13:32] a tool for fully automated installation [13:32] fhermeni: not the redhat tool? [13:32] yes [13:32] but usuable in ubuntu too [13:33] You can use that on Ubuntu? [13:33] !info kickstart [13:33] Package kickstart does not exist in karmic [13:34] https://help.ubuntu.com/9.10/installation-guide/amd64/automatic-install.html [13:34] use preseed [13:35] fhermeni: interesting - didn't know that. Perhaps "preliminary" means "not all working" :-) [13:35] yeah I suppose [13:35] but I was never able to boot using the configuration file of preseed [13:39] I've just tried passing a preseed file in the kernel but I have to screen to choose the language [13:43] so anyone already tryed preseed ? [13:47] in fact, the installer do not try to get the file (no log in apache access) [13:48] <_ruben> fhermeni: preseeding happens after locale/keyboard selection and network configuration (which can be configured using kickstart (never tried) or boot params) [13:51] ok, seems effective, still have the two first points (locale/keyboard) to manage which is not really possible [13:51] <_ruben> debian-installer/locale=en_US console-setup/layoutcode=us [13:52] ok thanks [13:52] <_ruben> add those as boot parameters (adjust if needed) [13:52] <_ruben> network shouldnt ask any questions if only one nic is available [13:52] <_ruben> and will try dhcp on it [13:53] yep, but it asks for a confirmation about the hostname [13:58] why does libaprutil1-dev depend on the old libmysqlclient16-dev? [14:00] Can rsync connect to a ftp server to pull the files to sync it locally or just through SSH? [14:00] _ruben: ok, seems to works. thanks for the tips [14:02] What do I need to do to make vmbuilder --add-pkg sun-java6-jdk work (sun-java6-jdk is in multiverse) [14:03] by [14:03] Should I create a boot script which adds multiverse to /etc/apt/sources.list and apt-get update/install from there? === Guest50488 is now known as JordanCook [14:12] New bug: #377474 in bridge-utils (main) "bridge-utils ifupdown script should disable IPv6 address autoconfiguration for underlying interface" [Wishlist,Triaged] https://launchpad.net/bugs/377474 [14:12] uvirtbot`: nick uvirtbot [14:12] soren: Error: You don't have the admin capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified. === uvirtbot` is now known as uvirtbot === dendrobates is now known as dendro-afk [14:19] the libaprutil1-dev issue mentioned above breaks up the hole anjuta package === Guest50488 is now known as JordanCook [14:25] i am getting this error in my mail.log.. "error: to submit mail, use the Postfix sendmail command" and "fatal: the postfix command is reserved for the superuser" [14:28] <_ruben> JordanCook: obviously something is trying to send mail in an inappropriate manner [14:29] ok [14:42] ttx: How much have you thought about the etckeeper-puppet integration thing? I have a few things I'm not sure how are supposed to work.. === dendro-afk is now known as dendrobates [14:43] soren: it's more a spec from Mathias, but I should be able to answer [14:43] ttx: For one, the spec says that it should support two modes: a) overwrite local changes, or b) abort on local changes and output the diff. [14:43] In b), what does "output" mean? Where should I put it? [14:43] (Also, the diff etckeeper subcommand is listed as a "Later" work item, but I needed it for this, so I went ahead and implemented it. [14:44] ) [14:44] Wow. I almost forgot the closing parenthesis. [14:45] * soren suddenly thinks of all the parentheses people have opened on IRC, but never closed. [14:45] The parts of my brain that speak LISP are twitching. [14:45] soren: i'm not sure what Mathias meant exactly ) [14:45] soren: since I agree "output" doesn't make a lot of sense in that context [14:46] soren: better ask him, should ba available in a few [14:51] hi, I got a problem similar to this hi i got a problem similar to this one http://www.mail-archive.com/samba@lists.samba.org/msg102818.html and i have samba 3.4.0. . It say is improved in samba 3.4.1. somebody know about that? how i can migrate to samba 3.4.1? [14:57] Roxyhart0: You can download Samba source and compile. The link you sent is referencing a fedora/RH Samba package - which may be different to the "official" 3.4.1 however (additonional patches). You would need to check. [14:58] i am wonderin if there are Ubuntu version [14:59] sombody has change or update to samba 3.4.x? x >0 [15:03] ttx: Alright, I'll talk to mathiaz. Thanks :) [15:05] Roxyhart0: http://packages.ubuntu.com/lucid/samba [15:07] i got karmik it is ok if i install it on 9.10? [15:09] Roxyhart0: Probably not - dependencies. You might break your system if you force it. [15:10] Roxyhart0: But I don't know. [15:10] :S so what i can do i need solve this problem [15:10] just move to lucid? === dendrobates is now known as dendro-afk [15:53] I got issue, I did install Open LDAP server according to this and I did not get promt for password, is there any way to recover it ? [15:53] https://help.ubuntu.com/9.10/serverguide/C/openldap-server.html [15:53] no, there's no password [15:54] if you are root on local machine [15:54] so I can still set it later on ? [15:54] you should be able to create DIT and work with it [15:54] invoks does SSH connection count as "root on local" ? [15:54] if you can do sudo, then yes [15:55] ok [15:55] this document is buggy [15:55] instead of -x -D cn=admin,cn=config -W [15:55] you should be able to use -Y EXTERNAL [15:56] do I need to define something for ldap server in /etc/hosts ? [15:56] so... ldapsearch -Y EXTERNAL -H ldapi:/// [15:56] nope [15:56] New bug: #522185 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/522185 [15:59] I get this error ldap_bind: Server is unwilling to perform (53) [15:59] additional info: unauthenticated bind (DN with no password) disallowed [15:59] when I did try to run [15:59] ldapsearch -xLLL -b cn=config -D cn=admin,cn=config -W olcDatabase={1}hdb [15:59] have you not read what i said? [15:59] instead of -x -D cn=admin,cn=config -W [15:59] you should be able to use -Y EXTERNAL === dendro-afk is now known as dendrobates [15:59] so... ldapsearch -Y EXTERNAL -H ldapi:/// [16:00] ldapsearch -Y EXTERNAL -H ldapi:/// olcDatabase={1}hdb [16:01] ivoks, in the Cluster Testing, how come there's no openais configuration? [16:01] i mean [16:01] corosync [16:02] or the default config and replacing the bindnetaddr is the only things that needs to be configured [16:03] RoAkSoAx: that's the only thing that's needed [16:04] gy, i changed the IP iin the interfaces file of my ubuntu-server, restarted networking service and ifconfig shows the new IP [16:04] but, if i try pinging to the IP, it doesnt ping [16:04] also, i cannot ssh to the new IP [16:04] please advise how i can make this work? [16:05] restart sshd [16:05] but why can't i ping to the new IP? [16:06] is restarting sshd mandatory? [16:06] firewall? [16:06] what that error I am getting is basically meaning ? [16:06] no, sshd should restart by it self [16:06] furythor: i told you 2 times to use -Y EXTERNAL, cause that doc is outdated and buggy for ubuntu 9.10 [16:07] no, i'm trying to ping the IP from within the network [16:07] i can ping other servers on the same network [16:07] swift: wrong gateway? [16:07] check it out with route -n [16:08] ok [16:08] ivoks: is there some better guide or should I just dump open LDAP for while ? [16:09] the gateway is correct [16:09] the server was getting a dhcp IP [16:09] i changed it to static in the interfaces file [16:09] and gave it a new IP [16:09] Say, anyone here got experience with setting up irssi in screen ? [16:10] also, I uninstalled network-manager [16:11] swift: paste your /etc/network/interfaces [16:12] swift: on pastebin [16:12] furythor: the guid is ok, except the auth part [16:13] sommer: here? [16:13] ok [16:13] what's the url for development version of help.ubuntu.com? [16:13] is there some way to change admin password ? [16:13] http://pastebin.com/m66d0d335 [16:13] thats the portion i added [16:14] ivoks:http://pastebin.com/m66d0d335 [16:14] swift: you are missing gateway there [16:14] i need to specify one in the interfaces file? [16:14] swift: yes [16:14] otherwise, routes won't be added [16:15] furythor: yes, but why would you do that? [16:15] furythor: do you understand how ldaps work? [16:15] oh! [16:15] ivoks honestly said : NO [16:15] furythor: there's config DIT [16:15] furythor: and there's your_data DIT [16:16] okay [16:16] furythor: config DIT doesn't need admin account, unless you would like to change configs from other machines [16:16] your_data DIT can be populated with whatever you want [16:16] ok [16:16] now, check this out: [16:16] http://www.howtoforge.com/install-and-configure-openldap-on-ubuntu-karmic-koala [16:17] this guide creates dc=example,dc=com (that's your_data DIT) [16:17] Can I use OpenLDAP to define access control to various directories on my server ? [16:17] directories? [16:17] like folders on disk? [16:17] yeah [16:18] thanks ivoks!!!!! [16:18] u rock!!! [16:18] it worked! [16:18] what's wrong with doing it with decades old chmod? [16:18] swift: great [16:18] RoAkSoAx: how's it going? [16:19] because I am building setup like this, I got my home server which I will use on multiple purposes, some include like designing websites for clients, and others get confidential data on there, so I want my server to be secure before I introduce it more wider population [16:19] you don't need ldap for that [16:19] mkdir /tmp/my_private_dir [16:19] chmod 700 /tmp/my_private_dir [16:19] done [16:19] only you can access it [16:21] well I'd wish it would be THAT simple however it is not, my intention is to offer files via webserver, so visitors need to be authenticated before getting access to files [16:21] so? [16:21] .htaccess? [16:21] and password through it ? [16:21] htpasswd [16:22] htdigest [16:22] openldap is overkill for what you want [16:22] yeah that would be one solution too... well I don't need server quite yet... so I will go through options and see what I choose [16:22] but thanks for the help [16:23] eh [16:25] RoAkSoAx: ? [16:30] ivoks, yers? [16:30] RoAkSoAx: i tought you were testing [16:31] ivoks, yeah i am but trying to configure a loadbalancer with pacemaker and ldirectord , but went to take a shower first :P [16:31] hehe ok [16:31] why ldirectord? === dendrobates is now known as dendro-afk [16:32] haven't we decide to use keepalived? [16:32] ivoks, yes, I've already done simple config with keepalived and it works, not I wanna test pacemaker/ldirectord/ipvsadm and do some comparisons with failover [16:33] s/not/now [16:33] ok [16:33] like how much time each takes to failover [16:34] i'll do a call for testing by the end of this week [16:34] ok [16:34] feature freeze is very near [16:37] New bug: #521231 in nagios3 (main) "package nagios3-cgi (not installed) failed to install/upgrade: trying to overwrite '/usr/lib/cgi-bin/nagios3/avail.cgi', which is also in package nagios3 0:3.0.6-5ubuntu3" [High,Triaged] https://launchpad.net/bugs/521231 [16:37] New bug: #522196 in postfix (main) "package postfix 2.6.5-3 failed to install/upgrade: subprocess new pre-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/522196 [16:37] ivoks, yeah well i rather have keepalived 1.1.17 instead of 1.1.19 since debian maintainer does not want that version either due to it introduces several bugs [16:38] so I rather have an stable version in lucid [16:38] then fix those bugs [16:39] ivoks, yeah Im tracking down the devel list, there's a patch for a bug there but I've not have the time to test it yet, I wanna do that tonight [16:48] argh, UEC needs images, damn them, images!!!!!!!!!!!!! [16:53] hey guys, i understand that its impossible to update the kernel on a ubuntu ec2 instance. so i was wondering if it ever makes sense to apply kernel updates. my guess is no, and i should have apt ignore all kernel updates. anyone have any thoughts on this? [16:56] New bug: #521312 in geronimo-j2ee-connector-1.5-spec (main) "Sync geronimo-j2ee-connector-1.5-spec 2.0.0-1 (main) from Debian testing (main)" [Undecided,Fix released] https://launchpad.net/bugs/521312 === Guest50488 is now known as JordanCook [17:00] Looking to do some traffic shaping, to ensure that my ssh terminal doesn't die thanks to torrenting/ftp server use. What should I be looking into? [17:08] Why is libsvn-dev uninstallable? This causes the anjuta package to FTBFS: http://launchpadlibrarian.net/39209795/buildlog_ubuntu-lucid-amd64.anjuta_2:2.29.90.0-0ubuntu2_FAILEDTOBUILD.txt.gz This is only an amd64 issue, it installs fine on i386. [17:08] I think the issue is caused by libaprutil-dev [17:18] hi, anybody know any web server software that can be launched from cli and instantly server curent pwd as docroot? [17:19] only static files support [17:22] RoAkSoAx: please help with MIRs [17:22] When you get a ssh "connection timed out" is that the server or client giving that? [17:22] client cannot connect to server === dendro-afk is now known as dendrobates [17:22] I have this server I try to connect too that is just slow as hell for me to get to, I always get timeout's [17:23] ivoks, anyway to increase how long the client will try before it gives up? [17:24] kpettit: time out means no network connectivity [17:24] there is nothing on that IP [17:24] mirs are killing me [17:25] ughh. it's anoying becuase sometimes it works, i just seem to be at the tail end of that time out and simetimes I can get thorugh but mostly I can't [17:25] I know it's up becuase it's web services and such are running, ughhh. I think it's the dumb linksys router that's not routing correctly or taking to long [17:27] stas: iirc there's a python one-liner that does that. [17:28] kriel: thx, just tried python -m SimpleHTTPServer but it lacks some support [17:28] i cant get everything loaded into webpage dunno why [17:30] i thinks it has poor mime support [17:31] kriel: nvm, it worked, ty. I misspelled my docroot === liva is now known as livret_a [17:35] postfix seems to be going in a loop [17:36] PS: <3 wondershaper [17:36] it keeps relaying the mail localy === luis__lopez is now known as luis_lopez [17:54] How can I tell apt to ignore an unmet dependency ? [18:00] ivoks, ok will also look into it tonight [18:03] please [18:03] we need to finish them asap [18:03] you could just use one of MIRs as template for new requests [18:07] ivoks, yep im already reading the other filed bugs [18:12] RoAkSoAx: go from bottom to top [18:12] ivoks, ;) [18:12] i'm to tired to continue today [18:12] Hum; After rebooting a Karmic server with some (not root) LVM logical volumes on clustered storage, the server will drop to busybox on boot because it does not find liblvm2clusterlock.so; It does *not* fall back to local locking, even if this is enabled in lvm.conf .. [18:13] Thus not starting the local root logical volumes, and not booting. [18:13] doh... that's cman [18:14] baffle: is that /home on clvm? [18:15] ivoks: No, the CLVM-volumes are virtual machine LVs. [18:15] ivoks: hi [18:15] ivoks: I don't think MIR needs to be written before FF [18:15] mathiaz: hi [18:15] ivoks: the most of important thing wrt to FF to have all the new packages in the archive by FF [18:15] mathiaz: problem is that if we don't get MIR before FF, we'll need to get features after FF [18:15] ivoks: But this is in the initramfs; Since lvm.conf specifies "locking_type = 2" it tries to use the locking library wich is not included in the initramfs. [18:15] mathiaz: that's the thing [18:16] ivoks: And instead of gracefully falling back to local locking, it dies instead. [18:16] ivoks: Even if this is specified in the configfile. [18:16] mathiaz: we can't have new rhcs before pacemaker is in [18:16] ivoks: because rhcs (in main) needs to be built with new libraries (currently in universe)? [18:16] mathiaz: yes [18:16] ivoks: another way around is to demote rhcs to universe for now [18:16] ivoks: If I change /etc/lvm/lvm.conf to "locking_type = 1" while in initramfs, I can active the root/swap LVMs. [18:16] ivoks: and do the promotion to main after [18:17] ivoks: if rhcs was already in main, promoting it is faster [18:17] mathiaz: same goes for lvm2 and ocfs2-tools [18:17] baffle: eh, i haven't worked with clvm and cman yet [18:17] ivoks: lvm2? [18:18] baffle: i just started testing it with pacemaker [18:18] mathiaz: yes, clvm [18:18] mathiaz: lvm is build against libcman [18:18] mathiaz: if we demote rhcs, we will have to demote lvm too :D [18:18] mathiaz: and to avoid all that, it's just easier to finish MIRs before FF [18:19] ivoks: lvm like in folder-hook imaps://mail.canonical.com/ 'set signature=~/.mutt-ubuntu-signature' [18:19] ivoks: mh [18:19] ? [18:19] ivoks: lvm like in Linux Logical Volume Manager? [18:19] mathiaz: yes [18:20] Depends: libc6 (>= 2.4), libcman3, libdevmapper1.02.1 (>= 2:1.02.24), libdlm3, libreadline5 (>= 5.2), lvm2 (= 2.02.39-0ubuntu11) [18:20] that's clvm [18:20] Source: lvm2 [18:20] ivoks: right [18:20] we could drop clvm for alpha3 :) [18:21] but it is just easier to finish MIRs [18:21] and be done with it all [18:21] ivoks: ok [18:21] there's not that many of them [18:21] https://wiki.ubuntu.com/ClusterStack/MIR [18:21] i just need one more pair of hands [18:24] can anyone comment on which version would be better to run a xen host: jaunty because it's newer, or hardy LTS? [18:26] that's it for me for now [18:26] i'll be back later [18:26] bye [18:28] panfist: how linux-savvy is the person that's going to be maintaining this server? [18:29] i would say i'm intermediate [18:30] i know some scripting and i've experimented with debian, suse and solaris [18:30] You'd probably be fine with jaunty, and maybe get some cool new optimizations out of it. LTS will be 'more stable' by definition, but I'd think jaunty would be more than stable enough. [18:30] ok [18:32] kriel: If you're going with a non-LTS release, why not just go for Karmic? [18:33] baffle: panfist didn't ask for karmic. He asked choose([jaunty,hardy LTS]) [18:35] kriel: Yes..? [18:35] baffle: If you'd like to ask that question, ask it of panfist. I answered the question that he posed. [18:35] kriel: Oh, sorry, didn't notice it before now. :-) [18:36] panfist: If you're going with a non-LTS release, why not just go for Karmic? [18:38] i heard that there is no xen patched kernel for karmic yet [18:38] unless that's changed since november [18:39] panfist: Ah, that might be true. [18:39] panfist: Didn't think of that; I've moved over to KVM now. [18:40] i need a solaris guest, does kvm support that? [18:40] i have a zfs thing going on, and no spare hard drives to dump the data and rebuild a linux raid. besides, there is a reason i switched from linux raid to zfs in the first place... the reason i want to run the zfs of a guest is because my video card isn't supported in solaris [18:41] so i have two options, buy a new video card or motherboard, or virtualize [18:41] or buy 8 TB storage to rebuild the array [18:41] that's right out, for being too expensive [18:42] i figure i will try the virtualization thing, because it's free, and if that doesn't work out then get an nvidia graphics card [18:44] panfist: Seems like KVM supports solaris: http://www.linux-kvm.org/page/Guest_Support_Status [18:44] fully or paravirtualized? [18:45] i was going to set up the solaris vm as a paravirtualized guest and hopefully get close-to-bare-metal performance of the zpool [18:45] panfist: Using HVM, not sure if there are paravirtualized i/o yet. [18:46] ummm do i have my terminology wrong [18:47] is paravirtualization what i want? [18:48] panfist: Seems XEN might be the best way to go; I don't think there is support for paravirtualized i/o (virtio) in Solaris yet. So you will not get close to bare-metal performance. [18:48] well, since you seem to be experienced with KVM, maybe you can let me know, what kind of performance would i expect managing the raid from a guest os? [18:48] ok === tbc1 is now known as tbc [19:19] mathiaz: I have a few questions about the puppet-etckeeper integration. [19:19] soren: hi! [19:20] mathiaz: Apologies in advance if I go missing all of a sudden; Sofie's ill so she demands a lot of attention. [19:20] mathiaz: Oh, right: Hi :) [19:20] soren: no problem - a human being is more important than a puppet and a keeper :) [19:20] :) [19:21] mathiaz: The spec says that it should support two modes: a) overwrite local changes, or b) abort on local changes and o [19:21] utput the diff. [19:21] whoops. [19:21] Meh, you can join lines. [19:21] In b) where should this output go? [19:22] The diff etckeeper subcommand is listed as a "Later" work item, but this output depends on it, so I'm curious if it was deferred? [19:22] (I did actually implement that diff subcommand, so if we can think of a place to put the output, that's cool). [19:23] soren: hm - good question [19:23] soren: This is hasn't been really discussed [19:23] mathiaz: Alright. [19:23] soren: so - for the spec and lucid alpha3 it's not important [19:24] soren: for alpha3, the goal is to have a similar experience to what apt/etckeeper does for now [19:24] mathiaz: Ok. [19:24] soren: ie: run etckeeper before modifying any files, run etckeeper after puppet has modified all the files [19:25] mathiaz: Have you thought about where to put this code? Should it be in a new puppet-etckeeper package that attempts to do the right thing if installed, or should we make etckeeper a dependency of puppet and stick the hooks in there by default? [19:25] soren: I wouldn't make etckeeper a dependency [19:25] mathiaz: Ok. [19:26] soren: if etckeeper is installed, puppet should automatically integrate with it [19:26] Ok. So I could put the integration code in puppet and have the hook scripts check whether etckeeper is installed, and if not, just bail out. [19:26] soren: sounds like a good plan to me [19:26] Alright. [19:27] One more thing: [19:27] soren: and have puppet *suggest* etckeeper [19:27] Not even recommend? Ok. Your call :) [19:27] The other thing: "run etckeeper before modifying any files"... What to do there? Commit local changes? [19:27] soren: yes [19:27] soren: that's what apt does [19:28] mathiaz: Oh, it does it before doing anything as well? [19:28] I didn't realise. Cool. [19:29] soren: yes - http://paste.ubuntu.com/377070/ [19:29] Wicked. [19:29] Dum, da dum.. [19:30] Anyone care to help me look at a LVM/clvm/initramfs problem? [19:31] baffle: Feel free to shoot. If I can be of help, I'll pitch in. [19:31] * soren is not really clvm savvy, but still.. [19:32] mathiaz: I will consider failure of etckeeper non-fatal for now. [19:33] * mathiaz nods [19:33] soren: I have root and swap on a local pv/vg. I also have some virtual machine logical volumes on a "remote" pv/vg (SAN). This vg is clustered. [19:34] soren: This means I have specified locking_type = 2 in /etc/lvm/lvm.conf, so that clustering on these volumes work. [19:34] baffle: ok. [19:34] soren: I've also enabled fallback to local clustering and local locking (for local volumes only) in lvm.conf. [19:35] soren: When creating the initramfs (update-intramfs -u) it seems liblvm2clusterlock.so does not get copied to /lib/lvm2/ in the initramfs-image. [19:36] soren: Ofcourse, liblvm2clusterlock.so would not *work* during the initramfs boot, as the cluster is not running. [19:36] soren: But lvm now seems to error because liblvm2clusterlock.so does not exist, and exists instead of falling back. [19:37] soren: (During the boot phase) [19:37] soren: So the local (non-clustered) root-filsystems doesn't start. [19:38] soren: To boot the server, I have to edit lvm.conf in the initramfs manually and change it to a non-clustered config and manually activate the logical volumes (lvm lvchange -a y /dev/blah/root & swap) [19:39] baffle: Ok. [19:39] soren: So, I'm not really sure if this is a) A missing hook for copying liblvm2clusterlock.so into the initrd, or b) An error in lvm2 that doesn't handle missing dependencies. [19:39] soren: Or c) something totally unrelated. :) [19:39] baffle: Have you tried copying liblvm2clusterlock.so in there to see the effects? [19:39] soren: But basically the result is that the combination of root on LVM and clustered LVM breaks totally. [19:39] soren: No, is there an easy way to do that? [19:40] baffle: How did you edit the lvm.conf in initramfs? [19:40] Oh, from the initramfs prompt? [19:40] soren: Yes, from the initramfs busybox shell. :) [19:40] Ok. [19:41] soren: sed < /etc/lvm/lvm.conf 's:locking_type = 2:locking_type = 1:' > /etc/lvm/meh ; mv /etc/lvm/meh /etc/lvm/lvm.conf [19:41] Yay. :) [19:41] In /usr/share/initramfs-tools/hooks/lvm2, you see the code that copies the lvm2 related stuff into initramfs. [19:42] copy_exec is a function that's supposed to copy the binary itself and all its dependencies. [19:42] However, I imagine the clvm stuff is dl_open'ed at runtime rather than linked. [19:42] ...which would explain our problem. [19:42] s/our/your/, I suppose :) [19:43] So, if that will fix it, the solution will be to add a hook to the clvm package that copies the relevant libraries. [19:44] soren: Yes, it is dl_open'ed. [19:44] soren: Updating and booting the server now. [19:45] Well, there we go. That's why copy_exec doesn't to the trick. [19:45] It uses ldd to figure out what else to copy. [19:45] IIRC anyway. [19:45] Something to that effect. [19:46] soren: It takes a little while to boot, the servers are slow to boot. [19:46] Wow, that was redundant. [19:47] Fresh off the press from the department of redundancy department. [19:50] soren: Hmm, seems that didn't fix it. It doesn't complain about the missing library anymore. [19:52] baffle: Dang. [19:52] Too easy. [19:56] soren: I hoped it was something that easy. Bah. [19:56] useless redundancy: atm machine; pin number; this slow server is slow [19:57] MTecknology: :) [19:59] baffle: How did you copy in liblvm2clusterlock.so ? [19:59] baffle: Using copy_exec? [20:00] soren: No, using mkdir $dest/lib/lvm2 & cp -p /lib/lvm2/blah.so $dest/lib/lvm2. So that worked. It didn't output an error about that. [20:01] soren: But I just tought of something, maybe it will fix it. [20:04] soren: Just noticed that fallback_to_clustered_locking = 1 (as well as fallback_to_local_locking = 1) .. But even after changing to "fallback_to_local_locking = 0" pvscan etc fails. [20:04] baffle: No errors? [20:05] soren: Yes, 1 sec. [20:07] soren: http://i.imgur.com/Kjdf7.jpg [20:07] baffle: Oh, right clvmd. [20:07] soren: Yes, that is the locking library. :) [20:08] soren: But this has worked before I think? I mean, people *have* booted their clustered machines before. :-) [20:08] So they say. [20:08] :) [20:09] Is this Lucid? [20:09] soren: Maybe everyone uses ksplice. [20:09] soren: It is karmic. [20:09] * soren ponders [20:11] baffle: Bah. Someone who's actually ever used clvm should probably be able to say something remotely informed about this. Incidentally, I've never used it. [20:20] soren: Hmm, it just seems fallback doesn't work at all, even if I try to set locking_type to 3 (clustered locking, but internal instead of external) it still doesn't fallback to local.. [20:24] Hmm, I should check how it works on another release. [20:48] Folks: synaptic has a cool feature for choosing the fastest ubuntu apt repository. [20:48] But I want to do that for a server to which I have only ssh access -- no X. [20:48] Is there a textual, command-line tool that does the same thing? [20:49] I know that Debian has two -- netselect and apt-spy -- but Ubuntu seems to have neither. [20:51] zooko: I never thoguht about it, but that would be nice [20:52] mathiaz: lp:~soren/ubuntu/lucid/puppet/etckeeper-integration for your reviewing pleasure. [20:53] mathiaz: Also, I'm feeling quite under the weather. I may "call in" sick tomorrow, so if I'm not back yet to upload before FF, could you take care of that? [20:53] soren: sure [20:54] It's amazing, really. Going to all these UDS's, acquiring all sorts of weird viruses from all over the world, training up my immune defence... and my daughter brings home something from day care, and I'm out for days. :( [20:55] soren: I'm sure you'll survive ;) [20:55] soren: your branch looks god [20:55] soren: *good* [20:56] soren: I'll probably do another upload of puppet as I've run into an issue with puppetmaster [20:56] mathiaz: Do share. [20:57] soren: debian bug 570012 [20:57] Debian bug 570012 in puppetmaster "puppetmaster fails to start: puppet user missing" [Normal,Open] http://bugs.debian.org/570012 [20:58] Oh. [20:58] Yeah, puppet-common.postinst should totally be creating that user rather that puppet.postinst. [20:58] Oh, as you say yourself, apparantly :) [21:06] zooko: I'm not finding anything helpful... Sorry :( - maybe we need to package netselect or apt-spy for Ubuntu with a few tweaks to make it work. Email me and I'll look into it. [21:07] What's your email address? [21:08] * MTecknology @ubuntu.com [21:13] zooko: thanks, I keep my to-do list in email form :P === RoAk is now known as RoAkSoax === RoAkSoax is now known as RoAk === RoAk is now known as RoAkSoAx [21:31] how do you mount the persistence file on a LIVECD/USB key? [21:36] New bug: #522339 in openldap (main) "package slapd 2.4.18-0ubuntu1 failed to install/upgrade: le sous-processus script pre-removal installé a retourné une erreur de sortie d'état 1" [Undecided,New] https://launchpad.net/bugs/522339 [21:37] New bug: #520729 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/520729 [22:56] New bug: #522160 in qemu-kvm (main) "qemu 0.12.2 now available" [Undecided,Fix released] https://launchpad.net/bugs/522160