[00:16] hey [00:16] just because I am curious, how come http://cloud-images.ubuntu.com/releases/12.04.2/release/ doesn't actually contain 12.04.2 images? [00:17] looks like 12.04.3 was uploaded into it today [00:24] smoser: see pabelanger's question about http://cloud-images.ubuntu.com/releases/12.04.2/release/ containing 12.04.3 ... [00:40] is it adviseable to change files like passwd and group to 600? [00:40] Darkstar1: no, that's what shadow is for [00:41] I don't understand what you mean [00:42] Darkstar1: passwd used to contain the hashed passwords in addition to all the other data currently in passwd; but as offline brute-force attacks against the hashed passwords became more commonplace, the 'shadow' system was created to keep the hashed passwords in mode 640 files, reducing the availability of the hashed passwords [00:43] but ls -l, ps aux, and other utilities all benefit from having user-readable passwd [00:43] Ok [00:43] Thanks [00:57] I'm having trouble creating new libvirt kvm VMs, does this look familiar to anyone? :) http://paste.ubuntu.com/6034869/ === freeflying is now known as freeflying_away [01:06] 12.04 server is still supported right? [01:07] i'm getting issues on an OpenVZ VPS when installing iptables-persistent and i'm not getting any good output from the breaking apt-get installation [01:07] is there a way to get more than just this: [01:07] dpkg: error processing iptables-persistent (--configure): subprocess installed post-installation script returned error exit status 1 [01:09] TheLordOfTime: I don't think openvz guests get to use iptables do they? [01:09] sarnold, afaict they should be allowed to. [01:09] i can check with the provider, but it's not my openvz setup [01:09] sarnold, i know that i've been able to use iptables on openvz guests before on other providers [01:09] so maybe it's different between providers [01:11] sarnold, the fact i was able to put a POSTROUTING entry into the nat table makes me think there *is* iptables access [01:12] just a qq: http://askubuntu.com/questions/223944/how-to-automatically-restart-tomcat7-on-system-reboots the script in the accepted answer will run tomcat as the user that calls the script right? [01:17] Does anyone know how how to get a ubuntu (12.04+) to bind to an Open Directory hosted by a Mountian Lion server? Is it possible? [01:18] TheLordOfTime: did you figure it out? [01:19] Darkstar1: that'll start /usr/share/tomcat7/bin/startup.sh as root [01:19] sarnold, no, but I confirmed there is access to iptables on the box. [01:19] s/box/VPS. [01:19] s/box/VPS/ [01:19] Darkstar1: chances are very good that the tomcat7 configuration specifies a user for tomcat to change to once it has bound sockets [01:19] evening. [01:19] by opening an HTTP listener on a given port, rejecting it, and then later ACCEPT-ing it in iptables [01:20] TheLordOfTime: thanks for the correction :) [01:20] and that did as expected: blocks the traffic, and accepts it [01:20] sarnold: Figured as much. I don't think so but I'll try the script now. [01:20] hey, I'm looking to contribute to the ubuntu effort. I'm a linux admin. is there a place I can help here? [01:20] sarnold: I mean I don't think tc7 changes to another user after binding * [01:20] sarnold, but i still am trying to get more verbose error output as to *why* postinstallation failed [01:21] sarnold, any idea how i can get more verbose output? [01:21] Darkstar1: eww. :) [01:22] TheLordOfTime: check /var/log/dpkg.log? I wish I had a better answer, launchpad is full of those bugs :( [01:23] jasondotstar: welcome aboard :) there's a lot to do, the hard part will probably be figuring out what you'd like to do. :) [01:23] sarnold, it'll probably get another one sooner or later from me depending on what I get from the dpkg logs [01:23] TheLordOfTime: hehe [01:23] sarnold, what do you think the priority is for a "Package Fails To Install" bug... just, you know, out of curiosity. [01:23] sarnold thanks. is there a place where I can do some configuration management stuff? I've heard cool stuff abt juju [01:24] Does anyone know how how to get a ubuntu (12.04+) to bind to an Open Directory hosted by a Mountian Lion server? Is it possible? [01:25] sarnold: problem is if I try to run tomcat as a created user (with no home and no shell) I get propmted for password still. [01:25] jasondotstar: yeah, juju looks cool. I haven't used the newer go-based version yet, but the old stuff was cool and the new stuff has more features. :) check out #juju, there's lots of friendly folks there who can help you write charms, debug installations, etc [01:25] sarnold ack. [01:26] sarnold are you guys maintaining the ubuntu infrastructure here? or answering questions abt ubuntu server stuff here? or both? [01:26] jasondotstar: this channel is mostly for people developing server components, but we do help users with questions as well [01:27] Darkstar1: there's a -user command line option described here that looks related: http://tomcat.apache.org/tomcat-7.0-doc/setup.html#Unix_daemon [01:28] * Patrickdk_ just attempts to abuse users in here :) [01:28] sarnold i c. [01:28] atleast till ikonia beats me up over it [01:28] sarnold, nothing in dpkg.log :/ [01:28] Patrickdk_, heh [01:28] or until someone tells ikonia about it, in either of those cases you get beat up :P [01:30] sarnold: thanks I already read that one. But want to use a normal startup script. Looks like I'll have to use it in the end :) [01:35] Darkstar1: I was thinking you'd modify the startup script to use the -user .. [01:35] sarnold: except I always get asked for a password. but the user can't login [01:36] Darkstar1: what prompts you for the password? o_O [01:38] sarnold: su -c "$TOMCAT_HOME/bin/startup.sh" noone <====== when I run the script I get prompted for password because of this line [01:39] Darkstar1: check out the startup.sh file -- does it use 'jsvc' to start tomcat? [01:40] nope [01:41] Darkstar1: can you pastebin that startup.sh? [01:41] Darkstar1: (the pastebinit package can make that easy :) [01:43] http://fpaste.org/35331/ [01:43] that's from catalina.sh which the startup sctript calls [01:45] man that's gross. :) [01:51] Darkstar1: how about changing your su line to: su username --shell /bin/sh -c "/path/to/tomcat/home/bin/startup.sh" ? [01:52] Darkstar1: perhaps $TOMCAT_HOME doesn't work in that location. and perhaps you need a shell for the command to be executed at all. [01:52] if your user doesn't have a shell, you may need to specify one here. [01:53] sarnold: going to try out the daemons.sh : http://fpaste.org/35332/ which apparently can use jsvc [01:55] Darkstar1: oh yes! that's better. then you can just use --tomcat-user right there :) [01:57] hope so [01:57] * Darkstar1 crosses his fingers [01:59] sarnold: getting a Cannot set group id for user 'apache-tomcat' with that. [01:59] Error validating user 'apache-tomcat' [01:59] it would seem creating a user with no home and no shell is problematic [02:00] Darkstar1: yeah, I'm not -too- surprised. set the home to the data directory and shell to either /bin/false or /bin/sh .. [02:00] Darkstar1: did you provide a primary group id in the passwd file too? [02:01] sarnold: also su username --shell /bin/sh -c "/path/to/tomcat/home/bin/startup.sh" <== prompts for passwd [02:01] that was executed from command line [02:02] sarnold: no I didn't modify the passwd file when I created the user [02:03] but it created one [02:03] Darkstar1: just to check, you are running that 'su' command as root, right? [02:03] no [02:03] aha :) [02:04] run it as root :) [02:06] sarnold: thanks [02:06] I was able to run the daemons command which uses jsvc and runs tomcat as the other user [02:07] Darkstar1: nice. [03:02] Anyone know how I can run a command as a user AFTER startup home directory decryption has taken place? [03:18] put it in that user's .bashrc (or .profile or equivalent) === freeflying_away is now known as freeflying [07:10] i have some servers running ntpd and the time still drifts off for some strange reason [07:11] ntpq -p shows that there is no offset or jitter.. [07:12] ntpdate -q shows a offset of 19 seconds [07:12] if the offset and jitter are actuall 0, then you're not connected [07:13] there is always some offset and jitter [07:13] ok [07:13] then why does it list some servers? [07:13] and why doesn't it reconnect? [07:13] you're looking for a '*' at the start of the line in ntpq -p output - that will tell you you're actually syncing [07:13] ok [07:14] it lists the servers it's trying to sync with [07:14] usually the st column has a high number, like 16, if you're not syncing [07:14] yeah i got 16 [07:14] you want a low number (<=3) for it to be considered a good host for syncing [07:16] this can be anything - a firewall might be blocking ntpd but not ntpdate (ntpd uses source port 123, ntpdate uses a high source port. Some firewalls think this is bad and drop ntpd) [07:16] or the server you're trying to sync from is itself out of sync [07:16] i use public [07:16] firewall? [07:16] http://pastie.org/8276376 [07:16] nope [07:17] if i stop ntpd [07:17] and run ntpdate against 0.se.pool.ntp.org it works [07:17] it fixes the time.. and I start ntpd again and after a while it drifts again [07:18] is there something wrong with my ntp.conf? [07:20] run 'tcpdump port 123 -s0 -A -vvv -n' ; restart ntpd. Pastebin 60 seconds of output [07:20] I'm on the train, so I may drop off but I'll look when I get connectivity again [07:22] ok cool thanks [07:26] ohh crap [07:26] tcpdump listens on wrong interface [07:26] i have to specify bond0 [07:31] sgran 0 packets captured [07:31] 5 packets received by filter [07:31] 0 packets dropped by kernel [07:33] Raboo: what I suspect and am asking tcpdump to confirm, is that you are sending traffic out and not getting a reply because of some network device dropping traffic with a low source port [07:33] since no traffic is leaving your machine, I now am considering your machine to be the thing stopping the traffic [07:33] why does ntpdate set the time correctly against same servers then? [07:34] because it is a different program, and uses a different source port, and runs as a different user and ... [07:34] there's lots of things that could cause the difference [07:36] ok [07:36] do you use any RBAC like apparmor or selinux? [07:36] Do you have host based firewall rules? [07:37] if you strace ntpd, do you see it trying to make outbound connections? Do you see it succeed or fail? [07:37] these are the sorts of things I'd consider [07:37] ok [07:38] well there is a firewall since this is not a public ip.. But I didn't think it would block outgoing traffic.. [07:38] and iptables is not running locally [07:41] if I were you, I'd start by trying to get a capture of ntpd traffic. This involves the tcpdump above, and looking for traffic going out but not coming back in [07:41] that will tell you conclusively where to start [07:42] if traffic goes out and comes in, you have a bad peer or a config problem. If it goes out but not in, you have a firewall problem. If it doesn't go out, you have something like an RBAC or local firewall problem [07:42] sgran ok i will start with strace [07:43] kan i run ntpd as root? [07:44] you don't want or need to do that [07:45] ok so i can run strace as ntp user [07:46] as ntp it says "must be run as root, not uid 107" [07:50] ah it's a option to specify user === smb` is now known as smb [08:35] rbasak: the version in repositories isn't the newest (it's actually pretty old). my only experience is with chef and that installs around 40MB of stuff. it has a dependency on ohai, which is a daemon that runs every couple of minutes and scans system info (devices connected, IPs, etc). In my case, ohai was continuously spinning up disks that i wanted in standby so that bit was annoying but otherwise i would not go back to managing servers [08:37] disposable: 40MB really is not significant for most users, or can you give me an example of where this is a problem for a significant number of users? I can't speak for chef, but I know that in puppet you can configure the automatic update interval, or turn it off entirely. Surely chef can do that too? puppet 3 is in Saucy, BTW. Not old! [08:42] I have no problem with anyone who disagrees. But I don't think it's reasonable to recommend against chef and puppet to others, when in fact most of the world uses these tools. [08:46] rbasak: i did not recommend against anything. merely pointed out that for one-off configuration of multiple servers, they're overkill. especially if you have to learn them. [08:46] rbasak: since you use puppet, i'd like to know a few things. does the free version come with any king of web interface? [08:47] No. You must learn the DSL. [08:47] rbasak: s/king/kind [08:47] You can easily run it serverless, though. [08:47] puppet apply [08:47] That will do a one off run, and "make it so". [08:49] how does puppet set IPs? does it just use 'ip' or are there modules for persistent changes in /etc/network/interfaces? [08:51] http://docs.puppetlabs.com/references/latest/type.html lists the built in resource types available. I don't manage IP from puppet (I don't personally think that makes sense). But you could do it by just maintaining /etc/network/interfaces as a file managed by puppet, though, using the "file" resource type. [08:51] Oh. There is a sutiable type available: http://docs.puppetlabs.com/references/latest/type.html#interface [08:52] Hmm. Only for Cisco switches though! [08:56] rbasak: thanks. === exekias_ is now known as exekias [10:10] I'm running 12.10, and I can't seem to install the noip2 package. sudo apt-get install noip2 returns with unable to locate package. Any ideas? === freeflying is now known as freeflying_away [10:20] jamespage, zul: https://code.launchpad.net/~yolanda.robla/ceilometer/autopkgtests/+merge/182583 and https://code.launchpad.net/~yolanda.robla/horizon/autokpgtests/+merge/182168 [11:46] Hello. [11:49] I am looking for a tip on how to most efficiently get DNS servers to share information if they are in two different LANs. [11:50] stlu: details? [11:50] anyone good at reading strace? [11:50] !anyone [11:50] A high percentage of the first questions asked in this channel start with "Does anyone/anybody..." Why not ask your next question (the real one) and find out? See also !details, !gq, and !poll. [11:51] http://pastebin.com/yJHN1Zcf [11:51] my ntpd doesn't update time. It seems like it doesn't connect to the ntp servers [11:51] rbasak i was just about to :-) [11:52] but from what i can see in strace is that it tries to connect to the ntp servers? [11:52] Raboo: if you think it's a network issue, it's easier to isolate using tcpdump. Filter for the right protocol. Eg. "sudo tcpdump -nieth0 port 123 or icmp". [11:53] rbasak: I'm doing this all under virtualbox, as a learning project. [11:54] stlu: If 'information' is zone files, and the DNS server is bind9 - it can do AXFR's itself. [11:55] Daviey; rbasak: I read about the option to set a secondary dns, but if there are several Lans (say, like 8 or 9) then that will be a lot of transfers, no? [11:56] stlu: Yes, but it's designed to do this. Only transfers when a zone file is modified [11:56] I was thinking there might be a proper way to get the DNS in LAN #5, for example, to just point to the other DNS server addresses instead of being a secondary DNS for ALL of them... [11:57] erm [11:57] i think there is some confusion here... is this for domains under your control? [11:57] okay then, I was just thinking I might have had a gaping hole in my understanding. [11:58] It is true that there won't be many transfers after its up and running... [11:58] Daviey: this is a learning project entirely under virtualbox, so yes, these LANs are all my creations. [12:00] If you'd care to know, I will share the setup with you... [12:01] stlu: I mean, are you talking about stlu.com, for example - that you are trying to manage? [12:01] or is this to resolve google.com for example? [12:02] Well, my bedroom is the domain, so I am managing "myroom.local" [12:03] LAN #1 is vbox1.myroom.local, a network of virtual machines, with an accompanying DNS server for their addresses. [12:04] Ah.. erm. .local is usually not recommended.. that typically uses a different protocol called mDNS or Avahi [12:04] its actually all going to fall into an address block 172.24.0.0, each DNS and so-called "lan" will have a block of 256 addresses, that I treat as a /24 network. [12:05] 172.24.200.x is the LAN#1, and 172.24.201.x will be LAN#2, only special thing is seperate DNS servers. [12:07] right.. but i am suggesting you use something other than .local to avoid confusion [12:07] Daviey: oh, well, I lied. My home router provides the domain "gateway.2wire.net" so I am not actually using local (but I had other plans to) [12:08] I actually got the .local idea from a D-Link router that was defaulting to that. [12:09] I looked at the IANA list of suffixes, they don't have a suffix reserved for LANs like I am making. [12:09] stlu: Ubuntu should disable mDNS / Avahi, but not worth risking it whilst learning IMO [12:09] Daviey: totally agree, if I throw in too much unpredictability I won't gain any knowledge, just a headache. [12:11] With DNS, you don't usually have a separate DNS server per network. You might have a separate resolver per network, but they don't need anything special apart from being able to route to DNS servers. [12:12] I'd put one DNS server on a single LAN, and if you want to experiment with secondaries, then put than on a LAN that can route to it and use AXFR. Then you'll be learning something closer to what might be used in the real world. [12:12] You can make DNS as obtuse as you want, but it'd just be prone to problems and I'm not sure that you'd learn anything useful out of it. [12:13] Just make sure that either the LANs can all route to your DNS server, or that they use resolvers that can route to your DNS server. [12:14] rbasak: hmm [12:16] utlemming, people (pabelanger and sarnold) mentioning that http://cloud-images.ubuntu.com/releases/12.04.2/release/ has a banner that says 12.04.3 [12:16] i would personally it rather not say '.X' at all. [12:16] which seems new to me. [12:16] (they complained of the title, i think. i dont like the path as it breaks anything and is inconsistent. [12:16] I'm pretty sure it's always done that. At least since 12.04.1 I think. [12:17] I don't seem to know what a resolver would be... all I really have played with is /etc/resolv.conf to point to DNS. [12:17] rbasak, it would seem that *something* changed in .2 [12:17] http://cloud-images.ubuntu.com/releases/ [12:17] smoser: yeah, the banner on the release webpage takes the name of the latest release. even back to 8.04 [12:18] Oh [12:18] I was talking of the motd banner. You aren't talking about htat. [12:19] is ia-32libs (32bit compat libs) able to conflict with current software? I think it can't but I've never used it so I'm not really sure [12:19] it's ubuntu 11.10 [12:20] YOu know that 11.10 is EOL now, right? [12:21] We don't use ia32-libs any more. I'm not sure whether the change happened before or after 11.10. [12:21] yes, it happened after 11.10 iirc [12:22] it's a production environment, sadly I'm not allowed to update anything [12:24] rbasak: by "configuring a resolver", do you refer to the configuration of resolv.conf on a host? [12:24] Or is this a software package that does something more comprehensive? [12:26] roaksoax, adam_g, yolanda, zul: updates for havana ca - http://people.canonical.com/~jamespage/ca/havana/ [12:35] Ok, so I can set up a separate network for each virtual LAN. If the host OS for the virtual systems has been given a domain name of say "laptop2.gateway.2wire.net", would the household router direct *.laptop2.gateway.2wire.net, or fail? [12:36] *direct all traffic to the *.laptop2... [12:40] thats a definite fail on the subdomain idea... [12:41] Whats the best way to automate CLI tasks? Anything better than expect/pexpect ? (Things like entering passwords, keys, directories to command line programs that ask questions from standard input, to things like automating ssh tasks) [12:49] codepython777: depending on what you need, something like ansible may work. [12:49] codepython777: I was just thinking the same thing. Excepty I didn't even know about 'expect'. [12:50] stlu: I dislike it very much. [12:50] codepython: oh. [12:51] codepython: well I'll try it anyway, and if I don't like it either, well I'll have a better idea of what I want. [12:51] My friend had tried out 'Webmin' but I don't know if all it's features are really needed. [12:51] stlu: last update of expect webpage = 2009 [12:52] a bit stale. [12:52] I suppose you'd be looking for something simpler then Webmin too? And more easily automated I bet. [12:53] just because it hasn't been updated doesn't mean it's not still current. :) [12:53] !webmin | also [12:53] also: webmin is no longer supported in Debian and Ubuntu. It is not compatible with the way that Ubuntu packages handle configuration files, and is likely to cause unexpected issues with your system. [12:54] thx Pici [13:11] zul, jamespage? can you take a look at my MP for horizon and ceilometer and horizon? https://code.launchpad.net/~yolanda.robla/horizon/autokpgtests/+merge/182168 , https://code.launchpad.net/~yolanda.robla/ceilometer/autopkgtests/+merge/182583 [13:11] yolanda, I have been trying to test them but we have a broken cloud-utils in saucy right now - blocking testing [13:12] they look OK [13:12] ceilometer tests fail [13:12] at least in my local machine, but they are testing the right things [13:12] ceilometer-common isn't starting cleanly [13:32] rbasak: do you have any words left about my question? I think I should be able to install those ia32 libs [13:32] without running into conflicts [13:39] stuxlof: what error do you get it you try? === dduffey_afk is now known as dduffey [13:51] sarnold: ping https://bugs.launchpad.net/ubuntu/+source/msgpack-python/+bug/1207003 [13:51] Launchpad bug 1207003 in msgpack-python "[MIR] msgpack-python" [Undecided,New] [14:01] xnox, just spotted your comment on the mongodb openssl bug [14:01] * jamespage does a little dance [14:01] \o/ [14:01] I can sleep again.... [14:01] /o/ \o\ [14:02] jamespage: =)))) lol, ScottK was giving you nightmares =) [14:02] lol [14:04] zul, When will we see you? [14:04] smb: never hopeuflly ;) [14:05] zul, Have people returned your walls :) [14:05] smb: @starbucks today [14:06] zul, Ah so better coffee and better internet. ;-P [14:06] in theory ;) [14:07] * smb is reminded of those day working from the mall [14:07] actually days... === freeflying_away is now known as freeflying [14:34] geser: I'm not getting any errors, I just want to know if this can influence my currently installed stuff in a way [14:34] sorry for the slow response [14:35] I have a perc h700 raid controller on ubuntu 11.10 x64, and the software needs ia32 libraries to be compatible because it's 32bit software. But I've never used these, since this is an production environment I thought it wouldn't be wrong to ask for a 2nd opinion. I think it shouldn't matter [14:36] It uses either megacli or megasasctl, I've got megacli running but it keeps returning with an exit code 0x00 when I try to use it, so I guess megasasctl is the proper one [14:41] it shouldn't matter as ia32-libs installs additional 32bit libraries which can only be used by 32bit apps [14:45] yolanda, horizon tests MP _ test are fine but needs a changelog entry! [14:46] yolanda, ceilometer merged - thanks! [14:46] jamespage, will take a look at horizon, i created it but may be have lost in the way [14:47] zul, as we have 20 mins - http://people.canonical.com/~jamespage/ca/havana/ [14:48] I thought so thanks [14:49] jamespage: doh... [14:49] jamespage: i dont think we need to update pep8 anymore since its not being used anywhere [14:50] zul, oslo-config [14:50] python-quantumclient [14:50] oh - that can be dropped [14:50] jamespage: fuuuuu oslo-config [14:51] jamespage: you need python-itsdangerous for flask i believe === freeflying is now known as freeflying_away [14:54] jamespage: other than that +1 [15:01] Daviey, around? [15:11] jamespage, how do i get a bug tracked on server section of http://reqorts.qa.ubuntu.com/reports/rls-mgr/rls-s-tracking-bug-tasks.html ? [15:11] adam_g, raise a release specific task [15:11] and hope the package is on the list... [15:11] jamespage, against saucy? [15:11] yes [15:11] thanks === raininja is now known as Tech-1 === Tech-1 is now known as Tech-I === Tech-I is now known as denkijin [16:51] zul: msgpack-python is not forgotten, but not yet started. :( this is a busy week. :( [16:52] sarnold: cool [17:22] zul, what version of swift we are shipping /w havana/saucy? [17:23] 1.9.1 [17:24] zul, so the 1.9.3 release that is out is post-havana? [17:24] adam_g: i think so [17:25] oh, 1.9.3 isn't out yet [17:25] unless there is a bug fix release before then === fego is now known as Guest10438 [17:39] jamespage, those CA updates LGTM [17:47] hallyn_: http://paste.ubuntu.com/6037443/ [17:49] I would like to reinstall Ubuntu Server 12.04 and I have a mirror raid with two disks and two paritions on each with md0 being sda1 and sdb1, and md1 being sda2 and sdb2. If I just make sure to untick "format" when I choose partition layout will I then keep my data on those mount points with "format" unticked? [17:50] jonascj: the installer won't reformat md1 unless you do something stupid [17:50] jonascj: make sure you have a backup, though, things sometimes don't go as planned [17:56] zul: what was the exact command line? [17:56] zul: you were trying ot exec the container name for some reason [17:56] d'oh [17:56] zul: lxc_start - ' ' is initialized [17:57] i think you did something like "lxc-start -n " " containername [17:57] hallyn_: ? [17:57] you passed a space as the container name [17:58] hallyn_: oh doh...ill fix that [17:58] zul: ok [18:00] RoyK: Thank you. I don't know how I forgot it, but I forgot to tell you that on top of the raid I have LVM. So i would like my vg0/root formatted but I would like to keep my vg0/home and vg0/var logical volumes. So during install should I choose to mount /var and /home to those location but untick "format" ? [18:00] jonascj: same rules apply [18:02] jonascj: in installer, enter manual partitioning, activate all raids, activate VGS, then manipulate the volumes as needed. [18:02] jonascj: which ones to keep, which ones to format. [18:03] xnox and RoyK: and we are in agreement on this: those mountpoints for which I uncheck "format" will retain their data? [18:08] jonascj: and set to "do not use" [18:16] jonascj: yes [18:17] jonascj: but keep a backup nevertheless [18:17] RoyK: I will when it matters. I'm just fooling around. [18:17] jonascj: backups are usually most useful for people accidentially deleting their things, or some worm comes in and makes havoc [18:18] jonascj: ok, go on, then :) [18:19] RoyK: atm I'm actually such a big fool that I cannot remember why I have two raid devices. md0 is 2gb and md1 is 998gb and only md1 is part of my LVM. What the heck was I thinking to use those 2gb for :S Can I find out if something is on those 2gbs? [18:19] jonascj: you might have been trying to fit the /boot partition in the first N gigs of the disk for ease of use with lilo or similar.. [18:21] sarnold: you have high thoughts of me :P /boot is however on the LVM root logicial volume. Inside my system which resides on LVM on md1, can I peek at md0 and whatever is on there? [18:23] jonascj: I usually make a small mirror for boot, then a larger one for swap and then something bigger for root - if the system doesn't have more than those two drives, I'll make a larger one for /home [18:23] boot as in /boot [18:25] royk: sounds reasonable. What I have now is two physical disks and then two mirros. One which purpose I've forgot (2gb) and then one large for LVM on which I have different logical volumes (root, home and var) [18:26] quite possibly the swap on those 2GB [18:26] the installer should autodetect that [18:27] RoyK: sorry, i forgot swap, I have root, swap, home, and var as logical volumes. I think you are right, I probably intended them for a non-LVM /boot [18:27] but somehow failed that during install [18:28] jonascj: with current systems, it's not really that important to have /boot on a separate partition [18:28] jonascj: guess it's just bad habit [18:28] royk: and /boot on non LVM is also not necessary, grub2 handles lvm just fne [18:29] indeed [18:29] RoyK: /boot on a different partition might -also- be with the idea that it need smirroring but the data partitions might do fine with raid5 or raid6 or raid10.. [18:30] sarnold: ah yes :) but not in my setup :P [18:30] sarnold: somewhat, yes, but I never mix data disks and OS disks anyway [18:30] sarnold: IMHO having data and OS on the same drives is bad, having it on the same partitions isn't good at all === acrocity is now known as Guest5888 [18:32] RoyK: ah :D that's a nice rule of thumb. very nice. [18:33] sarnold: you'll have to reinstall one day or another, so better keep the data off the root === natefinch is now known as natefinch-afk === acrocity_ is now known as acrocity === natefinch-afk is now known as natefinch === dduffey is now known as dduffey_afk === lj1 is now known as lj === freeflying_away is now known as freeflying