=== andrewdell is now known as andrewl === andrewl_ is now known as andrewl [01:42] anyone here use nfs shares? [01:49] billybigrigger: sure [01:52] giovani, what do you use for mount options in your fstab? [02:03] hey guys im having a weird issue. I just had a power outtage that lasted longer than my UPS could handle. Anyhow the system came back up nicely but i have 2 raid arrays that arent mounted but when i try to mount the manually by running sudo mount -a it tells me that its either mounted or the mount point is busy. when i type mount it doesnt show up that its mounted and when i umount it it says its not mounted any advice? [02:10] billybigrigger: uhm, hard,nosuid,udp off the top of my head [02:10] billybigrigger: why do you ask? [02:12] oh and sec=krb5i [02:14] just finished setting up my pptp vpn server. how can I get access to smb and afp shares from outside once i connected through the vpn server? [02:21] giovani, because the wiki page for nfs looks outdated [02:23] hey guys, trying to make my own google docs thing with my own server any ideas? [02:25] billybigrigger: outdated how? === dendrobates is now known as dendro-afk [02:26] hey guys im having a weird issue. I just had a power outtage that lasted longer than my UPS could handle. Anyhow the system came back up nicely but i have 2 raid arrays that arent mounted but when i try to debfx [02:26] mount the manually by running sudo mount -a ittells me that its either mounted or the mount point is busy. when i type mount it doesnt show up that its mounted and when i umount it itn says its not mounted deegee any advice? [02:26] T3CHKOMMIE: I think a few of the general colaboration suites share some of Google Doc's features [02:26] i.e. Zimbra [02:26] not that I've used them [02:27] giovani, thanks! [02:27] but search "open source collaboration" [02:27] fengoffice.com too [02:31] giovani, you know of anything kinda free? like does oppen office suport any web/cloud stuff? [02:31] Zimbra is free [02:31] as is fengoffice [02:31] hm, [02:31] openoffice is just a desktop application [02:31] got it. [02:31] so no web/"cloud" stuff [02:36] doesnt look like zimbra is in the repos. [02:36] im a bit worried about having zimbra crash my dovecot email server :S [02:37] thanks for the headstart giovani, gota jam! === Doonz_ is now known as Doonz [02:59] hello! I'm thinking to put some virtual machines in my server. But I was thinking, how it works to install a new OS in a Virtual HardDisk if i'm at console? [03:10] X-warrior: you can create HDDs from CLI [03:10] and enable VNC in the VM [03:10] mount the iso image form cLI again [03:10] fire up the VM [03:10] and connection from a PC to the VNC [03:11] that is how i am doing it with virtual Box [03:11] clusty, so I need to create the virtualmachine in my desktop (for an example) and them configure a remote access system... shutdown it put in ubuntu server and start running. Right? === dendro-afk is now known as dendrobates [03:13] X-warrior: look at VBoxHeadless [03:13] VBoxManage [03:13] ... [03:13] read the doc [03:13] it's nicely written [03:13] clusty, ok, thanks I will take a look! [03:16] clusty, that is nice! i didn`t know it! thanks alot! [03:17] X-warrior: i have done the whole shabam only a few times myself [03:17] i prefer usually to config the VM graphically [03:17] and upload tot he server [03:18] X-warrior: just beware that the FOSS Vbox does not support VNC [03:18] you want the PUEL version [03:18] you gotta add another repo from oracle (ex Sun) :D [03:23] clusty, I don`t have sure if I understand. Are u saying that the normal repo version don`t support this feature? So I need to add another repo from oracle? [03:23] yes. last time i chacked (about 1 year ago) the ose version did not support [03:24] http://www.virtualbox.org/wiki/Linux_Downloads [03:24] pick the linux version you are running [03:25] clusty, yes i'm already at this page [03:25] ;) [03:25] the PUEL is free only for personal use [03:52] Hi [04:03] mm [04:03] Talk? [04:03] I have a virtual server that, due to some network misconfiguration, gives me a bunch of messages on boot such as "init: network-interface (eth2) pre-start process (491) terminated with status 1" and then just a blinking cursor: I can't seem to get a shell or login or anything else prompt. How can I recover from this? [04:04] !network [04:04] Wireless documentation, including how-to guides and troubleshooting information, can be found at https://help.ubuntu.com/community/WifiDocs [04:04] Hmm [04:04] Ok sorry [04:05] Have you tried restarting network? [04:05] What, on the working host, or on the broken VM? [04:06] Perhaps I wasn't clear: I have a VM that will no longer boot, and I can't figure out how to recover it. (This is a 10.04 minimal VM install.) [04:06] Is it openvz? [04:06] Holding down shift or pressing ESC doesn't bring up a grub boot menu. [04:06] openVZ? or Xen? [04:06] VirtualBox even? [04:07] It's standard ubuntu KVM. [04:07] Oh, sorry i have no experience with KVM [04:07] Do you have access to the node? [04:08] cjs: What happens when you try to boot? Have you tried booting with Super GRUB2 Disk? [04:09] cfs: Is this a node or just a computer/server running ubuntu and kvm?, do you have root access to the host? [04:09] I boot the node, it gives me: "fsck from util-linjux-ng 2.17.2", "/dev/sda1: clean ..." and then messages similar to the one I posted above. Then a blinking cursor on the next empty line. The only keypress that seems to do anything is ESC, which repeats all that from the "fsck from ..." part. [04:09] Yes, I've got full access to everything. [04:09] Ok well [04:10] I don't know much about KVM but, if you can some how mount the image? [04:10] That way you can edit the vms files [04:11] Hm. I suppose I could try that. It seems a bit...frustrating that a linux boot could get into a state where the kernel loads and userland starts, but you can't recover the system. [04:12] Well if you can get into the vms files [04:12] "VMS files"? [04:12] The vm itself, from the host [04:12] You can backup some files too [04:12] Ah, you mean the filesystem the virtual machine is booting from. Yes. [04:12] Yes [04:13] Like i said, don't know much about KVM [04:13] But if you have access to the file system :) [04:13] Well, this doesn't really appear to be a KVM problem; it appears to be an Ubuntu server userland problem. [04:13] More than likely yes [04:14] fsck from util-linjux-member:ng 2.17.2, isn't that the boot image? [04:15] As in, "WTF won't it go single user if something is wrong, rather than effectively wedging?" [04:16] I would presume that the startup scripts run fsck fairly early on. [04:16] mhm [04:16] probably [04:16] Sorry, i can't personally help you [04:16] Someone else might be able to :) [04:17] and remove personally from that message [04:17] :) [04:17] well what i mean is [04:17] i don't know how to fix this certain problem [04:20] Ha, looks like it's a udev/rules.d/70-persistent-net.rules that's on crack. [04:21] ah [04:21] Nope, that didn't fix it. *Sigh*. [04:25] Hi kokozedman :) [04:25] Lol [04:25] What was your question? [04:25] heheh [04:25] the NTP does not seem to work: all servers have their own time [04:25] which is causing a lot of problems [04:25] i thought Ubuntu already came with NTP already working [04:26] https://help.ubuntu.com/10.04/server/C/NTP.html [04:26] Smooch: page not found 404 [04:26] whoops [04:26] Ok, simple enough. When I edited /etc/network/interfaces, I had "auto eth1" followed by config lines, but forgot the "iface eth1 inet static" between them. [04:26] This one https://help.ubuntu.com/7.04/server/C/NTP.html [04:27] try running ntpdate ntp.ubuntu.com does it work? [04:27] https://help.ubuntu.com/10.04/serverguide/C/NTP.html [04:27] :P [04:27] kokozedman: Was your local time very, very close to the servers when you started ntpd? If it's too far out, and especially if your hardware clock has a large skew, ntpd will not sync. [04:28] BTW, "ntpdate -d " will give you debug output indicating what time the remote server thinks it is. [04:28] Smooch: yes the ntpdate works all the time, but after some days, it will be weird again... [04:28] isn't it running automatically? [04:28] i mean, isn't it supposed to be so? [04:28] kokozedman: If doing ntpupdate works.. why not set a cron job? [04:29] maybe set the timezone again? [04:29] Smooch: i thought so, but isn't that supposed to be the case in Ubuntu Server… listening to you talk makes me think that it's not the default behavior [04:29] timezone is already set… but i have another problem with that also [04:29] I'm not sure if its the default behaviour actually [04:30] Something must be resetting the time zone [04:30] Are you connecting to another dns server? or through proxy? [04:30] these are new servers, so i don't know what is that "something" that is causing the reset [04:31] How many servers are there? [04:31] i am connecting to the server through a HTTP proxy (thru corkscrew) but it's the server itself which is the proxy [04:31] 4 servers [04:31] and the server that its connecting thru a proxy has the right time zone? [04:32] yes, all of them have the right time zone… but it's not about the right timezone which is troubling, but the timezones that the applications actually see [04:32] i have several CGI and Django applications on each servers [04:32] and they all have a very weird time issues [04:32] sometimes the TZ environment variable is set to "America/Chicago" [04:33] and sometimes they are "Asia/Karachi" [04:33] the latter one is the one that is correct [04:33] the former one is totally incorrect [04:33] and related to this timezone issue is the time that the applications receive [04:34] rarely do the application get the right time, but most of the time, it's using the time in America (Chicago i suppose) [04:34] https://help.ubuntu.com/community/UbuntuTime ? [04:34] whoops [04:35] thats desktop one :P [04:35] i've followed that many times already [04:35] lol [04:35] try opening /etc/cron.daily/ntpdate ? [04:35] whats in there? [04:35] it's a bash script [04:35] shouldn't be [04:35] wait [04:36] i don't have that [04:36] whats in /etc/ntp.conf ? [04:36] i only have ntp [04:36] i only have /etc/cron.daily/ntp [04:36] not ntpdate [04:36] try ntupdate [04:36] i guess that has something to do with the 10.04 [04:36] command not found [04:37] apt-get install [04:37] you mean ntpdate? [04:37] apt-get install ntpdate? [04:37] mhm [04:37] sudo first of course [04:37] i'm already root [04:37] thats fine then [04:38] man! ntpdate is already the newest version. [04:38] enter man ntupdate does anything come up? [04:38] but i don't have /etc/cron.daily/ntpdate with that [04:39] kokozedman: For these applications, is the time zone supposed to be the correct one for the location of the server, or always the same regardless of where the server is located? [04:39] an entry is found for man ntpdate, but not for man ntupdate [04:39] (Although a better question is: why would you application care about what time zone you're using? Log stuff in UTC if you work across multiple time zones; it will make life much happier.) [04:40] kokozedman: ntpdate is normally run just once at boot, and perhaps once in a while when you bork your system. Otherwise ntpd should be taking care of keeping things in sync on a continuous basis. [04:40] cjs: it's supposed to be always "Asia/Karachi", no matter where the servers are located… the server are serving applications to Pakistan and it has to be honoring the time there… but at this time they are having time pointed to Chicago [04:41] kokozedman: In that case, set the TZ variable in your startup script. [04:41] where does the server ip look up to? [04:41] e.g what country/place do they say in whoius [04:41] *whois [04:41] cjs: do you mind telling where is that startup script? [04:43] You're the guys starting the application, not me! :-) [04:43] Is this an app you developed, or is it something someone else wrote that you installed? [04:44] Keep in mind, I'm talking about the *application* startup script. The idea is that the application overrides whatever it's given when it starts. [04:45] cjs: i've already set the correct timezone (TZ) inside /etc/apache2/envvars … when i do, it has the correct timezone for about 2 or 3 requests, and then it goes back to the weird Chicago one after [04:46] i have developped it, i intended to receive the local computer's time and not care about what timezone is the script in, or is the time correct … i just assume that it will get whatever time the system is in [04:46] it has worked flawlessly in Gentoo Linux for about 4 years now [04:46] and since i'm leaving control of the server to a less proficient person, i decided to change the system to Ubuntu Server to make things easier for the person [04:47] Hm. Sounds like something is going wrong with apache or its startup scripts, then. I can't really help with that; I've not used apache in years, and never under Ubuntu. [04:47] but now, Ubuntu's time is getting us crazy [04:47] But you might consider reconfiguring things so that your application deals with the timezone, and doesn't depend on the environment, which is harder to control (as we've seen). [04:48] Maybe changing back to gentoo would make things easier. :-) [04:48] Yeah [04:48] cjs: lol hahahaha… we paid $75 for each servers to be switched to Ubuntu… no way we're paying that again to switch back to Gentoo!!! [04:49] i just assumed that Ubuntu is stable enough to have weird problems like that [04:49] but i was wrong then! :-/ [04:49] Ubuntu is great actually [04:49] But sometimes, you have to compile things from source.. if you know what i mean ;) [04:49] yes i know… but i don't know why it's doing this kind of weirdness in the servers [04:50] Did you transfer it all to ubuntu? [04:50] Perhaps you should back it up and re install ubuntu [04:50] i'm actually using a lot of Ubuntus in my VMware Fusion [04:50] If you want stability, don't change stuff. :-/ [04:50] cjs: actually, i was sick and tired of Gentoo's package system… and Ubuntu's update and security is super [04:50] Actually [04:51] I'm not a big fan of Ubuntu, but it doesn't seem much less stable than anything else out there that gets updated regularly. [04:51] Maybe its the resolv.conf? [04:51] kokozedman: Yes, that's one of the main reasons I use it. [04:51] Smooch: what's about it? [04:51] And why I live with stuff like installing 10.04 LTS may go and trash disks on your KVM virtual hosts. [04:52] Well, there the nameservers right. [04:52] If your ntp'ing to the server [04:52] perhaps ubuntus ntp server is getting the wrong ip [04:52] well the wrong location [04:52] i don't think so [04:52] it's using the OpenDNS [04:52] all of them [04:55] Oh [04:55] Umm [04:55] I'v read tons of things saying that openDNS is not good [05:00] oh? [05:00] then, i'll use Google DNS then [05:03] try it :) [05:05] erm [05:05] I have a question of my own :) [05:06] er [05:06] kokozedman: openDNS lies actually, instead of giving you an error when a domain doesn't exist and letting your browser decide it lies and tells you that it exists, and points you at a server with advertisment on it [05:06] how come apt-mirror takes so much cpu? [05:07] amstan_: i don't really consider that a harm ;) … it's just business, they need to advertise to run the servers right? [05:08] kokozedman: except it's their business at our expense, it sometimes breaks things [05:08] true [05:08] when you're trying to do dns stuff for example, and you want to ping to check if a dns works, you'll always get a response [05:08] yeah [05:09] i didn't think that far :P [05:09] might make my own dns servers [05:10] would anyone be interested :P [05:11] heheheh [05:11] Smooch: not really. i would probably get crappy pings [05:11] hehe [05:12] google dns is pretty decent these days, except that my ISPs DNS servers are faster(TekSavvy) [05:12] is there a way to check that crontab syntax is correct? [05:12] send your crontab via pastebin [05:12] and ask if its correct? [05:12] kokozedman: put it on pastebin? [05:12] hehehe [05:12] lol [05:12] lol [05:14] http://paste2.org/p/903943 [05:15] it is the last 3 lines which is causing things to not work [05:15] *; ? [05:16] oh, i guess i see [05:16] lol [05:16] for the username [05:16] hehe [05:17] hehe === Smooch is now known as Smooch_away === Smooch_away is now known as Smooch [05:20] kokozedman: Why not just run your own DNS servers? It's as easy as "aptitude install bind9". [05:21] cjs: what's wrong with Google's? i'm using that now [05:21] It's yet one more external dependency to go wrong. [05:21] How do I generate the UUID for a hard drive? [05:21] Not to mention that it lies from time to time, as mentioned above. [05:21] cjs: to go wrong? DNS would be the last thing to go wrong [05:22] GhostFreeman: hard drives don't have UUIDs. Perhaps you're thinking of a partition or something like that? [05:22] cjs: there's usually 3 DNS servers [05:22] ok, a partition [05:22] amstan_: DNS is one of the first things to go wrong. See above about Google's DNS lying to you. [05:22] what is the weapon of choice. [05:22] GhostFreeman: you don't generate it, you just read it [05:22] GhostFreeman: What kind of partition? [05:22] I need to read it so I can add it to fstab [05:22] GhostFreeman: ls /dev/disk/by-uuid [05:23] a Linux LLVM part [05:23] GhostFreeman: Ah! You want to know the UUID of an existing partition! "blkid" is your very, very best friend, here. [05:23] what is an example command for blkid [05:23] the comments in fstab are very vague [05:23] "blkid". (As root, though.) [05:24] got it [05:24] cjs: meh.. my way doesn't need root [05:24] amstan_: yes, but your way doesn't give the partition types as well. [05:25] cjs: so.. what about google lying? [05:25] Oh, that was OpenDNS. Oops. [05:25] google doesn't do the advertisment on nonexisting domains [05:25] got disconnected :P [05:25] once I edit fstab, how can I reload it? [05:25] GhostFreeman: mount -a [05:26] GhostFreeman: It's not continuously "loaded" by anything. The mount command (and various others, such as fsck) look up data in it when they start. [05:26] cjs: so, i'm good with Google's DNS then ;) [05:26] thanks everyone [05:26] kokozedman: yes [05:26] GhostFreeman: If you want to mount the filesystem you just added, just type "mount /foo" or whatever its mount location is. [05:26] :D cool [05:26] kokozedman: I guess. I just reckon, why rely on yet another external service when it's trivial to do yourself. [05:26] kokozedman: mount -a just reads the fstab and tries to mount anything not already mounted [05:26] cfs great point [05:27] One last question: What's a good disk partition editor that's not complicated like parted? [05:27] js sorry [05:27] Gparted [05:27] Heck, I even run separate resolving servers on my virtual hosts, even though they could use the one on the host host. [05:27] cjs: hosting dns will induce extra delays on dns lookups [05:27] amstan_: ? [05:27] ...that works in a shell [05:27] (so no gparted) [05:27] fdisk [05:28] cjs: well.. the local dns server will have to check the database, and if not there already check with the upper servers, why not just skip the local dns and make all comps use those external servers? [05:28] amstan_: you mean, resolving stuff yourself is slower than having someone else do it for you? Only if it's cached in theirs and not yours. If you both have it cached, using the local copy will be slightly faster. [05:29] It's also a security issue, of course. You have to trust Google quite a lot to use their DNS. [05:29] but you have to trust someone either way [05:29] local dns doesn't magically make you not need to ever trust another dns server [05:29] amstan_: Your characterization of a DNS lookup is not correct; walk through it slowly. [05:30] amstan_: And the issue is not that you have to trust someone, it's that you have to trust Google *as well as* anybody else you'd also have to trust. [05:30] isn't a local dns server just a cache? so you have to tell that server to use another server to lookup the unknown domains? [05:31] It's a decision you have to make, of course. If typing "aptitude install bind9" is really that much work, or the result uses up too much disk space, or whatever, sure, go ahead. But you are certainly (if possibly only in a very tiny way) increasing your changes of DNS resolution failing and security problems. [05:31] Any resolving DNS server is just a cache. Google's is no different from yours (though they might be using something other than bind9). [05:32] ok.. so knowing that, what dns resolver do you give to your local dns server? [05:32] But when you query their server on a.b.c.d rather than your own on 127.0.0.1, either way the server will go through the exact same process to resolve the name. [05:32] amstan_: There is no DNS resolver for your local server. That's the point. Your local server is the resolver. [05:33] It only ever asks for non-recursive answers, and follows the chain itself. [05:33] cjs: but where does it get its info from? [05:33] cjs: how do i make the ntp (which is starting ntpd) service started on start-up? [05:33] sorry, i didn't mean to highlight cjs [05:33] :P [05:33] amstan_: how does it know the root servers? /etc/bind/db.root. It's a set of "well known" data. [05:34] kokozedman: If you just did standard apt-get of it, it should start itself. Check first for /etc/init.d/*ntp*, and then /etc/rc?.d/*ntp* [05:34] kokozedman: according to http://www.howtoforge.org/perfect-server-ubuntu8.04-lts-p7 it's enough just to install it [05:34] amstan_: well, i think 10.0 [05:34] kokozedman: same thing for this particular thing [05:34] 10.04 have it installed by default [05:34] right? [05:34] i don't think so [05:35] no way man! i have not installed that service on my own… but it's there [05:35] in all servers [05:35] i mean: /etc/init.d/ntp exists in all servers [05:35] idk then, i did not say i'm certain [05:36] but here is what i noticed: when i start that service, i will not be able to run ntpdate [05:36] and thing is: i can run ntpdate freely on ALL servers [05:36] It depends on which flavour of server install, too. There's, e.g. "minimal" vs. regular, and minimal has that extra "Basic Ubuntu server" option in the package install screen in the installer which I don't know just what it does. [05:37] which means, they have not been run at start-up [05:37] kokozedman: Yes. You can run ntpdate with -d, to see the difference, but if you try to run it without that it tries to use the port that ntpd is already listening on. [05:37] cjs: it's the normal [05:37] And anyway, you always want to stop ntpd before running ntpdate, and start it again afterwards. [05:37] yeah, i just tought to check an existing install @hypertriangle.com and i get 5 Jul 00:41:51 ntpdate[7191]: no servers can be used, exiting [05:38] erm [05:38] cjs: exactly!! i did not have to stop ntpd and ntpdate worked well [05:38] Yes, but is ntpd still happy? [05:38] if ntpd is running, then i'll get something like: 5 Jul 09:38:50 ntpdate[12209]: the NTP socket is in use, exiting [05:39] what you mean? [05:39] And by how much did ntpdate adjust the time? [05:39] kokozedman: That message means that ntpdate didn't do anything. [05:39] the point i'm trying to proove is that ntpd is NOT start at boot time [05:39] Well, to prove that, reboot and then type "ntpdc -p". [05:40] otherwise the first time i tried ntpdate, i would get that error because ntpd is already running [05:40] Though I have six Ubuntu 10.04 servers here that say otherwise. :-) [05:41] i'll have to wait for tonight to reboot, because people are using the services :P [05:41] ntpdate does not work correctly, use ntpdate-debian [05:41] kokozedman: Oh, BTW, do you use etckeeper? If not, you really want to. You really, *really* want to if you might have to come back to the installation after someone else has been administering it for a while. [05:41] it's because of how the package is configured [05:41] kokozedman: Pffft. Users. Don't let such unimportant things get in the way of your work. [05:42] lol [05:42] kokozedman, ^ [05:43] that problem stumped me at first as well [05:43] qman__: i'll try that… but please confirm: does ntpdate-debian run ON STARTUP? [05:43] ntpdate is present but will not work correctly, at all [05:43] ntpdate-debian runs periodically [05:43] it's in cron [05:43] qman__: thanks for confirming this problem [05:44] :) [05:44] cool [05:44] I'm also fairly certain it runs post if-up [05:44] you may see that i'm randomly doing emotions [05:44] its so i don't get disconnected [05:44] so it would effectively run on startup [05:44] Smooch: lol… why the disconnection? [05:45] kokozedman: irc servers auto disconnect idle connections some times [05:45] not this one [05:45] ??? not on mine [05:45] ok [05:45] also [05:45] to the errors i see, it is actually your connection which is closing [05:45] does anyone know how to successfully setup a irc server with nickserv on ubuntu 10.04? [05:46] I have not done it myself [05:46] neither did i [05:46] not sure if i should reinstall a vps with centos of ubuntu for this [05:46] why would you want to anyway? [05:46] I've heard that unrealircd is best, but I can neither confirm nor deny [05:46] i need to create an irc channel for a commercial project [05:46] oh… i see [05:47] gtg [05:47] about your disconnection issues [05:47] make sure your client is sane [05:47] mIRC has a habit of doing that, but I can see you're not using it [05:47] collyquy mac [05:47] colloquy [05:48] Smooch: running on ADSL? [05:48] i mean ADSL router? [05:48] not right now [05:48]