[00:34] hello all. I am just getting ready to install server 12.04 on my main house server and I wanted to know except for backing up home anything else I should do? all the data is on separate drives. [00:38] xpistos: Might be worth taking a copy of /etc, just in case you have any relevant settings there you want to go back and take a peak at. [00:38] xpistos: Otherwise it really depends on what used to run on the server. Any daemons storing anything relevant under /var perhaps? [00:38] andol: thanks! I copied my fstab and exports so I think I am good there [00:39] deluge, and I think that is about it? [00:39] No if I can just find a blank cd!!!! [00:40] xpistos, why are you reinstalling? [01:05] escott: upgrading to 12.04 [01:23] xpistos, and you just want to reinstall for the heck of it? [01:24] escott: No whenever I upgrade to a new version I always do a clean reinstall [01:25] xpistos, well be aware of kernel modesetting before you start [01:26] not sure what that it [01:26] is [01:28] xpistos, some graphics cards have issues. for a server you can just throw nomodeset into the boot if you dont need a gui [01:32] I might but only like icewm or awesome window manager tupe stuff [01:43] i just found that dspam fell over in my postfix+dspam+dovecot mail system. i've taken dspam out of the delivery train, and new mail is being delivered, but i can't get the mail in postfix queue to flush out... any tips? === robin is now known as Guest61182 [04:17] nathwill: postqueue -f didn't do it? === jeremydei_ is now known as jeremydei [04:42] sbeattie, nope. o_O same errors... it's acting like the mail's still destined for dspam. === th0mz_ is now known as th0mz === th0mz_ is now known as th0mz [08:01] how come our Ubuntu 10.04 servers did not apply any leap second adjustments? when midnight UTC came nothing happened (not message in dmesg, like in Ubuntnu 12.04 servers), other than the time was one second out of sync with the upstream NTP servers. any ideas? === Lcawte|Away is now known as Lcawte [09:11] Hey guys. Does anyone know which channel I could find ubuntu admins/maintainers. There seems to be a outage on a set of repos at the moment === cpg is now known as cpg|away [11:17] what's the current fix for the leap second bug? restart, or update kernel? === Smaug_ is now known as AMR_ === robin is now known as Guest85110 === fenris is now known as Guest79195 === Guest79195 is now known as ejat === bastidra1or is now known as bastidrazor === th0mz_ is now known as th0mz [15:59] First, which files to move from Windows to the server - the "obvious" is to mount "My Documents" on the server, but then the "Local Settings" and "Application Data" are all still local (and so need to be backed up separately). Is it better to put the whole of \Documents and Settings\ onto the server? [15:59] (Sorry - cut-and-paste error!) [15:59] I have a question about configuring windows clients to use a linux server (actually not Ubuntu, but I couldn't find any other linux-server-oriented IRC channels, and the question is about general structures anyway) [16:00] I'm moving data from a collection of workstations to a linux server. [16:01] The question is (as accidentally pasted above) which files to move from Windows to the server - the "obvious" is to mount "My Documents" on the server, but then the "Local Settings" and "Application Data" are all still local (and so need to be backed up separately). Is it better to put the whole of \Documents and Settings\ onto the server? [16:02] Secondly, where to put it on the server - each user will have a linux account, so it could go in the linux home directory and be shared from there, but would it be better in a separate "shared" tree? [16:07] kyuzo, what are you trying to accomplish? Sounds like you want a domain controller and give everyone a domain account. they you can log into each individual workstation with your domain credentials and have your individual user settings (I think that you can configure windows to work that way... not really a windows networking guy) [16:09] I think he wants roaming profiles [16:14] threel8ti: patdk-lap: The goal is to just move the data off the client systems onto an ECC+raid-protected server. Roaming profiles would be nice, but not essential. [16:14] isn't there a dropbox alternative that came out recently... the big sellign point was you hosted the "dropbox" stack on your own hardware. [16:15] patdk, thanks, I couldn't come up with that name. [16:16] I've heard that roaming profiles can be tricky to configure and not 100% effective, even with windows servers. [16:16] How would dropbox help? [16:16] to backup your data... [16:18] you've referred to "the data" a couple of times, and this is what we call the trouble with pronouns; -what- data are you talking about? What is your end goal? What are you trying to accomplish by moving "the data"? [16:18] OK. The fileserver option does have the advantage that users aren't tied to one workstation, though. Dropbox would be tied to one workstation though, woudn't it? [16:19] "The Data" (TM) is a whole lot of "Stuff" (TM). :-) [16:19] (I think "ownCloud" or TeamShare are the projects I've run across recently that are dropbox alternatives) [16:19] It's a collection of MS-Office documents, source code, you name it. "User Data". [16:20] no, you can access your data from "anywhere" re: Dropbox would be tied to one workstation though, woudn't it? [16:20] brb [16:22] Also in "The Data" are a collection of Virtual Machine disk images, though they will probably need to be handled differently for performance reasons. [16:27] what hypervisor are you using? [16:31] A mixture of VMWare Server 2 and MS Virtual PC [16:31] Performance isn't a major issue for most of the VMs - they're just for "clean environment" software testing. [16:43] kyuzo, I think you need to take a step back and redefine your problem. it seems to me that you have a couple things going on and are looking for a silver bullet to all of your projects instead of engineering a proper solution to each problem individually. [16:50] how come our Ubuntu 10.04 servers did not apply any leap second adjustments? when midnight UTC came nothing happened (not message in dmesg, like in Ubuntnu 12.04 servers), other than the time was one second out of sync with the upstream NTP servers. any ideas? [16:59] daff: did you check machines not using the ubuntu ntp servers? [16:59] machines as in ntp clients [16:59] * RoyK doesn't know how well ntp handles leap seconds or how long it'll take to sort it out [17:02] I've found leap time causes problems with ntp [17:02] eg: the jump moves beyond the 5 minute threshold so never syncs [17:02] never seen that [17:02] you have to stop ntp, ntpdate, then restart ntp [17:02] RoyK: the machines are all running chrony and adjusted the time after a while (I issued a "service chrony restart" via puppet on all machines, just to be sure) [17:02] some more info here http://www.eecis.udel.edu/~mills/leap.html [17:03] but I was wondering why only 12.04 machines had the dmesg message "Clock: inserting leap second 23:59:60 UTC" [17:03] which are also running chrony [17:03] chrony? [17:03] !chrony [17:03] a time daemon, like ntpd [17:04] oh [17:04] I just use ntpd [17:04] !ntp [17:04] Information about using and setting your computer's clock on Ubuntu can be found at https://help.ubuntu.com/community/UbuntuTime - See https://help.ubuntu.com/10.04/serverguide/C/NTP.html for information on usage of the Network Time Protocol (NTP) [17:04] perhaps ntpd has a bug there [17:05] only problem with debugging that stuff irl, is that leap seconds only happen every 18 months :þ [17:05] I mean it's probably for the best that there was no leap second inserted, because apparently at least java and mysql go screwed by it [17:05] but I was wondering why 10.04 (with chrony) did not insert one, while 12.04 with chrony did [17:06] oh - they attempt to keep the time themselves? [17:06] sounds like a pathologically bad idea [17:07] I don't know the details yet, but half the internet crashed last night because of java and mysql issues with the leap second [17:07] ROTFL [17:07] http://serverfault.com/questions/403732/anyone-else-experiencing-high-rates-of-linux-server-crashes-during-a-leap-second [17:07] seems my blog is up, and it's a wp blog mysql in the back [17:08] I experienced the mysql problem on my home machine. at 00:00 UTC it started using 100% CPU, increasing the load to around 15 [17:09] looks pretty decent from here http://munin.karlsbakk.net/munin/localdomain/localhost.localdomain/index.html [17:09] that's a Lucid box [17:10] like I said, our lucid machines had no problem, since there was no leap second inserted [17:10] oh [17:10] but I find it strange that there wasn't one inserted. I thought that's the kernel's responsibility? [17:10] this is a Precise box http://door.nilu.no/mrtg/munin/DMZ/door.nilu.no/index.html [17:11] looks good as well [17:11] this not so much: https://blog.mozilla.org/it/2012/06/30/mysql-and-the-leap-second-high-cpu-and-the-fix/ === th0mz_ is now known as th0mz === th0mz_ is now known as th0mz === cpg|away is now known as cpg === jussi01 is now known as jussi === cpg is now known as cpg|away [20:36] i removed the /var dir by mistake.. this is a new server.. short from re-intsalling can I anyother way to restore it? i installed another server on same hardware just few mins ago.. can I just rsync the /var dir? [20:37] axisys: spare the hassle and reinstall, there are a lot of symlinks there... [20:37] guntbert: sure.. np [22:43] I've asked this before but didn't get a real answer. maybe there's someone here now who might know this: how come none of our Ubuntu 10.04 servers applied any leap second adjustments last night? when midnight UTC came around nothing happened (no message "Clock: inserting leap second 23:59:60 UTC" in dmesg). only the clock was one second out of sync with the upstream NTP servers after 00:00 UTC. any ideas? === Lcawte is now known as Lcawte|Away [22:54] daff, maybe cause you screwed it up? cause my ubuntu 10.04 servers say that [22:54] [9780753.205624] Clock: inserting leap second 23:59:60 UTC [22:54] root@patbox:/home/dswett# uname -a [22:54] Linux patbox 2.6.32-37-generic-pae #81-Ubuntu SMP Fri Dec 2 22:24:22 UTC 2011 i686 GNU/Linux [23:09] patdk-lap: interesting, but how? we run hundreds of lucid servers and I haven't found one that applied the leap second [23:09] the all run chrony, not ntpd, but I don't know if that matters [23:10] all servers had the latest tzdata package from march 2012 installed [23:11] it does matter [23:11] how so? [23:11] tzdata is useless, if there is nothing to adjust the kernel using it [23:11] I dunno what chrony does [23:11] but ntpd notifies the kernel about the leap second [23:13] looks like 1.24 added leap second support for linux [23:13] I see. that explains that, then. [23:13] and the verson of chrony in ubuntu 10.04 is lower than that, so it doesn't have support for leap second [23:14] yeah, lucid has 1.23.x, so no leap second support [23:15] thanks for clearing that up for me, until now I wouldn't have thought that there was much difference between ntpd and chrony (other than chrony generally doing an easier and better job of keeping the time in sync) [23:15] odd, I have never heard of chrony [23:15] but I have been using ntpd forever, without issues [23:17] we have had issues in the past with ntpd (long time ago), with the clock of VMs drifting around like crazy, and ntpd not keeping up. maybe we didn't use it right, but simply installing chrony and removing ntpd seemed to have magically fixed that issue. so we stuck with chrony [23:18] hmm, time keeping on vm's is a totally different thing [23:19] yeah, the consenus back then was to use ntpd or chrony to keep time in sync. that's on lucid hosts and VMs (using KVM) [23:19] has that consenus changed? [23:19] dunno, I don't use kvm [23:19] I see. well, thanks for your input and help, now I can sleep better :) [23:21] Anyone know where I can find sun java 5 debs for 10.04? [23:23] smw_: I don't think there are any official debs around anymore, certainly not for 1.5. but you can get old versions from oracle's archives, starting here: http://www.oracle.com/technetwork/java/archive-139210.html [23:23] thanks [23:24] daff, I think I used to have it... then I did an update recently [23:24] daff, I had not updated for god knows how long [23:24] then after the update java5 was gone [23:24] I can tell you my other ones are not getting upgraded until I find out what the hell happened ;-) [23:25] ubuntu and other distros are no longer allowed to package and distribute java [23:25] so an update removed what was already there? [23:25] that does not seem right [23:25] daff, I use KVM in 12.04 and haven't noticed a time drift, however, on previous version of Ubuntu I use ntpd. [23:25] no, that should not have happened [23:25] yeah, still trying to figure it out [23:25] daff, this seems to be perferred for kvm: http://www.linux-kvm.org/page/KVMClock [23:26] smw_ my systems still have sun java on them [23:26] smw_: may look here for clues: https://wiki.ubuntu.com/LucidLynx/ReleaseNotes#Sun_Java_moved_to_the_Partner_repository [23:26] so a update didn't *remove* it [23:26] daff thanks [23:27] smw_: and then here: https://lists.ubuntu.com/archives/ubuntu-security-announce/2011-December/001528.html [23:29] patdk-lap: our 10.04 KVM VMs all have kvm-clock set as time source, yet without chrony (or probably ntpd) the clock starts drifting wildly. maybe this is better in 12.04 as guest and/or host, like three18ti suggests [23:30] I use vmware, so have all mine set to vmware host time [23:30] haven't had an issue [23:30] did have issues with xen, but the documentation on that is pretty wide [23:31] who knows. seems to me that slowly adjusting the time was a better solution for us than inserting a leap second. the consequences of that event have been quite interesting (looking at java, mysql and big sites like reddit) [23:32] daff: dunno about that - I always run ntpd on my guests [23:32] hehe, wose than the Y2K bug apparently, re: seems to me that slowly adjusting the time was a better solution for us than inserting a leap second [23:32] Tribaal: why not? [23:32] heh, I haven't had any issue [23:33] daff, I think it has more to do with the kernel than KVM as far as the drift appearing to have improved over previous versions of KVM. [23:33] mysql, linux, ... [23:33] I don't really do any java [23:33] Tribaal: y2k was a code issue, the leap second issue was on another ballpark [23:34] RoyK: I think you mean three18ti, not Tribaal, no? :) [23:34] patdk-lap: the bug was mysql only and java, not linux [23:34] royk, there is a linux bug also [23:34] RoyK: this suggests otherwise, no? https://lkml.org/lkml/2012/6/30/122 [23:34] kernel issue [23:34] oh, for leap sec? [23:34] yep [23:35] daff, I think so, royk, I was making a joke I understand it's not a coding issue; what do you mean it was "on another ballpark"?