[00:57] <miceiken> Hey guys
[00:57] <miceiken> when I try to telnet localhost 25 to test postfix i get this error after the "."
[00:57] <miceiken> 421 4.3.0 collect: Cannot write ./dfs1O0umsx031011 (bfcommit, uid=0, gid=110): No such file or directory
[04:13] <Nautilus> I have a 12.04LTS VPS, the fs was in read-only mode so I rebooted. When it was coming back up it detected fs problem(s). I hit the "F" to fix them. Is there a way for me to tell if something is messed up now?
[04:53] <Nautilus> anyone awake that can helkp me with making certs? Per https://help.ubuntu.com/lts/serverguide/certificates-and-security.html#generating-a-csr I've made a self-signed root cert, made a key with a passphrase and am trying to make a key w/o a passphrase for daemons. The message I'm getting is that it must be 4 to 8191 characters.
[04:54] <Nautilus> oh wait, nm, it's asking me the passphrase for the root cert
[04:55] <Nautilus> sorry. that worked of course
[05:58] <cornernote> hi, i have an ubuntu server and it doesnt send cron output to an external email account
[05:58] <cornernote> in "crontab -e" MAILTO=me@my.com
[05:58] <cornernote> but it gets sent to root, not me@my.com
[08:16] <Diceroll> hi... i ve got a problem since the last update on 13.10 : vsftpd won't start anymore... when i launch it manually it said : 500 OOPS: munmap
[08:16] <Diceroll> any clue ?
[08:16] <Diceroll> nothing in /var/log/vsftpd.log
[08:17] <twb> Diceroll: does it say anything interesting before that?
[08:17] <twb> munmap is the opposite of mmap(2) -- a syscall that lets a program slurp a file into memory
[08:17] <Diceroll> nope... at least not where i looked
[08:20] <Diceroll> how can i safely remove the current linux-image and revert to the previous one (i ve only ssh access to the box)
[08:23] <Diceroll> is it ok if i remove with : apt-get purge and then just restart ? i mean is it safe to do that (cause it sounds pretty ugly) ?
[08:26] <twb> Diceroll: if you're using grub, I vaguely remember there is a command like "on the next boot, default to <this>"
[08:26] <twb> You can just purge the current kernel, but make triple sure that grub.cfg ends up with the default you expect
[08:26] <twb> Oh, wait, if you're *running* the kernel at the time, don't remove it.
[08:27] <Diceroll> i do indeed
[08:27] <twb> If you can't find the grub thing, you could also try kexec-tools
[08:27] <Diceroll> ok i'll try the grub thing
[08:27] <twb> (Although lately vendors have been disabling kexec by default -- I don't know if ubuntu has yet.)
[10:42] <jamespage> zul, couple of swift bugfixes - https://code.launchpad.net/~james-page/swift/add-object-expirer/+merge/207883
[12:08] <jamespage> zul, https://code.launchpad.net/~james-page/cinder/hpclients-fixup/+merge/207910
[12:15] <robot_38> test
[13:20] <rabbel> hey guys ... i want to do a preseed install of ubuntu server precise. I would like to do it via USB install. So I'll have to make my own customized iso, right?
[13:20] <rabbel> Any suggestions?
[13:54] <zul> jamespage:  oh you already backported nose-test-config
[13:54] <jamespage> zul, yes - lefthandclient should backport now
[13:54] <zul> jamespage:  k
[13:54] <jamespage> was just waiting for that to build in the ppa
[13:56] <jamespage> zul, you kicked off neutron?
[13:56] <zul> jamespage:  yeah
[13:56]  * jamespage waits for the box to explode
[13:56] <zul> jamespage:  doh
[13:56] <jamespage> load averages at about 18 when running the test suite now
[13:56] <zul> *sigh*
[14:34] <esde> how can i enable rdrand (or check to see if it is enabled) in 12.04? system has i7 3770k processor that shows rdrand flag in /proc/cpuinfo
[14:44] <robot_38> Anyone ever tried Opendedup?
[14:45] <smoser> med_, in general ".X" things are pointless.
[14:45] <smoser> you shoudl just ignore them.
[14:45] <smoser> so i'm not terribly concerned about inconsistencies like that.
[15:30] <hallyn> smb: hi
[15:30] <smb> hallyn, re-hi
[15:30] <hallyn> smb: i did some testing with nested qemu.  My findings:  http://paste.ubuntu.com/6987525/
[15:30] <hallyn> smb: is there an existing kernel bug you have open to track this?
[15:31] <hallyn> maybe I should use https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1278531
[15:32] <smb> or potentially bug 1208455
[15:32] <hallyn> smb: I consider this one very important.  nested kvm worked will in precise.  I'ts been on the decline, and doesn't work in trusty.  Do you have time to look into it, or is this a spare-time thing?  (If you dont' then I"ll dig into kernel causes)
[15:32] <jamespage> coreycb, afternoon
[15:32] <jamespage> re you merge for neutron
[15:32] <hallyn> smb: no, that one seems different,
[15:32] <jamespage> looks generally good - however you needed to target the branch you checked out for the MP
[15:32] <jamespage> lp:~ubuntu-server-dev/neutron/icehouse
[15:32] <hallyn> it may be same root cause in the end, but it's i386-only
[15:32] <jamespage> I resubmitted against the right one for you
[15:32] <hallyn> I'm not saying jodh is less important :)
[15:33] <smb> hallyn, Yeah, you say it hangs
[15:33] <smb> hallyn, Oh wait
[15:33] <coreycb> jamespage, ok thanks I'll take a look
[15:33] <jamespage> coreycb, zul provided a bit of feedback as well re bug tracking
[15:33] <smb> hallyn, 32bit kvm in Trusty kernel has some issue even without second level
[15:34] <smb> hallyn,  bug 1278531
[15:35] <smb> hallyn, too many tabs... I meant 1268906
[15:35] <hallyn> smb: right
[15:35] <hallyn> smb:  I could be wrong, but am claiming that is less important
[15:35] <smb> hallyn, But that would be a host on T issue... So S host
[15:35] <hallyn> smb: all of my testing has been 64-bit only
[15:36] <coreycb> jamespage, ok I'm adding the bug# to the changelog.  thanks for reviewing!
[15:37] <jamespage> coreycb, np
[15:37] <smb> hallyn, Ok, I think you add a new class maybe
[15:37] <zul> hallyn:  so libvirt 1.2.2...its suppose to be out the next week right?
[15:37] <zul> hallyn:  oh wait im not on holiday next week never mind :)
[15:38] <smb> hallyn, Ok, and bug 1278531 was actually for that.
[15:38] <hallyn> smb: eh, it's worse than that
[15:38] <hallyn> zul: lucky you :)
[15:38] <hallyn> smb: so it's actually only qcow2 that has a problem (maybe).  But ubuntu release is irrelevant
[15:38] <hallyn> smb: it's purely to do with kernel version
[15:38] <hallyn> precise host with saucy kernel hangs with 100% cpu just like a saucy host does;  precise host with trusty kernel fails to boot qcow2 just like trusty does
[15:39] <smb> hallyn, Well I would think anything with qcow2 would be in the qemu code...
[15:39] <hallyn> smb: well it may be differnet root cause, but yeah.  ok so i'll track it there.
[15:39] <hallyn> smb: i'll spend some time this week digging.  if you have time to look into it pls keep me uptodate
[15:39] <smb> hallyn, Let us use you bug to put info in
[15:40] <smb> hallyn, *your I meant
[15:40] <smb> hallyn, what kind of system are you using for your tests? amd of intel cpu?
[15:42] <hallyn> smb: intel cpu (vostro laptop).  i originally ran into it on my intel cpu server (precise userspace with trusty kernel)
[15:42] <hallyn> smb: i'll add the info to that bug in a few mins - thx
[15:42] <smb> hallyn, ok.
[15:42] <smb> hallyn, thanks, will then follow up on that
[16:00] <hallyn> smb: noh it's not just qcow2
[16:01] <hallyn> just a few more tests then i'll comment
[16:01] <smb> hallyn, ack
[17:08] <Meatplow> Morning.    Curious how 13.04 is on older power edge 2950
[17:10] <RoyK> Meatplow: better use LTS for servers
[17:10] <RoyK> Meatplow: IIRC precise (12.04) works well on an old 2950
[17:11] <Meatplow> pangolin 12.04 ?
[17:11] <Meatplow> Thanks RoyK
[17:14] <RoyK> Meatplow: that is - don't think I've tried it on 12.04, but running older RHEL/CentOS on similar hardware, so 12.04 should work well
[17:14] <Meatplow> RoyK:  -   that seems to be what I've skimmed in a couple support links. as well
[17:23] <RoyK> Meatplow: I've tried 10.04 on an r300, and it worked well
[17:23] <RoyK> a little newer hardware
[17:23] <RoyK> old hardware is very rarely a problem
[17:23] <RoyK> erm
[17:23] <RoyK> 12.04, not 10.04
[17:24] <Meatplow> cool.   I'm starting install in a few.      I expect it to be straight.
[17:25] <robot_38> Anyone ever tried Opendedup? Is it stable enough for production server?
[17:45] <Lcawte> Hi, I'm having a bit of a problem with my Ubuntu Server install (13.10)... it's connected to my router via ethernet, but when I boot it gets to "waiting for network connection..." and doesn't manage to pick anything up. I haven't changed any settings just rebooted to a newer kernel version (tried the older versions which appear to have the same problem) ... it doesn't to pick up the eth interface
[17:48] <pmatulis_> Lcawte: troubleshoot manually (once booted).  and did you check your cables?
[17:49] <Lcawte> pmatulis_: I've tried switching the cables and ports on the router, all of which work on my laptop running Ubuntu Desktop...
[17:50] <Lcawte> lshw shows the network device correctly, what else can I do to try and debug this?
[17:55] <Tempeszt> Hi folks.  I've got a 12.04LTS server that I'm running Minecraft server on, and I used "at" to start the server, which is normally rather interactive on the console.  now, however, it's running, but it doesn't have a pts associated with it, and I'd like to be able to control it.  any thoughts or advice?
[17:57] <smoser> Tempeszt, i'd suggest maybe running it via screen
[17:57] <Tempeszt> that's what I normally do, and I didn't know that "at" wouldn't start it interactively within the screen where I set it up.
[17:58] <smoser> at wouldn't do that, no
[17:58] <smoser> https://gist.github.com/smoser/1019125
[17:58] <smoser> thats something i wrote some time ago to have jobs (cron, but should work for at) start in a screen
[17:59] <smoser> i think the idea is/was that you could do:
[17:59] <smoser> inscreen --new-if-needed mincraft-screen-session minecraftd whatever-args-here
[18:02] <smoser> Tempeszt, http://askubuntu.com/questions/15244/how-can-i-create-a-cron-command-that-will-execute-a-command-inside-a-detached-sc
[18:03] <Tempeszt> that's neat. I may have to use that in the future and just kill it for now.
[18:03] <pmatulis_> Lcawte: maybe pastebin your interfaces file
[18:04] <marcoceppi_> hey jamespage, how likely is glusterfs MIR to make it for 14.04?
[18:04] <Lcawte> pmatulis_: I think I may have worked it out (did some googling, found out about ethtool and tried a couple of different eth<insert number> interfaces, looks like for some reason it decided to change interface) doing some tests now
[18:05] <smb> hallyn, I updated the bug as well, could you have a look and sanity check whether I understood the setups correctly. I would go and look into those in more detail tomorrow
[18:06] <hallyn> smb: will look, thanks
[18:06] <jamespage> marcoceppi_, that depends entirely on the resourcing in the MIR team
[18:06] <pmatulis_> Lcawte: good stuff
[18:06] <marcoceppi_> jamespage: last the bug said, it needed a security review, so I'm not sure who to ping for status on that
[18:07] <jamespage> jdstrand, do you have a likely timeframe around reviewing MIR bug 1274247?
[18:07] <jamespage> bug 1274247
[18:07] <jamespage> bot?
[18:07] <jamespage> grrrrr
[18:07] <jamespage> https://bugs.launchpad.net/ubuntu/+source/glusterfs/+bug/1274247
[18:09] <jdstrand> oh, I missed that assignment
[18:10] <jdstrand> I'm going to assign to sarnold for the time being-- he is busy finishing up an apparmor upload and then I think he can get back to code reviews
[18:10] <jdstrand> sarnold: let me know if that's not the case
[18:10] <jdstrand> sarnold: ^
[18:11] <sarnold> I do'nt think I'll get to glusterfs this week.. there's already four (I think? :) MIRs ahead in line..
[18:11] <hallyn> jstrand: bug 1278531, I'm hoping (but not confident) that will also solve your qemu problem
[18:11] <sarnold> and I'm skeptical of glusterfs, the handful of folks I've known who've given it a shot tended to give up and headed towards lustre, last I knew..
[18:13] <jamespage> sarnold, quite likely
[18:13] <semiosis> sarnold: this gives me confidence
[18:13] <semiosis> ...not
[18:14] <jamespage> jdstrand, sarnold: hopefully one of you has juju-core + deps on your list as well
[18:14] <sarnold> yikes :)
[18:14] <jamespage> https://bugs.launchpad.net/ubuntu/+source/juju-core/+bug/1267393
[18:14] <jamespage> sarnold, that includes some mongodb fun as well
[18:15] <semiosis> sarnold: please keep an open mind toward gluster.  it's come a long long way in the last couple years and plenty of people are using it on ubuntu
[18:15] <semiosis> sarnold: i'm the upstream deb packager btw, so if i can be of any help please ping me
[18:16] <sarnold> semiosis: cool, thanks! :)
[18:16] <semiosis> my pleasure
[18:26] <zul> jamespage:  mir filed for python-hplefthandclient and python-nose-testconfig
[18:26] <jamespage> zul, +1 thanks
[21:02] <hazmat> anyone played with http://www.ktap.org/?
[22:13] <hallyn> cyphermox: stgraber: bug 1205086, does removing --strict-order seem like the right thing to do?
[22:16] <cyphermox> hallyn: well, it will likely fix the problem for the reporter
[22:16] <cyphermox> hallyn: not sure if there is an actual reason why --strict-order is set there
[22:16] <hallyn> will it cause problems for anyone else
[22:17] <hallyn> i'm nto sure either
[22:17] <cyphermox> assuming stgraber is the one who set this up the way it is, I'd wait for his answer :)
[22:18] <stgraber> from what I remember, strict-order makes dnsmasq query the first server, on failure (NXDOMAIN isn't a failure), it'll then fallback to the next, etc...
[22:18] <stgraber> removing strict-order, makes dnsmasq query all servers
[22:18] <stgraber> which for most people would mean doubling or tripling the DNS traffic
[22:19] <hallyn> really i just don't understand why there isn't a "ping the other servers if first doesn't have an entry for this" option
[22:19] <hallyn> removing strict-orer makes it random order right?
[22:19] <hallyn> not query-all
[22:19] <stgraber> it makes it query all
[22:20] <stgraber> unless that changed, dnsmasq changed quite a bit since we dealt with that stuff in 12.04
[22:21] <hallyn> In comment #2 (point 2) in that bug, i was going based on documentation i think.
[22:22] <cyphermox> my bofh brain says the right fix is for that VPN dns to become recursive...
[22:23] <stgraber> right, reading current dnsmasq manpage, dropping strict-order should be fine, however this wouldn't solve this bug
[22:23] <stgraber> it'd just make it random instead of reliably wrong
[22:23] <hallyn> well i did think comment #4 was "the right fix"
[22:23] <hallyn> right
[22:23] <hallyn> though he claims it did fix it for him.  <shrug>
[22:24] <stgraber> strict-order guarantees the dns servers are processed the same way the libc would, without it, the servers are taken in random order, but if the first queries server returns NXDOMAIN the request will still fail
[22:24] <stgraber> well, it could be that the libc resolver attempt the request twice or something along those lines
[22:24] <stgraber> and by chance the order is sufficiently random that trying twice is enough?
[22:25] <hallyn> which is NOT sufficient to make that change to the package :)
[22:26] <stgraber> right. If the reporter is using a desktop machine, the real fix is to use NetworkManager which will properly setup dnsmasq to only use the VPN dns server for requests relevant to it
[22:30] <stgraber> hallyn: ok, so just did some tests. The problem there is clearly that the remote dns server is misconfigured. Trying with mine, I get NXDOMAIN for an invalid domain from a recursive server (as I should) but SERVFAIL for a domain outside the scope of a non-recursive server.
[22:30] <stgraber> SERVFAIL causes dnsmasq to query the next server, NXDOMAIN doesn't.
[22:31] <stgraber> SERVFAIL is nsd's response when non-recursive. REFUSED is bind's response when non-recursive. Both work with dnsmasq.
[22:32] <hallyn> so he has a non-recursive dnsd responding as though it was recursive?
[22:32] <stgraber> if he's getting NXDOMAIN, yes
[22:37] <stgraber> it looks like there's no way of properly turning off recursion in dnsmasq, no-resolv only prevents it from getting upstream servers (so it'll rightfully return NXDOMAIN) and dns-forward-max is meant as a tweak for long resolving chains, not as a way to block recursion
[22:37] <stgraber> so my suggestion there is, don't use dnsmasq