[02:17] <tohuw> Where can I find the authoritative documentation on creating a service in Ubuntu Server 12.04? Related: should I be creating the service as an "upstart" service? This isn't clear to me.
[02:19] <sarnold> tohuw: hopefully useful: http://upstart.ubuntu.com/cookbook/
[02:19] <sarnold> I don't know if it is The Definitive Source, but it sure is useful :)
[02:21] <tohuw> sarnold: Thank you, I'll read through this.
[02:21] <sarnold> SpamapS: hey, your @canonical.com is still on the header at http://upstart.ubuntu.com/cookbook/   :)
[02:53] <SpamapS> sarnold: there's a bug link at the bottom... ;-)
[02:53] <sarnold> SpamapS :)
[05:54] <Senor> I am trying to deploy one web server cluster , can I put the shared mem-cache on one single server ?
[09:16] <Daviey> jamespage: Hey, do you fancy merging the nova and keystone security uploads into the pending SRU ones.. or shall we wait for adam_g ?
[09:16] <jamespage> Daviey, I'll take a look in a bit
[09:17] <Daviey> jamespage: ok, we should push out the sec uploads directly to the UCA.
[09:18] <jamespage> Daviey, agreed
[09:26] <rurufufuss> hi guys, I can't seem to do ./somelocalbinary even though its in my path
[09:31]  * koolhead17 pokes Daviey 
[09:38] <rurufufuss> anyone?
[09:41] <jamespage> Daviey, keystone point release already includes the security fix
[09:41] <jamespage> so do I just need to document that in the changelog? I guess so
[09:42] <Daviey> jamespage: Probably fair to merge changelogs and make comment to it, without using valid LP synatx.
[09:43] <Daviey> so, SRU orig changelog, jdstrand's then your brief message with LP: FOO ?
[09:44] <jamespage> Daviey, okay
[09:47] <jamespage> Daviey, "Make sure both public and admin API work with 1-way ipv6 & SSL. ... SKIP: Eventlet doesn't support IPv6, lp 1176204"
[09:47] <jamespage> thats going to be an issue
[09:48] <jamespage> (for our IPv6 testing plans that is)
[09:48] <Daviey> jamespage: I'm not sure it is
[09:48] <Daviey> I saw that go past, and discounted it as we are focusing on floating ip ipv6 this cycle, rught?
[09:48] <Daviey> right*
[09:48] <jamespage> Daviey, maybe
[09:49] <jamespage> yeah - I guess thats enough
[09:49] <Daviey> jamespage: There is a partial patch for eventlet ipv6
[09:49] <Daviey> it was on bitbucket, but they moved to github(? or somewhere), and lost the bug history
[09:53] <jamespage> Daviey, moved on github somewhere
[09:54] <Daviey> jamespage: can you review this please? Folsom CA + Security update, http://pb.daviey.com/7sc7/  (quantal-updates -> folsom CA, http://pb.daviey.com/2yV8/)
[09:56] <jamespage> Daviey, so the security update will pull in the new point release as well?
[09:56] <jamespage> is that what we did in distro?
[09:56] <Daviey> and for giggles, *.changes http://pb.daviey.com/tNAr/
[09:56] <Daviey> jamespage: No
[09:57] <jamespage> Daviey, duh - yeah - sorry getting my point releases confused
[09:57] <Daviey> yeah, it's a little confusing - which is why i wanted a double check :)
[09:57] <jamespage> Daviey, the changes needs -v2012.2.3-0ubuntu2 I think
[09:57] <Daviey> hm
[09:58] <Daviey> Ah yes
[09:58] <Daviey> good catch
[09:59] <jamespage> omg the keystone test suite takes some time
[10:07] <koolhead17> jamespage: what is missing? Some new dep got added :P
[10:07] <jamespage> koolhead17, nothings missing I think
[10:08] <koolhead17> jamespage: happy to hear that :)
[10:09] <Daviey> jamespage: pushed to -staging, when it's published - can you review into -proposed & fire off a jenkins job?
[10:09] <Daviey> (keystone & nova for folsom and grizzly)
[10:10] <Daviey> Oh curse.
[10:10] <Daviey> keystone_2013.1-0ubuntu1.1~cloud0.dsc: Version older than that in the archive. 1:2013.1-0ubuntu1.1~cloud0 <= 1:2013.1.1-0ubuntu1~cloud0
[10:11] <jamespage> Daviey, yeah
[10:11] <jamespage> I thought that might happen
[10:11] <jamespage> you'll need to drop it from the PPA first
[10:11] <Daviey> just did
[10:12] <jamespage> Daviey, the reprepro sync from the proposed and staging ppa's will work OK with this right?
[10:13] <jamespage> (I think it will - reprepro does not work quite like PPA's)
[10:14] <jamespage> Daviey, OK - I've uploaded a rebased keystone for raring; just testing nova now
[10:15] <Daviey> jamespage: it will at least work to -updates
[10:15] <Daviey> as it isn't published there
[10:17] <Daviey> jamespage: (we might need to go straight to updates, is what i think you are saying)
[10:17] <jamespage> Daviey, I think it will be OK via proposed
[10:19] <jamespage> Daviey, hmm - I just spotted a potential issue
[10:19] <jamespage> not sure how much of a problem it actually is
[10:19] <jamespage> Daviey, we don't push the new point releases into dev prior to SRU
[10:19] <jamespage> which I think makes sense (as dev will shift to havana)
[10:20] <jamespage> but we are in that pre-first milestone phase with upstream right now
[10:20] <Daviey> jamespage: I am not terribly concerned by that TBH
[10:20] <jamespage> Daviey, OK - nova/raring also rebased and re-uploaded
[10:20] <Daviey> jamespage: if you want to upload it there aswell, i wouldn'tgrumble
[10:21] <jamespage> Daviey, I'll do the same for folsom now
[10:21] <Daviey> jamespage: where did you upload?
[10:21] <jamespage> Daviey, raring-proposed
[10:21] <Daviey> ok, cool
[10:22] <Daviey> jamespage: Security uploads re-done to staging
[10:22] <jamespage> Daviey, okies
[10:37] <jamespage> Daviey, keystone/quantal-proposed uploaded; just waiting on nova/quantal tests to complete
[10:39] <jamespage> Daviey, I think we need a less awkard way to deal with security fixes into the UCA
[10:39] <jamespage> Daviey, like the security team do we should have XXX-security PPA -> XXX-updates
[10:39] <jamespage> bypassing -proposed and avoiding all this deleting package non-sense
[10:44] <Daviey> jamespage: that sounds a good idea
[10:44] <Daviey> jamespage: preparing the updates is quite gentlemanly now
[10:45] <Daviey> jamespage: folsom-security-staging ?
[10:45] <jamespage> Daviey, yeah
[10:47] <jamespage> Daviey, OK - nova/quantal-proposed uploaded as well
[10:47] <jamespage> I updated the lab branches with the revised changelogs
[10:49] <jamespage> yolanda, care to review -https://code.launchpad.net/~james-page/ceilometer/refresh-patches/+merge/164341
[10:49] <jamespage> ta
[10:49] <yolanda> sure
[10:51] <yolanda> jamespage, we will be adding mongo as default?
[10:51] <jamespage> yolanda, no
[10:52] <jamespage> Daviey, OK - this sucks as for some reason the tooling is still seeing the old source packages in the staging PPA
[10:53] <Daviey> jamespage: might need to wait for the publisher?
[10:53] <jamespage> Daviey, maybe
[10:54] <Daviey> +----------+-----------------------------------------------------+----------+
[10:54] <Daviey> | package  | staging                                             | proposed |
[10:54] <Daviey> +----------+-----------------------------------------------------+----------+
[10:54] <Daviey> | nova     | 2012.2.3-0ubuntu2.1~cloud0                          | None     |
[10:54] <Daviey> | keystone | 2012.2.3+stable-20130206-82c87e56-0ubuntu2.1~cloud0 | None     |
[10:54] <Daviey> looks right to me?
[10:55] <jamespage> Daviey, the grid is OK but when I query the changelog it brings back the most recent versioned one
[10:55] <Daviey> ah
[10:55] <Daviey> indeed
[10:58] <Daviey> jamespage: I suspect this is a bug in LP tbh.
[10:59] <jamespage> Daviey, I think its just the changelog retrieve bit; the copy-package call under the hood appears todo the right things
[11:00] <jamespage> Daviey, trying now - it will send the wrong email content but I can live with that for the time being
[11:01] <Daviey> jamespage: hmm
[11:02] <Daviey> wait a min
[11:02] <jamespage> Daviey, I think I see the problem in the code as well
[11:02] <Daviey> jamespage: It grabs the Publication by name
[11:03] <jamespage> Daviey, yeah - it does not include the version
[11:03] <jamespage> [0] is latest normally - which in this case is incorrect
[11:03] <Daviey> yeah
[11:04] <Daviey> We know the version, so matching on version should do the trick... i'd think
[11:07] <Daviey> jamespage: I'd argue that LP isn't DTRT here, as the original version is no longer Published.
[11:07] <jamespage> Daviey, OK -  I have a fix for the tool
[11:07] <Daviey> super
[11:07] <jamespage> version=XXX on getPublishedSources
[11:07] <jamespage> I'll sync up the grizzly security fixes to proposed now
[11:11] <jamespage> Daviey, okies - that is working
[11:11] <jamespage> tooling pushed
[11:11] <jamespage> I also added the security-staging -> updates workflow
[11:13] <jamespage> Daviey, hmm - lab seems unhappy
[12:32] <Daviey> jamespage: unhappy?
[12:41] <bjf> jamespage, bug 1164739
[12:42] <bjf> jamespage, is that something you could verify as fixed?
[12:42] <bjf> smoser, ^ ?  (i see you have comments on that bug)
[12:43] <ak5> hi guys, do you know how I can get mod_proxy_fcgi for apache?
[12:43] <smoser> bjf, i'll get that done.
[12:43] <bjf> smoser, awesome
[12:44] <TheLordOfTime> ak5:  i believe there's a package to install it...
[12:44] <TheLordOfTime> but i might be wrong
[12:45] <ak5> TheLordOfTime: I see, I just reralized I am in apache 2.2
[12:45] <TheLordOfTime> that might be the problem
[12:45] <ak5> are there repos for 2.4 for 12.04 ?
[12:47] <TheLordOfTime> ak5:  i dunno, but i don't use apache so... xD
[12:47] <ak5> nginx then?
[12:47] <TheLordOfTime> yep i'm an nginx user :P
[12:48] <TheLordOfTime> ... that reminds me... note to  self: push patch to nginx devel branch ppa...
[12:49] <TheLordOfTime> ak5:  you could always search "apache 2.4 ppa ubuntu" and see what comes up on google
[12:49] <TheLordOfTime> you might find a PPA for it.
[12:49] <TheLordOfTime> but...
[12:49] <TheLordOfTime> ~ppa
[12:49] <TheLordOfTime> erm
[12:49] <TheLordOfTime> !ppa
[12:49] <TheLordOfTime> ^ that
[12:49] <TheLordOfTime> shut up uvirtbot
[12:49] <jamespage> Daviey, dunno - I know adam_g and smoser where working on some sort of issue yesterday
[12:50] <ak5> http://ubuntuforums.org/showthread.php?t=2011603 <- this doesn't look very promising
[12:50] <Daviey> jamespage: ah ok
[12:50] <Daviey> they should be here soon
[12:50] <smoser> jamespage, we're closer
[12:55] <TheLordOfTime> Daviey:  about the nginx support thing, would the mailing list be a good place to start the discussion, or should I wait until Tuesday?
[12:55] <TheLordOfTime> i ask because Tuesday's election day where I am, and I might be busy
[12:55] <TheLordOfTime> (and might miss the server team meeting :/)
[12:56] <ak5> How do backports work? I uncommented them in sources.list, now I apt-get update'd
[12:56] <ak5> now will it just install all kinds of backports on upgrade?
[12:56] <ak5> or do I need to choose specific ones
[12:56] <ak5> /s/need/can/
[12:56] <TheLordOfTime> ak5:  what version of ubuntu are you on
[12:56] <ak5> 12.04
[12:58] <ak5> TheLordOfTime: ^
[12:58] <TheLordOfTime> i believe the policy is written that unless you specify the package from backports it won't install from backports
[12:58] <TheLordOfTime> but i can't be certain
[12:58] <TheLordOfTime> but i'm not certain, since if I want something from backports i usually dump it into a PPA of my own
[13:00] <ak5> TheLordOfTime: ok, then how do I search backports?
[13:01] <TheLordOfTime> ak5:  if you're looking for backported apache 2.4 there is none
[13:01] <ak5> TheLordOfTime: jesus
[13:01] <ak5> from source it is I guess
[13:01] <TheLordOfTime> ak5:  not everything get's backported
[13:01] <TheLordOfTime> ;P
[13:01] <ak5> this seems like so big of a deal
[13:03] <rbasak> apache 2.4 isn't in Ubuntu at all. In Debian it's in experimental.
[13:04] <rbasak> Presumably because of the wheezy freeze.
[13:05] <rbasak> When was 2.4 released as stable upstream? I wonder if we should sync 2.4 from experimental, but it's quite an involved package, there may be upgrade path issues and I'm not sure it's worth going out on a limb for that.
[13:06] <rbasak> Oh and I guess we have a delta to forward port too.
[13:11] <patdk-wk__> well, in itself, using the debian package worked well for upgrade
[13:12] <patdk-wk__> but attempting to sync in all the other packages, like php and all the modules, become a real pain
[13:12] <patdk-wk__> I put 2.4 in a ppa for me to play with over a year ago
[13:21] <Daviey> TheLordOfTime: mailing list is always a good place :)
[13:22] <patdk-wk__> hmm, I have apache 2.4 for 12.04 in my ppa :)
[13:22] <patdk-wk__> the real question is, why are you looking for apache 2.4?
[13:23] <patdk-wk__> if it's for pci compliance, your solving this the wrong way
[13:23] <patdk-wk__> I have had soo many emails asking about my apache 2.4 ppa cause of people attempting to solve pci issues
[13:29] <rbasak> patdk-wk__: that's interesting. Why does using < 2.4 cause people PCI concerns?
[13:32] <patdk-wk__> it shouldn't :)
[13:32] <patdk-wk__> the issue is, the pci compliance companies do scans
[13:32] <patdk-wk__> and to be quick and efficient, they only check the version of the software in use
[13:33] <patdk-wk__> they don't check if you patch it for known CVE's
[13:33] <patdk-wk__> therefor you always fail, unless you have the latest version, and you have to file a report to them, showing the ubuntu CVE patchs made to your version
[13:34] <patdk-wk__> so when they get the failure report, and it says, fixed in apache 2.4, they attempt to install that :) instead of knowning that the version they have is already fixed
[13:41] <rbasak> Oh I see - thanks
[13:45] <zul> uh we just had an earthquake tremors
[13:47] <rbasak> zul: nothing on http://earthquake.usgs.gov/earthquakes/map/
[13:48] <zul> rbasak:  dude my house was shaking :)
[13:48] <zul> like 2 minutes ago
[13:49] <rbasak> Computer says no :-P
[13:50] <patdk-wk__> zul, could easily be mining operations
[13:50] <patdk-wk__> road construction, ...
[13:50] <zul> rbasak:  https://twitter.com/CANADAquakes
[13:50] <zul> no it was an earthquake
[13:51] <zul> Automatic detection of seismic event: magnitude 4.8 - 17 May 9:43 EDT - BRAESIDE, ON region
[14:20] <vlad_sta_> Hello everyone.
[14:20] <vlad_sta_> Need emergency help with recovering RAID1 on Ubuntu 12.04
[14:21] <vlad_sta_> Having 2 RAID1: md0 and md1
[14:22] <vlad_sta_> 1 disk was down few hours ago
[14:22] <vlad_sta_> The file system became read-only
[14:22] <vlad_sta_> cat /proc/mdstat
[14:22] <vlad_sta_> Personalities : [raid1] [linear] [multipath] [raid0] [raid6] [raid5] [raid4] [raid10]
[14:22] <vlad_sta_> md0 : active raid1 sdb1[1]
[14:22] <vlad_sta_>       975296 blocks super 1.2 [2/1] [_U]
[14:22] <vlad_sta_>       
[14:22] <vlad_sta_> md1 : active raid1 sdb2[1]
[14:22] <vlad_sta_>       975653696 blocks super 1.2 [2/1] [_U]
[14:22] <vlad_sta_>       
[14:22] <vlad_sta_> unused devices: <none>
[14:23] <vlad_sta_> then I made  mdadm --assemble -scan -v
[14:23] <vlad_sta_> and mdadm -A --scan
[14:23] <vlad_sta_> after that
[14:23] <vlad_sta_> cat /proc/mdstat
[14:23] <vlad_sta_> Personalities : [raid1] [linear] [multipath] [raid0] [raid6] [raid5] [raid4] [raid10]
[14:23] <vlad_sta_> md126 : active raid1 sda1[0]
[14:23] <vlad_sta_>       975296 blocks super 1.2 [2/1] [U_]
[14:23] <vlad_sta_>       
[14:23] <vlad_sta_> md127 : active raid1 sda2[0]
[14:23] <vlad_sta_>       975653696 blocks super 1.2 [2/1] [U_]
[14:23] <vlad_sta_>       
[14:23] <vlad_sta_> md0 : active raid1 sdb1[1]
[14:23] <vlad_sta_>       975296 blocks super 1.2 [2/1] [_U]
[14:23] <vlad_sta_>       
[14:23] <vlad_sta_> md1 : active raid1 sdb2[1]
[14:23] <vlad_sta_>       975653696 blocks super 1.2 [2/1] [_U]
[14:23] <vlad_sta_>       
[14:24] <vlad_sta_> unused devices: <none>
[14:24] <vlad_sta_> Where the md126 and md127 appeared from?
[14:24] <vlad_sta_> This is not my raid as I only had md0 and md1
[14:24] <RoyK> !pastebin | vlad_sta_
[14:25] <vlad_sta_> Sure. Sorry about that
[14:26] <RoyK> vlad_sta_: doesn't look too good
[14:26] <RoyK> I'd say, first run smartctl -x on /dev/sda and /dev/sdb
[14:26] <RoyK> pastebin that
[14:26] <RoyK> !pastebinit | vlad_sta_
[14:27] <RoyK> vlad_sta_: looks like it's rather confused - md126-127 is apparently the other side of the mirror
[14:27] <vlad_sta_> http://pastebin.com/mLgtWEvD
[14:28] <vlad_sta_> RoyK: Could you guide me a bit about how serious is this problem and what are the first aid here?
[14:29] <RoyK> vlad_sta_: check the drives first. something has happened. you shouldn't just assemble a raid if a drive is kicked out
[14:29] <RoyK> btw, pastebin dmesg output
[14:29] <vlad_sta_> smartctl -x /dev/sda
[14:29] <vlad_sta_> bash: /usr/sbin/smartctl: Input/output error
[14:30] <rbasak> jamespage: bug 1181135 - looks like he's using the quantal HWE stack on Precise. Is our position that he needs to downgrade to the 3.2 kernel if he wants to build open-vm-dkms?
[14:30] <vlad_sta_> dmesg
[14:30] <vlad_sta_> Traceback (most recent call last):
[14:30] <vlad_sta_>   File "/usr/lib/command-not-found", line 21, in <module>
[14:30] <vlad_sta_>     from CommandNotFound.util import crash_guard
[14:30] <vlad_sta_> ImportError: No module named CommandNotFound.util
[14:30] <vlad_sta_> Error in sys.excepthook:
[14:30] <vlad_sta_> Traceback (most recent call last):
[14:30] <vlad_sta_>   File "/usr/lib/python2.7/dist-packages/apport_python_hook.py", line 66, in apport_excepthook
[14:30] <vlad_sta_>     from apport.fileutils import likely_packaged, get_recent_crashes
[14:30] <vlad_sta_> ImportError: No module named apport.fileutils
[14:30] <vlad_sta_> Original exception was:
[14:30] <RoyK> !pastebin | vlad_sta_
[14:30] <vlad_sta_> Traceback (most recent call last):
[14:30] <vlad_sta_>   File "/usr/lib/command-not-found", line 21, in <module>
[14:30] <vlad_sta_>     from CommandNotFound.util import crash_guard
[14:30] <RoyK> vlad_sta_: stop spamming!
[14:30] <vlad_sta_> ImportError: No module named CommandNotFound.util
[14:31] <vlad_sta_> oops too much
[14:31] <vlad_sta_> RoyK: absolutely sorry about that
[14:31] <RoyK> vlad_sta_: generally, more than 2 lines should go to pastebin
[14:31] <RoyK> I guess you'll have to reboot that thing
[14:31] <RoyK> try 'reboot'
[14:31] <vlad_sta_> RoyK: I got it first time you pointed me. This time was a mistake
[14:32] <vlad_sta_> reboot
[14:32] <vlad_sta_> bash: /sbin/reboot: Input/output error
[14:32] <RoyK> if that doesn't work, press and hold alt+sysrq (print screen) and type S U B while holding
[14:32] <vlad_sta_> RoyK: I have remote KVM there, but I'm on Mac and don't know where is sysrq
[14:32] <rbasak> jamespage: it might be an idea to document this on https://wiki.ubuntu.com/Kernel/LTSEnablementStack too
[14:33] <RoyK> vlad_sta_: then 'echo s > /proc/sysrq-trigger'
[14:33] <RoyK> then echo u .. then echo b
[14:33] <RoyK> as above
[14:33] <disposable> does anybody run a mdraid-1 array of Samsung 840PRO? what speeds are you getting?
[14:34] <RoyK> vlad_sta_: in the state it's now, echo b > /proc/sysrq-trigger should do
[14:34] <vlad_sta_> I made remotely by ssh echo s > /proc/sysrq-trigger && echo u > /proc/sysrq-trigger && echo b > /proc/sysrq-trigger
[14:35] <vlad_sta_> after that I lost connection to the server
[14:35] <RoyK> yes
[14:35] <vlad_sta_> oh rebooting
[14:35] <vlad_sta_> fine
[14:35] <RoyK> b is reboot
[14:35] <vlad_sta_> RoyK: this magic is awesome. Don't know what it was
[14:37] <RoyK> vlad_sta_: http://en.wikipedia.org/wiki/Magic_SysRq_key
[14:39] <RoyK> vlad_sta_: so, did it boot correctly_
[14:39] <RoyK> ?
[14:39] <vlad_sta_> No
[14:39] <vlad_sta_> How can I paste image here?
[14:39] <RoyK> !imgpaste
[14:39] <RoyK> hm. dunno
[14:40] <RoyK> !screenshots | vlad_sta_
[14:40] <vlad_sta_> http://cl.ly/image/1W1k2v3e2O1e
[14:41] <RoyK> try rebooting into single
[14:41] <vlad_sta_> http://cl.ly/image/1r0r2V432u3o
[14:41] <RoyK> type exit
[14:42] <RoyK> default config in ubuntu is for some moronic reason not to boot correctly if a raid is degraded
[14:42] <vlad_sta_> To be clear, config is 2 HDD + RAID1 + encryption + LVM
[14:42] <RoyK> shouldn't matter
[14:43] <vlad_sta_> RoyK: the reason is that raid is encrypted. I had the same result when I was installing this server and testing how raid works. I set explicitly to load in degraded mode, but 12.04 LTS has a kinda bug with that (as read in forums)
[14:44] <vlad_sta_> RoyK: rebooting in single (recovery) mode
[14:44] <RoyK> works for me, though
[14:44] <vlad_sta_> RoyK: what version is your ubuntu?
[14:44] <RoyK> lots :)
[14:45] <vlad_sta_> Lucky you then :-)
[14:46] <RoyK> but typing exit from the recovery prompt should make it boot normally
[14:46] <vlad_sta_> the same thing
[14:46] <vlad_sta_> it waiting for encryoted and busybox then
[14:47] <vlad_sta_> while loading it showed error reading ata2 drive
[14:47] <vlad_sta_> I'll reboot and try to screenshot it
[14:47] <RoyK> probably a bad drive, then
[14:47] <vlad_sta_> RoyK: I think it could be that
[14:48] <RoyK> well, that's why you're using raid, isn't it? ;)
[14:48] <vlad_sta_> Sure!
[14:48] <RoyK> anyway - whatever you have on top of the md raid shouldn't matter
[14:49] <vlad_sta_> http://cl.ly/image/2Q2O2i2w113a
[14:50] <vlad_sta_> RoyK: ok. I'm just newbe to Raid and just started digging deeper inside Linux (reading huge book now)
[14:50] <RoyK> ok
[14:50] <RoyK> for linux raid specific stuff, it's sometimes easier to get good answers on #linux-raid
[14:51] <vlad_sta_> RoyK: Thanks a lot!
[14:52] <vlad_sta_> Could you recommend some other good channels on linux and related?
[14:52] <RoyK> not sure - it all depends what you want to know. there are hundreds
[14:53] <jamespage> rbasak, otp right now - give me 10 mins
[14:53] <rbasak> np
[14:53] <vlad_sta_> RoyK: I'm not ready to answer this question right now
[14:53] <vlad_sta_> RoyK: so I have to go to obtain new hdd and then goto data-center...
[14:54] <RoyK> it should be possible to boot the system with one drive in that mirror
[14:54] <vlad_sta_> RoyK: so having looking to the last boot screen, what do you think had happen?
[14:55] <RoyK> just did a wee test on a raid test vm on precise (12.04), set boot_degraded=on and removed a drive, dd'ed a bunch of zeros on it and rebooted
[14:55] <RoyK> came up
[14:55] <vlad_sta_> RoyK: If it possible it would be awesome. Currently the system falls down to busybox (initramfs)
[14:55] <RoyK> type exit
[14:55] <RoyK> it should continue
[14:59] <vlad_starkov> oops timeout
[14:59] <vlad_starkov> RoyK: http://cl.ly/image/1B1U42450j2T
[14:59] <resno> im tryin to write a bash script to capture the hostname - HOSTNAME=`hostname` isnt working
[14:59] <resno> i also tried "`hostname`"
[15:02] <RoyK> resno: just tried HOSTNAME=`hostname`; echo $HOSTNAME
[15:02] <RoyK> works for me...
[15:02] <resno> hmm, then ive done something else wrong
[15:02] <hXm> i have many /usr/sbin/apache2 -k start process, are normal?
[15:03] <RoyK> hXm: no
[15:03] <RoyK> hXm: pastebin 'ps axf'
[15:04] <hXm> http://pastebin.com/cBPUVDbx
[15:04] <RoyK> oh
[15:04] <hXm> the tclsh script is allowed, the SCREEN too
[15:04] <RoyK> those are ok
[15:05] <RoyK> it's just apache starting a some listening processes
[15:05] <hXm> ah
[15:05] <resno> RoyK: i missed typed the variable lol
[15:05] <resno> mistyped*
[15:05] <RoyK> ;)
[15:05] <hXm> good, thanks
[15:05] <RoyK> hXm: apache normally starts a few processes to balance incoming requests
[15:06] <hXm> ah ok, so i dont scare
[15:06] <RoyK> if traffic grows, the number of processes will grow, and then lessen again if traffic decreases
[15:11] <resno> as long as apache isnt going "crazy"
[15:11] <RoyK> resno: never seen that ;)
[15:12] <resno> havent seen it today ;)
[15:12] <RoyK> resno: what do you mean "go crazy"?
[15:13] <resno> when to many people hit it and it makes more processes then the machine can handle
[15:13] <resno> default installation with little tinkering of workers
[15:13] <resno> and the machine just doesnt kill them off as it should
[15:17] <RoyK> resno: if you have low memory, reduce MaxClients in the apache config
[15:17] <RoyK> resno: it's not apache's fault, it's the admin's fault, as usual ;)
[15:17] <resno> but of course
[15:18] <resno> im a admin learning as i go :)
[15:18] <resno> i've tested my setups enough to realize a few tweaks to make apache not kill a server
[15:18] <mndo> hi, does kvm support cpu hotplug?
[15:19] <RoyK> mndo: afaik, no
[15:19] <RoyK> mndo: there's been some talk about it, but I don't think there's anything there yet. try #virt @ OFTC.net
[15:19] <patdk-wk__> I can't remember if the linux kernel supports cpu hotplug even
[15:20] <RoyK> patdk-wk__: it's been supporting cpu hotplug for a decade or so
[15:20] <RoyK> http://www.linux-kvm.org/page/CPUHotPlug
[15:21] <patdk-wk__> ah, like memory, still need to online them manually
[15:22] <RoyK> should be doable with some kvm guest utils
[15:22] <patdk-wk__> I wonder if a udev rule could be made to bring hotplug memory/cpu online automatically
[16:01] <patdk-wk__> ok, udev rules work to online cpu/memory on add, seems to work good, as long as the memory hotplug module is loaded before you add ram
[16:06] <krababbel> Hi, can I upgrade using apt-get upgrade, so that also recommended packages are installed? They are left out and --install-recommends won't work on upgrade it seems.
[16:06] <Pici> krababbel: Why do you think packages are being left out?
[16:07] <patdk-wk__> krababbel, no you can't
[16:07] <patdk-wk__> the whole point of upgrade is NOT to install new packages
[16:07] <krababbel> Pici: I tried aptitude before, but didn't install, just looked in the list it would install. I showed packages which would be recommended after the upgrade, which would not be installed by an apt-get upgrade.
[16:08] <krababbel> patdk-wk__: OK, I see.
[16:08] <patdk-wk__> you need dist-upgrade to install new stuff, but not sure if you will have the *same* issue still
[16:08] <Pici> krababbel: 'upgrade' will never install new packages, only new versions for existing packages.
[16:08] <patdk-wk__> upgrade will only upgrade things that don't need new items :)
[16:08] <krababbel> Thanks, I will try dist-upgrade too.
[16:08] <patdk-wk__> I don't believe dist-upgrade will, but it solves half the issue :)
[16:09] <patdk-wk__> I purposely don't install recommended packages for many things, cause I don't want them
[16:09] <patdk-wk__> and dist-upgrade doesn't bring them in
[16:09] <patdk-wk__> but probably an option to do that
[16:09] <Pici> You can just re-do an install and include --install-recommends there.
[16:10] <krababbel> OK, thank you for your help. :)
[16:16] <jamespage> rbasak, OK
[16:21] <jamespage> rbasak, there are lot of those on 12.04 with the enablement kernel
[16:21] <jamespage> (s)
[16:21] <jamespage> the DKMS modules package should be fixed to work with the newer versions - that what we did with ovs and iscsitarget
[16:29] <hallyn> stgraber: if you're not playing with the staging tree right now, i coudl quickly revert, add your ack, and push --force
[16:30] <stgraber> hallyn: not playing with it, so if you want to add the ack, go ahead
[16:31] <hallyn> k done
[16:33] <olivier_bK> hi
[16:38] <rbasak> jamespage: OK, thanks. Presumably using the original kernel, if possible, is a suitable workaround then?
[16:38] <jamespage> yes
[16:38] <rbasak> Got it.
[16:38] <jamespage> I think there is a segment of users who want todo this sort of stuff who don't really need the HWE kernels
[16:39] <jamespage> they should stick on 3.2 IMHO
[16:48] <patdk-wk__> I have attempted to upgrade to the newer kernel on a few vm's, so that I could get the newer network stack, as it helps a lot
[16:49] <patdk-wk__> but open-vm-tools dkms breaks badly doing that
[16:49] <patdk-wk__> haven't revisited since that failed attempt
[16:54] <rbasak> jamespage: some discussion of DKMS and HWE in #ubuntu-kernel atm
[16:56] <rbasak> jamespage: http://paste.ubuntu.com/5674611/
[16:57] <rbasak> jamespage: 17:57 <rbasak> openvswitch is probably the module that we care about the most. I'll ask jamespage about it.
[17:07] <zul> adam_g: derp https://code.launchpad.net/~zulcss/cinder/cinder-ftbfs-pbr/+merge/164464
[17:10] <rbasak> jamespage: you should probably read the #ubuntu-kernel logs when they appear. And lurk there, too :)
[17:11] <rbasak> Sorry if I confused matters.
[17:19] <zul> adam_g:  fixed
[17:22] <foxbuntu> anyone in here got much experience with ZFS?
[17:23] <foxbuntu> got an issue with trying to expand zfs pool/fs on top of a hardware raid5 that has been expanded (JBOD isnt an option because of the controller)
[17:23] <foxbuntu> ...and cant remove the disk from the raid5 now
[17:31] <RoyK> foxbuntu: has the raid capacity been extended?
[17:31]  * RoyK knows a wee bit about zfs
[17:32] <irv> any easy way to unencrypt a home dir?
[17:32] <irv> done using the installation time encryption
[17:32] <RoyK> irv: rsync -avPHA /somewhere/else # ;)
[17:32] <irv> ahh that's what i figured :P
[17:32] <irv> kk
[17:32] <irv> spinnin up a new vm now
[17:32] <irv> :D
[17:33] <foxbuntu> RoyK, yes, the raid5 logical volume has been extended
[17:33] <foxbuntu> RoyK, so we went from 3tb to 4tb
[17:35] <RoyK> virtualization is the new thing! I mean, IBM hasn't been doing it for more than three decades ;)
[17:35] <RoyK> foxbuntu: zpool set autoexpand=on
[17:35] <RoyK> iirc
[17:36] <sarnold> RoyK: but they cheated, they designed their chips to be virtualizable without any extra hassle. :)
[17:36] <foxbuntu> RoyK, then just export/import? or is it all automagic?
[17:36] <RoyK> should be no need for export/import
[17:37] <RoyK> try zpool list
[17:37] <foxbuntu> RoyK, cool, thanks. I will give it a shot
[18:10] <zul> adam_g:  ping it got resubmitted to the right branch can you have a look please?
[18:38] <RoyK> foxbuntu: any luck?
[18:51] <balloons> so forgive me because it's friday afternoon here, but I seem to remember something about the ubuntu cloud images no longer being produced. Is this correct?
[18:53] <sarnold> balloons: timestamps still look new: http://cloud-images.ubuntu.com/
[18:53] <foxbuntu> RoyK, havent had a chance to test it yet...gotta try off hours...
[18:53] <balloons> sarnold, ty :-)
[18:56] <vlad_starkov> RoyK: Hello again! I changed faulted HDD to the new one and partitioned it as it should be: 1GB boot,raid + 999GB raid. No file system formatted.
[18:58] <RoyK> vlad_starkov: perhaps something borked in the encryption part
[18:58] <RoyK> dunno
[19:13] <JoeyJoeJo> If I build a kernel and some modules on one machine how can I package it all up into a deb file for redistribution onto different machines with the same hardware?
[19:14] <sarnold> JoeyJoeJo: investigate the kernel-package pacakge.
[19:14] <JoeyJoeJo> thanks
[20:48] <vlad_starkov> Question: mdadm. I'm in initramfs shell. I just successfully launched recovery process. No it's about 10% done. Can I type exit and reboot to normal system and be sure that everything will be recovered properly?
[21:09] <vlad_starkov> Question: can anyone explain me how it is possible that after I changed faulty HDD in my RAID1 and launched recovery process, after system boot I have clear system with no software and my data like?????
[21:22] <RoyK> vlad_starkov: what do you mean?
[21:25] <vlad_starkov> RoyK: I have a server with 2 HDD built to RAID1+encryption+LVM. One HDD was failed. mdadm was shoing that raid is stopped. I changed it with the new one and reassembled the raid. After that mdadm showed that raid is active and new hdd is recovering now.
[21:26] <vlad_starkov> I reboot to normal mode and I see there is no software
[21:26] <vlad_starkov> actually there is only 1 user of several that system was supported before disk fail
[21:26] <vlad_starkov> I can't understand what's going on....
[21:27] <vlad_starkov> RoyK: As I think the system should boot normally OR do not boot at all.
[21:37] <RoyK> vlad_starkov: seems to me you had a faulty array in the first place
[21:38] <vlad_starkov> RoyK: hmmm....
[21:38] <vlad_starkov> how to check it?
[21:38] <RoyK> vlad_starkov: I guess it was dead for some time and that the data resides on the other drie
[21:38] <phillw> Hi good people, do you have a zsync link for 12.04.2 ? I'm struggling.
[21:38] <RoyK> drive
[21:38] <vlad_starkov> RoyK: this looks the most reasonable cause of the problem...
[21:41] <RoyK> vlad_starkov: try to remove the other drives and boot on the old one
[21:41] <RoyK> vlad_starkov: I've seen similar stuff happening before, but not on linux
[21:44] <vlad_starkov> RoyK: the recovering process is still on. I think of this strategy: remove old drive from the server. One more time repartition the new drive that I installed after raid failure. Then clone the drive, that I removed initially and thought that IT WAS failed, to external disk. Then put it back to the server and reassemble the array.
[21:56] <RoyK> vlad_starkov: try booting on the old drive first
[21:56] <RoyK> if the data is there, the new stuff won't matter
[21:57] <RoyK> make a backup, as in copying it to somewhere else, and fix the mirror
[21:57] <RoyK> don't try to do magick with the existing stuff
[22:06] <vlad_starkov> RoyK: what is the better way to make backup from old drive?
[22:45] <vlad_starkov> RoyK: Can I "hot swap" a regular HDD by unplug it from motherboard cords and then plugin another HDD? Is it fine for working server?
[22:46] <RoyK> if the drivers support it, yes
[22:46] <RoyK> SATA is hotplug by design
[22:46] <RoyK> vlad_starkov: but in your case, better reboot into the old drive
[22:47] <RoyK> you can't hotswap mounted filesystems on single drives
[22:47] <vlad_starkov> RoyK: I'd like to make a backup copy of the old drive before recover the RAID
[22:47] <vlad_starkov> RoyK: I mean boot with USB flash (Ubuntu Live) and hot swap disks
[22:49] <RoyK> vlad_starkov: raid isn't backup
[22:49] <RoyK> keep that in mind
[22:49] <RoyK> backup your data to something and then try to recover the raid
[22:49] <vlad_starkov> RoyK: Yep, it's like insurance
[22:49] <RoyK> no, it's not, backup is insurance
[22:50] <RoyK> or the step before it
[22:50] <RoyK> raid is like safety belt
[22:50] <vlad_starkov> RoyK: I have free healthy HDD that i'd like to use to clone old disk to it. And then begin recovery process
[22:50] <vlad_starkov> if recovery fail the old disk for some reason, I'll have cloned disk
[22:51] <vlad_starkov> RoyK: nice
[22:54] <vlad_starkov> RoyK: I made a short script that I used to successfully launch recovery process. Did I miss something? http://pastebin.com/irJ3JxFU
[22:56] <RoyK> I wonder why you should need that
[22:56] <RoyK> md devs are automatically assembled
[22:56] <RoyK> except nested ones, a bug
[22:58] <RoyK> see bug 1171945
[23:01] <vlad_starkov> RoyK: On Ubuntu 12.04 Server md devs don't automatically assemble after fail if md is encrypted
[23:07] <phillw> secondary server for 12.04.2 is now mirrored at http://phillw.net/isos/ubuntu-server/
[23:35] <vlad_starkov> RoyK: When I make sudo ddrescue -v /dev/sda /dev/sdb it says  Output file exists and is not a regular file. Use `--force' if you really want to overwrite it, but be  aware that all existing data in output file will be lost.
[23:35] <vlad_starkov> RoyK: should I add -f key?
[23:36] <sarnold> vlad_starkov: do you truly intend to overwrite the hard drive /dev/sdb and not just make a new file in the filesystem that is already on /dev/sdb?
[23:36] <vlad_starkov> sarnold: oh. sure I do
[23:36] <vlad_starkov> thanks!
[23:36] <sarnold> vlad_starkov: then go for it :)
[23:39] <vlad_starkov> sarnold: cloning...
[23:41] <vlad_starkov> sarnold: Don't you know, Ubuntu determine the drive name (sda,sdb,sdc...) just randomly and it doesn't bounded to SATA port on motherboard?
[23:42] <vlad_starkov> RoyK: it's cloning 1Tb... for the moment, I'd like to thank you for your help and advices. It was extremely helpful for me.
[23:43] <sarnold> vlad_starkov: this is currently under discussion: https://lists.ubuntu.com/archives/ubuntu-devel/2013-May/037084.html
[23:43] <sarnold> vlad_starkov: but I believe the current behaviour will be retained for the time being...
[23:43] <vlad_starkov> For the moment there is no errors while cloning.... It will be funny if I've removed the wrong drive initially)
[23:44] <sarnold> hehe, "funny" :)
[23:45] <vlad_starkov> sarnold: so the current behaviour is like a just wrote above?
[23:46] <sarnold> vlad_starkov: as I understand it, yes; /dev/disk/* has 'stable' names available..
[23:46] <vlad_starkov> sarnold: mmmm, I didn't understand then
[23:47] <sarnold> vlad_starkov: ls -l /dev/disk/by-id/ will explain :)
[23:47] <vlad_starkov> sarnold: ok then
[23:48] <vlad_starkov> Am I right thinking that instead of doing md array recovering, I can just clone current disk to another and plug it in?