[00:54] <SpamapS> smoser: well my theory seems to not be true.. I added a 50ms delay to the bit of udev where I thought the race was happening but it didn't seem to make the race more likely. :-/
[00:56] <ubuntucloud523> hello
[01:41] <smoser> SpamapS, are you able to reproduce at will ?
[01:42] <SpamapS> smoser: no, thats what I was hoping
[01:43] <smoser> ah.
[01:54] <SpamapS> smoser: hmm.. found another spot to try actually
[10:35] <koolhead17> hi all
[10:42] <flaccid> suup
[10:54] <kim0> koolhead17: howdy
[10:54] <kim0> flaccid: monrning :)
[10:54] <flaccid> evening
[10:55] <kim0> oh hehe
[10:55] <kim0> flaccid: you're working with cloudstack right
[10:56] <koolhead17> hi kim0 flaccid
[10:56] <flaccid> not much yet
[10:56] <flaccid> i work at RightScale
[10:56] <kim0> ah .. is that based on openstack, or a separate product
[10:56] <flaccid> separate
[10:56] <kim0> ah I see
[14:49] <keyz182> Hi, has anyone here successfully set up Puppet on a UEC  cloud? I'm trying, but failing. on the puppet master, I'm getting "puppet-master[3153]: Could not resolve 172.19.1.3: no name for 172.19.1.3" in syslog. I have DNS_DISABLE set to no to enable DNS
[15:05] <kim0> keyz182: looks more like a #puppet question
[15:07] <keyz182> kim0; No, I'm pretty sure it's a networking/UEC issue. running hostname -f or dnsdomainname fail.
[15:07] <keyz182> I'm not sure however if the issue is misconfiguration on my behalf (probably) or a UEC problem
[15:23] <keyz182> Ok, sorted it now. I had to change /etc/dhcp3/dhclient.conf to supersede the domain-name and domain-search, and prepend the clc IP address as the primary dns
[15:24] <keyz182> you have to use the CLC internal/private address though, i.e. in my case 172.19.1.1, otherwise it sees queries returned from that address even though they were sent to the public address
[15:41] <keyz182> Are there any UEC/Eucalyptus developers in? I'm trying to find the best place to modify Eucalyptus so that custom dhcp options can be added (and not overwritten), e.g. setting the domain-search, and dns servers etc. To me, it looks like tools/eucanetd is the place to do it in, as it appears to be writing the dhclient.conf file, does that sound right?
[18:57] <smoser> mathiaz, i just uploaded builds to awstools ppa
[21:40] <navanjr> looking for help troubleshooting performance issues in my setup...  NC has 16cores and 24gb ram and a small instance with 1core and 1gb ram is performing poorly.  htop uses 30% of the CPU in the instance
[21:40] <navanjr> I have looked over the logs for any obvious errors
[21:42] <navanjr> and googled and read all i can find to help me trouble shoot.
[21:46] <navanjr> CPUs on NC are 2.4
[21:46] <SpamapS> navanjr: do you definitely have hardware virtualization support?
[21:46] <TeTeT> navanjr: yeah, check kvm-ok if possible
[21:47] <navanjr> yes... INFO: Your CPU supports KVM extensions, INFO: /dev/kvm exist, KVM acceleration can be used
[21:48] <TeTeT> navanjr: on the nc, when you grep for kvm, it is truely kvm running and not qemu by some strange config problem?
[21:50] <navanjr> yes, top says... 25026 libvirt-  20   0 1698m 1.3g 3564 R  100  5.4 109:39.03 kvm
[21:56] <SpamapS> navanjr: also is it actually performing poorly, or is it just showing high load?
[21:57] <SpamapS> navanjr: there have been bugs in the past where the kernel said it was busier than it actually was.
[21:58] <navanjr> it is very sloooow.
[21:58] <navanjr> http://pastebin.com/0iaNGTDb
[21:58] <navanjr> there is my pstree
[21:58] <navanjr> fwiw
[22:00] <navanjr> for my tests i have been running a pretty simple instance loading a few packages from user-data
[22:00] <navanjr> what should take 5 min takes over an hour
[22:30] <TeTeT> navanjr: how is disk i/o on the instance with dd or spew or bonnie? Any good or garbage? How is network IO?
[22:35] <navanjr> should i check this within an instance? or on the NC or CLC
[22:38] <TeTeT> navanjr: on the instance I think
[23:07] <TeTeT> navanjr: good luck, need to logoff now