Wednesday, 2011-07-06

*** littleidea has joined #openstack00:04
*** mgius has quit IRC00:09
*** ton_katsu has joined #openstack00:09
*** jinjax has joined #openstack00:11
jinjaxAny instructions manual for using XenServer 6.0 (beta) for instantiating VMs using Nova?  I am not married to XenServer 6.0, Xen 5.6 will do.00:13
*** Mandell has joined #openstack00:13
*** nijaba has quit IRC00:14
*** nijaba has joined #openstack00:16
*** nijaba has joined #openstack00:16
*** adjohn has quit IRC00:17
*** joearnold has quit IRC00:20
*** nijaba has quit IRC00:22
*** nijaba has joined #openstack00:23
*** nijaba has joined #openstack00:23
kpepplejinjax: start here
*** nijaba has quit IRC00:29
*** msinhore has quit IRC00:30
*** msinhore1 has joined #openstack00:30
*** clauden_ has joined #openstack00:30
*** nijaba has joined #openstack00:31
*** nijaba has joined #openstack00:31
*** littleidea has quit IRC00:32
*** ccc11 has joined #openstack00:34
*** nijaba has quit IRC00:35
*** nijaba has joined #openstack00:36
*** bluetux has joined #openstack00:36
*** clauden_ has quit IRC00:36
*** msinhore has joined #openstack00:37
*** msinhore1 has quit IRC00:37
*** nijaba has quit IRC00:46
*** nijaba has joined #openstack00:46
*** nijaba has joined #openstack00:46
*** vladimir3p_ has quit IRC00:47
*** msinhore has quit IRC00:48
*** mszilagyi has quit IRC00:49
*** dirakx1 has joined #openstack00:52
*** kashyap has joined #openstack00:56
*** jdurgin has quit IRC01:00
*** HugoKuo has quit IRC01:00
*** HugoKuo has joined #openstack01:07
*** ohnoimdead has quit IRC01:07
*** jakedahn has quit IRC01:12
*** lorin1 has quit IRC01:13
*** lorin1 has joined #openstack01:14
*** nijaba has quit IRC01:15
*** nijaba has joined #openstack01:15
*** nijaba has joined #openstack01:15
*** vodanh86 has joined #openstack01:18
vodanh86as i read about networing options in cactus doc : "In both flat modes, the network nodes do not act as a default gateway. Instances are given public IP addresses."01:22
vodanh86is it correct?01:22
*** jinjax has quit IRC01:27
*** nijaba has quit IRC01:28
*** nijaba has joined #openstack01:28
*** nijaba has joined #openstack01:28
HugoKuoI'm not agree with that01:29
*** jeffjapan has joined #openstack01:30
*** dysinger has joined #openstack01:30
vodanh86as I understand, in FLATDHCP mode, nova-network host ac as gateway, but how about FLAT mode?01:31
dysingerquestion: if I have a client that wants to hack on openstack code & they want to override some of swifts code, can they just deploy a python egg that would over-ride ubuntu package python code for swift?  Is it possible to just put the egg code in the path first before starting services ?  I am no python class-loader expert.01:32
HugoKuovodanh86 , in my knowing01:33
HugoKuothe difference between flat & flatDHCP01:33
HugoKuois the way to assign ip for instance01:33
*** nijaba has quit IRC01:34
*** nijaba has joined #openstack01:35
*** nijaba has joined #openstack01:35
*** cereal_bars has quit IRC01:38
*** nijaba has quit IRC01:38
*** nijaba has joined #openstack01:38
*** nijaba has joined #openstack01:38
*** Mandell has quit IRC01:41
*** iammartian has joined #openstack01:44
*** koolhead171 has joined #openstack01:45
*** shang has joined #openstack01:55
*** koolhead171 has quit IRC01:59
*** skraps has quit IRC02:02
*** berto- has quit IRC02:02
*** Mandell has joined #openstack02:05
*** ziyadb_ has joined #openstack02:11
*** ziyadb_ has joined #openstack02:11
*** ziyadb has quit IRC02:12
notmynamedysinger: that should work. depending on what part of swift you want to override, you may simply be able to do it with your own middleware02:16
dysingernotmyname: thanks02:17
notmynamedysinger: but essentially, Python uses a path system like a shell and the first package it finds is the one it uses. the shell var to control this is PYTHONPATH02:17
johnpuris anyone running swift on 11.04? i am seeing some wierd xfs timeout errors...02:17
notmynamejohnpur: sorry I didn't answer you earlier. I don't know. we're using LTS02:18
dysingernotmyname: sounds like ruby, java, erlang, etc - it's all the same - first code in the path that matches wins02:18
notmynamedysinger: indeed02:18
johnpurnotmyname: I know! Did you find any XFS wirdness in getting the LTS version to work?02:19
johnpurnotmyname: the system is throwing timeount eerors on the sync process02:19
notmynamejohnpur: can you paste a traceback?02:20
johnpurnotmyname: not tonight, but soon02:20
notmynamejohnpur: but you're sure it's xfs?02:20
notmyname(ie not swift throwing the timeouts)02:21
johnpurthe error is: INFO: task xfssyncd/sdb2/6694 blocked for more than 120 seconds02:22
notmynamejohnpur: not sure. perhaps tomorrow we can bug letterj or pandemicsyn about it. btw, cw is also an expert on xfs :-)02:23
johnpurnotmyname: sounds good... i am wondering if xfs is sensitive to ubuntu versions... a lot of testing has been done on 10.04, not so much on 11.04...02:25
notmynameand people wonder why we run LTS ;-)02:25
johnpurnotmyname: Sent you my contact info, let meknow if you don't get it02:28
*** [1]RickB17 has joined #openstack02:29
notmynamejohnpur: got it02:29
johnpurnotmyname: cool!02:29
*** jakedahn has joined #openstack02:29
*** jakedahn has quit IRC02:29
*** jakedahn has joined #openstack02:29
*** ziyadb_ has quit IRC02:33
*** lborda has joined #openstack02:36
*** skraps has joined #openstack02:36
*** adjohn has joined #openstack02:39
*** kashyap has quit IRC02:42
*** iammartian has quit IRC02:47
*** gaveen has joined #openstack02:50
*** johnpur has quit IRC02:50
uvirtbotNew bug: #806288 in nova "Unable to ping or ssh Fedora 15 instance" [Undecided,New]
*** j05h has quit IRC02:53
uvirtbotNew bug: #806289 in nova "Error launching RHEL 6 instance" [Undecided,New]
*** osier has joined #openstack02:58
*** PeteDaGuru has quit IRC03:00
*** lmi563 has joined #openstack03:01
*** kashyap has joined #openstack03:08
*** gohko has quit IRC03:14
*** aliguori has quit IRC03:16
*** openpercept has joined #openstack03:18
*** [1]RickB17 has quit IRC03:19
*** mdomsch has joined #openstack03:20
*** kashyap has quit IRC03:20
*** dgags has quit IRC03:22
*** ewindisch has quit IRC03:22
*** kashyap has joined #openstack03:22
*** saju_m has quit IRC03:29
*** toytoy has joined #openstack03:30
*** toytoy has quit IRC03:35
*** toytoy has joined #openstack03:35
*** toytoy has joined #openstack03:35
*** toytoy has quit IRC03:36
*** toytoy has joined #openstack03:36
*** GeoDud has quit IRC03:36
*** lorin1 has left #openstack03:39
*** lborda has quit IRC03:40
*** lmi563 has left #openstack03:41
*** deepa has quit IRC03:42
*** AimanA is now known as HouseAway03:44
*** deepa has joined #openstack03:46
*** jakedahn has quit IRC03:49
*** mihgen has joined #openstack03:50
*** rchavik has joined #openstack03:51
HugoKuohi guys ,03:52
HugoKuoIs there any better way to update DB while upgrade nova ?03:52
*** mszilagyi has joined #openstack03:52
HugoKuofrom Cactus to Trunk ?03:53
*** jakedahn has joined #openstack03:56
*** Ephur has joined #openstack04:02
*** katkee has joined #openstack04:04
*** osier has quit IRC04:07
*** katkee has quit IRC04:08
*** katkee has joined #openstack04:10
*** MarkAtwood has quit IRC04:12
*** Ephur has quit IRC04:13
*** negronjl_ has joined #openstack04:13
*** Ephur has joined #openstack04:13
*** negronjl has quit IRC04:15
*** deepest has joined #openstack04:17
deepestHi everyone04:17
deepestNice to see you all04:18
deepestI'm writting from Vietnam04:18
deepestI have something to ask you guy about Openstack04:18
deepestIf possible, please give me some suggestion04:18
*** MarkAtwood has joined #openstack04:18
deepestI'm a beginner in OpenStack04:19
deepestI want to know about Storage04:19
deepestusing solaris04:19
deepestanyone can support me?04:20
*** jakedahn_ has joined #openstack04:24
*** jakedahn has quit IRC04:28
*** jakedahn_ is now known as jakedahn04:28
kpeppledeepest: you want to use solaris with nova-volume or swift ?04:28
deepestI want to use solaris with nova-volume04:29
deepestdo you have any document or source code to describe about that?04:30
deepestin my company now04:30
deepestwe have a server with solaris04:31
deepestand I want to use this server with solaris for nova-volume04:31
deepestwhat do I need to do?04:31
deepestand How can I do this job?04:32
kpeppledeepest: there is a SolarisISCSDriver in nova/volume/ that should help you. follow the directions in that file at line 144.04:33
deepestwhere can I see this file?04:34
kpeppledeepest: if you've installed from the source code, it's in trunk/nova/volume/  If you installed from package, it's probably where ever they installed the nova egg.04:35
kpeppledeepest: let me see if I can find it online for you04:35
deepestok thank you so much04:36
*** jakedahn has left #openstack04:37
kpeppledeepest: read this file ( starting at line 144 for instructions on setting up the Solaris server04:38
deepestthank you, Kpepple04:41
deepestdo you have an email address?04:41
kpeppleno problem04:42
deepestcould you give me? If I have something to ask, I will send mail to you04:42
kpeppledeepest: my contact info is at
deepestas we know, we must to install nova-volume with LVM (logical volume manager)?04:45
deepestand now, if I want to use solaris then where can I install nova-volume?04:46
deepestin solaris server or in any host04:46
kpeppledeepest: i think you need to install it on another host (it uses ssh to contact the solaris iSCSI server)04:47
deepestand about Solaris Server04:48
deepestdo they need LVM to contact to nova-volume?04:48
kpeppledeepest: i don't think they need LVM04:48
*** mszilagyi has quit IRC04:50
deepestI think my acknowledge is not enough04:51
deepestso, do you have any document to describe more details about  Storage?04:51
deepestI read the tutorial file " compute adminguide cactus" and saw some information about LVM04:53
kpeppledeepest: the best tutorial i've read is but that does not talk about Solaris04:53
deepestand they said that "To set up Compute to use volumes, ensure that nova-volume is installed along with lvm2."04:53
deepestbut I think, the firstly I must understand deeply in Storage04:54
deepestand after that I can change to Solaris04:55
kpeppledeepest: yes, it is best to start with something easier like linux then try solaris04:55
deepestI will try my best to do this job04:56
deepestin the future, I think I will have many question. I will send email to you,If possible, please support me.04:57
*** f4m8_ is now known as f4m804:57
*** Mandell has quit IRC05:00
kpeppledeepest: good luck05:00
*** vladimir3p_ has joined #openstack05:07
deepestthank you05:08
*** Espenfjo has quit IRC05:09
*** shentonfreude has joined #openstack05:27
*** vladimir3p_ has quit IRC05:31
*** miclorb_ has quit IRC05:39
*** miclorb has joined #openstack05:39
*** ewindisch has joined #openstack05:52
*** dirakx1 has quit IRC05:56
*** ewindisch has quit IRC05:56
*** ewindisch has joined #openstack06:00
*** ccc11 has quit IRC06:09
*** katkee has quit IRC06:12
*** ewindisch has quit IRC06:18
*** mihgen has quit IRC06:23
*** mihgen has joined #openstack06:23
*** gohko has joined #openstack06:29
*** wcang has joined #openstack06:32
*** Jedicus2 has joined #openstack06:35
*** po has joined #openstack06:35
*** Jedicus has quit IRC06:36
*** berto- has joined #openstack06:40
*** dragondm has joined #openstack06:45
*** mihgen has quit IRC06:45
uvirtbotNew bug: #806338 in nova "recursive zone calls hang when at least 1 child API call hangs" [Undecided,Confirmed]
*** saju_m has joined #openstack06:47
*** MarkAtwood has quit IRC06:48
*** murkk has quit IRC06:58
*** berto- has quit IRC06:59
*** murkk has joined #openstack07:01
*** reidrac has joined #openstack07:02
*** reidrac has left #openstack07:03
*** reidrac has joined #openstack07:03
*** reidrac has left #openstack07:03
*** berto- has joined #openstack07:03
*** reidrac has joined #openstack07:04
*** zenmatt has quit IRC07:04
*** Ephur has quit IRC07:06
*** skraps has quit IRC07:08
uvirtbotNew bug: #806340 in nova "All VLANs/bridges created even if it's unnecessary " [Undecided,New]
*** BK_man has quit IRC07:13
*** BK_man_ is now known as BK_man07:13
*** berto- has quit IRC07:21
*** willaerk has joined #openstack07:22
saju_mjpgeek1: now both ping and SSH not working
saju_mping to instance was working before flat-network configuration.07:26
*** npmapn has joined #openstack07:31
*** mancdaz1203 has quit IRC07:32
*** mancdaz has joined #openstack07:33
saju_mjpgeek1: hi, r u busy ?07:38
*** ccc11 has joined #openstack07:40
*** duffman has quit IRC07:40
*** duffman has joined #openstack07:40
vodanh86i plan to deploy as in the picture!/?cid=a996ef8137167635&sc=documents&uc=1&id=A996EF8137167635!166!cid=A996EF8137167635&id=A996EF8137167635!216&sc=documents07:45
vodanh86is it posible?07:45
*** nacx has joined #openstack07:48
*** saju_m has quit IRC07:50
*** dobber has joined #openstack07:54
*** ibarrera has joined #openstack07:58
*** katkee has joined #openstack08:01
*** berto- has joined #openstack08:03
*** mancdaz has quit IRC08:04
*** saju_m has joined #openstack08:05
*** mancdaz has joined #openstack08:05
*** saju_m has quit IRC08:06
*** nagyz has joined #openstack08:06
nagyzI have a two-compute node setup. At the first node, everything is OK, but at the second, the ubuntu cloud image behaves like it hasn't got an IP. however, both in db and in the tools, I see the assigned IP. I'm using flat manager.08:06
nagyzany ideas?08:06
matttnope, but i have a question ... w/ a two compute-node setup, does each compute node need to run nova-compute?08:12
nagyznova-compute IS the compute component.08:12
matttyeah, that's what i figured, until i read some answers saying that nova-compute does not need to run on the dom0, which threw me off.08:12
nagyzI don't know about Xen, but I guess it should be the same with KVM: it should run on each compute node.08:13
nagyzhence the name. :)08:13
*** nacx has quit IRC08:13
*** ibarrera is now known as nacx08:13
*** saju_m has joined #openstack08:14
*** jeffjapan has quit IRC08:17
matttnagyz: think i misunderstood what was asked and said when i read it late last night :)08:17
*** dragondm has quit IRC08:19
nagyzit sais it needs to run on each physical host in dom008:20
nagyzobviously. :)08:20
*** deepest has quit IRC08:21
*** saju_m has quit IRC08:25
*** Vasichkin has joined #openstack08:30
nagyzhm, checked /etc/network/interfaces inside the image, and that seems ok08:33
nagyzand I can't ping a VM on compute node 1 from compute node 208:35
*** bluetux has quit IRC08:35
nagyzwith tcpdump I see the ARP request arriving in the vnet interface belonging to the VM08:35
nagyzah, there's a mention of an iptables prerouting DNAT08:38
nagyzthat doesn't change a thing.08:41
*** circloud has joined #openstack08:49
*** chemikadze has quit IRC08:55
nagyzhm, with a natty image, it isnt even booting08:57
*** chemikadze has joined #openstack08:57
doude_Hi all, I get an error with one of my nova-compute nodes. When it tries to fetch image from Glance, Glance responds with 300 http code: and nova-compute fails with this log:
nagyzdoude_, I'm trying to set up my second compute node. did you have any problems with networking?08:57
doude_it's strange, I get this error only with this node. With others nodes, it works nicely. I use he same version (rev 1244) on all nodes08:59
doude_nagyz: no networking problem. Which network mode do you use ?08:59
*** dobber has quit IRC09:00
nagyzand according to all the logs and the injected interfaces file, the VMs on the second compute node is ok09:01
nagyzbut in reality, they don't work09:01
doude_nagyz: I use VLAN mode. What's your problem ?09:01
nagyzthe VM starts up, the interfaces file looks good, populated with the IP09:01
nagyzhowever, the VM stops booting and starts throwing metadata timeout messages09:01
nagyzhere they mention I should do a DNAT for UEC images09:02
nagyzbut I did that, and no packet gets caught in the rule09:02
doude_nagyz: do you have boot log of your VM ?09:02
nagyzlet me start up one09:02
doude_nagyz: ok09:03
nagyzwell, now I don't see the timeout messages, but I see this:09:04
nagyzinit: plymouth main process (48) killed by SEGV signal09:04
nagyzinit: plymouth-splash main process (296) terminated with status 209:04
nagyzthis seems very similar09:05
*** dobber has joined #openstack09:06
nagyzshould I switch to FlatDHCPManager?09:06
*** morfeas has quit IRC09:11
doude_nagyz: your VM (uec image) try to acces the matadata on IP You must forward the of this IP to ec2_dmz_host:ec2_port of your IaaS. In flatDHCP and VLAN mode, an iptables rules is automaticaly added, but in flat mode it isn't done. So you must do it by yourself09:12
nagyzI've added it per instructions of the linked docs09:13
nagyzbut with iptables -t nat -L -v -n, I see zero packets on that rule09:13
nagyzah, it should be done on the default gw09:14
*** miclorb has quit IRC09:14
nagyznot on the compute node09:14
nagyzok, added on the default gw, still no hits on the rule09:16
*** mancdaz has quit IRC09:18
doude_nagyz: yes you must add this rule on each compute nodes09:19
doude_nagyz: in flat mode09:19
*** mancdaz has joined #openstack09:19
dsockwellwin 2709:21
nagyzI've added it on the gw the VMs are using09:21
nagyzthere's no difference sadly09:22
nagyzactually, all my console.logs look like that, even if it's working...09:22
nagyzat least I can ping those, but nothing else09:23
*** jpgeek1 has quit IRC09:24
*** 13WAAEE7D has joined #openstack09:24
doude_nagyz: where is your nova-api server ? Behind your gateway ?09:24
nagyzwe're all in the same net :) is the gw, .3 is cn1, .4 is cn209:24
nagyzthe VMs are using
*** ccc11 has quit IRC09:26
*** Capashen has joined #openstack09:28
nagyzi've opened up vnc on the vm, and all I get is a blinking cursor :(09:33
nagyzit used to work with one compute node, now even that doesn't work... :)09:33
*** berto- has quit IRC09:35
doude_nagyz: so the compute nodes and api server are on on the same subnet ?09:36
nagyzas I told you, already is in
nagyzlet me do a clean install...09:37
doude_nagyz: I think you must set the NAT rule on all node09:37
nagyzI already did09:38
nagyzno packet matched the rule09:38
nagyzset it up on the second compute node and the default gw09:38
nagyzdidn't change a thing09:38
*** KAM has joined #openstack09:39
*** irahgel has joined #openstack09:40
KAMIs there a way to execute commands on a running server using open stack ?09:42
KAMdoes 'what' means that there's not ?09:44
*** BMRU has joined #openstack09:45
KAMI want to execute certain scripts on a running machine using open stack, Is this possible ? ( without using a server running inside the machine )09:46
*** Deirz- has joined #openstack09:47
*** mgoldmann has joined #openstack09:49
sorenKAM: I think you're looking for something like Puppet.09:50
*** MarcMorata has joined #openstack09:51
doude_nagyz: I don't understand your problem. The VM boot ?09:52
nagyzI'm reinstalling the compute node now, and starting from scratch.09:52
nagyzbut yes, the VMs do boot, I can ping it, but nothing happens except that plymouth segfaulted.09:53
*** BMRU has left #openstack09:53
KAMsoren: puppet requires an internal server to run inside the VM, no ?09:54
sorenKAM: This is the first time you mention VM's.09:54
KAMsorry, I meant server09:55
nagyzKAM, what you're saying doesn't even makes sense.09:56
nagyztry to rephrase it if you expect answers. :)09:56
sorenKAM: I think you need to start over and explain whatyou're trying to do.09:56
KAMOk, I'm trying to execute certain scripts on a running server ( instance ) using open stack.09:57
KAMIs it clear now ?09:58
sorenNot really.09:58
sorenWhy don't you just... you know... do it?09:58
sorenIf I want to run a command I just... you know... run it.09:59
sorenNow, if I have A MILLION machines on which I want to do this... Then it's a different story.09:59
soren...and this brings us back to the "explain what you're trying to do" bit.09:59
*** adjohn has quit IRC10:00
Deirz-hello there. I set up nova on two machines: On first runs api, network, scheduler, compute and glance. On second - only compute. I put into nova conf these lines "--glance_host= --image_service=nova.image.glance.GlanceImageService", and on first machine it works as intented, but on second `euca-describe-images` says "no route to host"10:00
KAMah, I meant I'm launching this server ( instance ) for somebody else, then while he is working I want to execute some scripts on the background.10:00
KAMWithout me actually accessing the machine.10:00
sorenKAM: So why don't you?10:00
sorenOk, so I'll go back to my original answer: Puppet.10:01
Deirz-but `wget -q -O -`, invoked on second machine, shows me a list of images10:02
KAMsoren: Yup, but that requires other software to be installed on the server. Something that actually listens puppet, no ?10:02
dsockwellpuppet connects out10:03
dsockwellKAM: if you were using openvz you might use vzctl to inject commands/processes into the machine10:03
dsockwellwhat's your goal? what do these scripts do?10:03
sorenKAM: Yes.10:03
sorenKAM: Nothing you've said suggests that this is a problem.10:04
sorenKAM: Erl, well, as dsockwell says, puppet connects out (usually). But it needs to run inside the vm.10:04
dsockwellyou could put a key into /root/.ssh/authorized_keys to allow yourself root access to the vm10:05
dsockwellthat would be part of the image file, unless that's particularly forbidden10:05
KAMdsockwell: Just executing monitoring scripts based on what the user is doing.10:05
dsockwellyou could inject the scripts into the image using init.d or rc.local10:06
dsockwellhey, does any part of a typical compute node require direct access to the database that nova-api keeps?10:08
dsockwellthat would be a silly thing to allow, wouldn't it?10:08
*** saju_m has joined #openstack10:13
*** wcang has quit IRC10:18
*** chemikadze has quit IRC10:22
*** circloud has quit IRC10:23
*** circloud has joined #openstack10:25
*** chemikadze has joined #openstack10:28
*** dobber has quit IRC10:32
*** toytoy has quit IRC10:37
*** duffman has quit IRC10:47
*** duffman has joined #openstack10:48
*** adjohn has joined #openstack10:58
*** dobber has joined #openstack11:03
*** toytoy has joined #openstack11:04
*** 13WAAEE7D has quit IRC11:07
*** toytoy has quit IRC11:09
*** miclorb_ has joined #openstack11:13
*** adjohn has quit IRC11:25
sorendsockwell: It does require that, actually. For now, at least.11:30
*** ctennis has quit IRC11:37
freeflyingcan I just use public IP for multiple node installation of nova?11:42
*** ziyadb has joined #openstack11:47
*** ziyadb has joined #openstack11:47
*** ctennis has joined #openstack11:50
*** ctennis has joined #openstack11:50
mpatelI am having issues with virbr0 interface  and seems SNAT is not working properly and breaking network connections11:51
nagyzwhen I try to query from the second compute node, it gives me a python exception, saying 'no floating ip for address address_of_my_compute_node'11:53
nagyzdo I need to define floating IPs?11:53
*** katkee has quit IRC11:55
nagyzI don't want to use floating IPs...12:00
*** PeteDaGuru has joined #openstack12:00
vodanh86nagyz no, if you query above url from any compute node, exception will occurs12:01
nagyzthis is my exact same problem12:01
*** markvoelker has joined #openstack12:01
vodanh86show your compute node ifconfig12:02
vodanh86and "ip addr"12:03
nagyzthe other compute node is, it's running nova-api too12:04
nagyzand nova-network, and nova-scheduler12:04
nagyzthis is how the console looks like for the affected VM:
*** msivanes has joined #openstack12:05
nagyzthe VM's default gw is .129, which is an IP on the other compute node, and the other compute node has the iptables prerouting rule12:06
nagyzdone automatically by nova-network, I guess12:06
*** Krooks has joined #openstack12:07
vodanh86does you enable ipv4 forward?12:07
nagyzsince vnet0 is part of the bridge, why should I?12:08
nagyzat least on cn212:08
nagyzon cn1, I didn't touch anything; nova-network handles that12:08
*** Linex has quit IRC12:08
mpatelthe network gets messed up when the instance is started12:10
nagyzdoesn't qemu assign vnet0 to it?12:10
*** ton_katsu has quit IRC12:10
nagyzinteresting, if I do a tcpdump on vnet0, I see ARP requests for
nagyzwhich is the default gw for the VM, and one of the IPs of the other compute node12:11
nagyzwhy can't it find it using ARP?12:11
nagyzon br100, I see the arp reply12:11
nagyzbut not on vnet0...12:12
*** j05h has joined #openstack12:17
nagyzhas anyone seen something like that?12:21
doude_Hi, how I can install the last version (2.5.7) of python-novaclient over my distrib version (2.4.2) ?12:23
doude_I get las code from launchpad. If I install it with (python install), I have the 2 versions on my system12:24
*** katkee has joined #openstack12:30
*** mies has quit IRC12:30
doude_Ok I solve my problem12:30
doude_but I got another one12:31
*** ccc11 has joined #openstack12:32
doude_I just create an admin user and I try toask the OSAPI with his credentials (nova list for example), but I get this error:
*** chetan has joined #openstack12:34
*** miclorb_ has quit IRC12:34
doude_in nova-api logs, I can see successfully authentication12:34
chetanhow to save instance to relaunch its as seperate VM in openstack  ?12:35
nagyzthis whole networking mess is so annoying.12:36
chetanOr  EBS bootable images possible in OPenstack ?12:36
*** aliguori has joined #openstack12:36
mpatelI am also having problem with network12:37
nagyzthe VM on the second compute node doesn't get any ARP replies12:37
nagyzso it can't talk to the metadata server12:37
nagyzso it won't boot.12:37
mpatelvirbr0 breaks all the network connectivity12:37
mpatelyes same here it seems the SNAT rules are broken in iptables12:38
nagyzI'm talking about ARP.12:38
*** dysinger has quit IRC12:38
nagyzthat has nothing to do with SNAT/DNAT12:38
*** Eyk^off is now known as Eyk12:40
*** gaveen has quit IRC12:40
*** HugoKuo_ has joined #openstack12:44
*** vodanh86 has quit IRC12:45
*** HugoKuo has quit IRC12:48
*** wcang has joined #openstack12:50
*** hggdh has quit IRC12:53
*** hggdh has joined #openstack12:55
*** shentonfreude has quit IRC12:59
*** marrusl has joined #openstack13:02
*** ziyadb has quit IRC13:03
*** mdomsch has quit IRC13:06
*** ziyadb has joined #openstack13:07
*** ziyadb has joined #openstack13:07
*** fulc has joined #openstack13:09
*** jaypipes has joined #openstack13:13
*** circloud has quit IRC13:14
*** ccc11 has quit IRC13:16
*** circloud has joined #openstack13:17
*** ccc11 has joined #openstack13:17
nagyzso for the second compute node, in case of the default vlan networking, should I add a network address to it's bridge?13:19
nagyzor shouldnt that be done automatically?13:19
sorenWhatever needs to happen will happen automatically.13:21
nagyzin the instance I keep seeing   13:18:42 [ 2/100]: url error [[Errno 101] Network is unreachable]13:21
sorenWhen I say "will" I of course mean "should".13:23
nagyzI see vnet0 being part of the bridge, automatically created (this is on the second node, so there's no nova-network running here)13:24
*** kbringard has joined #openstack13:24
nagyzhow could I find out what's wrong in this case?13:25
nagyzsince in this case it's set to dhcp, it could be that it doesn't see the dhcp reply/cant make the request13:26
nagyzthat would casue a network unreachable error13:26
*** kbringard has quit IRC13:26
*** kbringard has joined #openstack13:27
nagyzhow could I verify that the second compute node indeed sees the traffic in the vlan?13:28
*** bastichelaar has joined #openstack13:28
nagyzis it OK to assign an IP from the same subnet to br100 on it?13:28
kbringardyou could just tcpdump it13:28
bastichelaaris anyone using openstack with lxc?13:29
kbringardDHCP is broadcast, so if the machine you're running tcpdump on is in the same broadcast domain it should see the traffic13:30
*** evil_e_ has joined #openstack13:30
nagyzkbringard, let me give it a try.13:30
nagyzso on the second node, it's ok to tcpdump br100 :)13:30
kbringardshould be13:30
nagyzlets see13:30
kbringardtcpdump -i br100 -vvv -s 1500 port 67 or port 68... I think that'll do it13:31
fulcHello, I just managed to install a dualnode configuration using StackOps. The smart installer is at http://address:8888/, but when I try using the https://address:8773/ to actually have it running, i get this error (Error code: ssl_error_rx_record_too_long)13:31
kbringardoh, it looks like there is a program called dhcpdump too13:31
nagyzkbringard, right after reboot if I don't restart nova-network manually, it gives me an error message when the scheduler first wants to start a VM13:32
nagyzis that normal? :p13:32
nagyzafter restart, it's ok.13:32
kbringarddoesn't seem like it13:32
*** foxtrotdelta has joined #openstack13:32
nagyzright now I have eth0 without any bridge or anything on both hosts with ips, and eth1 specified as the vlan interface, and a subnet on it13:33
nagyzdoes that sounds correct?13:33
nagyzI mean, feasible.13:33
*** wcang has quit IRC13:33
kbringardyea, totally13:33
kbringardthat's what I do13:33
bastichelaarcan anyone help me with some issues with LXC?13:33
nagyzI see the DHCP request on br100 on the second compute node; there's a DHCP reply which I see at the FIRST compute node13:33
kbringardbastichelaar: sorry, never used LXC13:34
nagyzbut not at the second one...13:34
nagyzhm, wait, the mac addresses doesn't match13:34
*** msivanes has quit IRC13:34
*** PeteDaGuru has quit IRC13:34
*** evil_e has quit IRC13:34
*** klumpie has quit IRC13:34
*** klumpie has joined #openstack13:34
nagyzthat means I don't even see the dhcp requests on the first node, running nova-network and the dhcp server13:35
nagyzthat would mean they aren't in the same vlan?13:35
kbringardwell, how many networks do you have configured?13:35
nagyzonly one,
nagyzI can see it with nova-manage fixed list, and network list13:36
nagyzthis is the error I see if I don't manually restart nova-network on the second node:
nagyzand this is what it does after manual restart:
kbringarddo you have eth1 up with no IP configured?13:37
nagyzhm, doesn't seem so. that causes the first error then. let me add it to /etc/network/interfaces13:38
kbringardauto eth113:38
kbringardiface eth1 inet manual13:38
kbringardup ifconfig eth1 up13:38
*** mies has joined #openstack13:38
kbringardI have that in my /etc/network/interfaces file13:38
kbringardit'll bring up eth1 with no IP13:38
nagyzok, let me add it :-)13:38
kbringardyou'll need to add it to all your compute nodes as well13:38
kbringardbecause if the interface isn't there, the bridge and vlan can't attach to it :-)13:39
nagyzthat's true :)13:39
nagyzas an enhancement, openstack could set it to up state :p13:39
kbringardhaha, feel free to code it in :-p13:39
nagyzactually, I have some enhanced security functionality patches13:39
nagyzbut I'm waiting for legal to say if we can release it or not...13:40
nagyzwe're a big company :)13:40
nagyzhas been waiting for a month.13:40
kbringardit's just text... if it happens to get emailed to someone accidentally.. ;-)13:40
nagyzwe're releasing a paper proposing additional security to general IaaS clouds13:40
*** kbringard has left #openstack13:40
*** kbringard has quit IRC13:41
*** deepa has quit IRC13:41
*** deepa has joined #openstack13:41
*** kbringard has joined #openstack13:41
kbringardsorry about that13:41
nagyzit's ok13:41
kbringardI blame the VPN13:42
*** huslage has joined #openstack13:42
*** PeteDaGuru has joined #openstack13:43
nagyzI misspelled the syntax in interfaces, and not it got stuck at mounting the nfs share...13:43
nagyzgonna have to fix it with sweet init=/bin/bahs13:43
huslagemorning. i just found that one of the backing images for several servers has disappeared. coworkers yay. I've recreated it, but it has a different id. Should I just update the DB with the new image's ID for the effected servers?13:43
kbringardI hate that13:43
kbringardhuslage: you shouldn't have to...13:43
*** j05h has quit IRC13:44
kbringardin theory it should download the image again from glance13:44
huslagesomeone deleted it nagyz13:44
huslageout of glance13:44
huslageit's GONE13:44
kbringardand then use the image it just downloaded as the backing image for the instance it's building13:44
kbringardout of glance13:44
*** msivanes has joined #openstack13:44
huslagenow i'm seeing: (nova): TRACE: Error: Image 10 could not be found.13:44
kbringardyou still shouldn't have to worry about it13:44
*** npmapn has quit IRC13:44
huslagewhen i try to do anything to those servers13:44
kbringardthat's weird13:44
huslagei'm a little worried that they won't reboot13:44
kbringardin theory the running servers should keep running13:45
huslageand that would be bad mkay13:45
kbringardand new instances should launch, so long as you use the new image id13:45
huslageyes kbringard but the old instances will not reboot, am i correct?13:45
huslagewell i can't snapshot them to back them up at least13:45
kbringardhmmm, I'm unsure13:45
kbringardI've never had that happen :-)13:45
huslageme either :)13:46
kbringardis there one that is OK to lose?13:46
kbringardtry it :-D13:46
huslageedge case ftw13:46
huslageyes i'll do that13:46
kbringardI would think it'd be OK13:46
*** f4m8 is now known as f4m8_13:46
kbringardthe image is gone from _base as well?13:46
huslagelooks like it rebooted fine kbringard13:48
*** openpercept has quit IRC13:48
kbringardcool, that's good to know13:48
*** jatsrt has joined #openstack13:48
nagyzok, so now that eth1 is finally up13:48
kbringardI don't see why it wouldn't... there's an autonomous copy running in $instances/instance-$id/13:49
nagyznova-network doesn't have to be restarted manually13:49
jatsrtmorning all13:49
kbringardnagyz: w00t13:49
jatsrtor evening13:49
jatsrtbeen away for a bit working on another project13:49
*** vladimir3p_ has joined #openstack13:49
nagyzok, network is still unreachable13:49
jatsrtjust updated to latest head13:49
nagyzbe back in 1013:49
jatsrtand I am getting "trying to add VLAN #100 to IF -:None"13:49
jatsrtso some conf param changed13:49
huslagekbringard: i just can't snapshot it with 'nova image create'13:49
jatsrtthought it was vlan_interface to bridge_interface, but that did not seem to work13:50
jatsrtany thoughts?13:50
*** NeCRoManTe has joined #openstack13:50
*** openpercept_ has joined #openstack13:50
kbringardhuslage: I'm not sure, I don't use that command... I just qemu-img snapshot it13:51
huslagekbringard: k13:51
kbringardthat's how I do it13:51
*** vladimir3p_ has quit IRC13:51
kbringarddrop that on the compute node13:51
kbringardand run it with -i $instance-id13:52
kbringardjatsrt: reading back13:52
huslagecute kbringard13:52
kbringardjatsrt: hrmm, lemme look through the code and see if I can find it13:53
jatsrtfull trace13:53
jatsrtthis is what I get for taking a couple of weeks off :-)13:53
kbringardthat'll teach ya13:54
bastichelaaropenstack and lxc doesnt seem to be quite popular?13:54
kbringardwhat hypervisor are you using?13:55
jatsrtme? kvm, but this is on the nova-network node13:55
kbringardnova/network/'vlan_interface', None,13:55
jatsrtnova-network will not start13:55
kbringardit looks like it should still be vlan_interface13:55
jatsrtthat's what I thought13:55
kbringardI'm running 1244 with --vlan_interface and it works13:56
jatsrtbut then in the, everything says bridge_interface13:56
jatsrtmay have to go through code diffs to see if something changed13:56
jatsrtdef ensure_vlan(vlan_num, bridge_interface):13:57
jatsrtunless it thinks I am no longer clan network mode13:57
jatsrtmaybe that changed?13:57
kbringardyea, interesting13:58
kbringardnova/network/        _execute('sudo', 'vconfig', 'add', bridge_interface, vlan_num)13:58
kbringardthere you go13:58
kbringardprobably still working for me because I have not removed the vlan tagging since I updated13:58
kbringardcheck it13:58
*** Jedicus2 has quit IRC13:58
kbringardhrmm, n/m, that's in the tests13:59
kbringardbin/nova-manage:            bridge_interface = FLAGS.flat_interface or FLAGS.vlan_interface13:59
kbringardonly in nova-manage?13:59
kbringardseems like a bug? :-/14:00
jatsrtthe morning I try to update :-(14:00
kbringardI'm really not sure... don't know why it would have changed14:00
kbringardwhich is to say, I don't know what the reasoning for a change was14:00
kbringarddid you do a db sync?14:01
kbringardI see a migration that adds the bridge_interface column14:01
jatsrtyes I did14:01
mpatelkbringard:  when the instances are started, the network gets screwed up and when I remove the network configurations for virbr0 interface the external network works fine, so what could be the wrong with virtual network configuration14:02
jatsrtI'll look, maybe that is what it is, not populated in the net table14:02
kbringardjatsrt: I didn't look in depth at how it's used, I just happened to notice it while grepping14:02
*** marrusl has quit IRC14:02
kbringardmpatel: hmmmm14:02
kbringardhow many NICs do you have in the machine?14:03
jatsrttracing through the code that might make sense, looks like that brindge_interface param that is None is coming from the DB14:03
jatsrttwo nics14:03
kbringardjatsrt: sorry, I was asking mpatel :-)14:03
kbringardjatsrt: I wonder if you recreate your networks14:04
jatsrtyeah, going to try now14:04
kbringardthat could be why it's only in nova-manage14:04
*** zenmatt has joined #openstack14:04
kbringardin theory that would populate that field14:04
mpatelbrctl shows br100 and virbr0 interface14:04
kbringardmpatel: what are they attaching to?14:05
kbringardand what does virbr0 belong to?14:05
kbringardis that the default network that is coming up with libvirt?14:05
jatsrt"update networks set bridge_interface = 'eth1';" did it14:05
kbringardjatsrt: hot14:05
mpatelbr100 is attached to eth1 and virbr0 to none14:05
kbringardwell done sir14:05
*** dprince has joined #openstack14:05
* kbringard updates his DBs too14:06
*** marrusl has joined #openstack14:06
kbringardmpatel: is virbr0 part of your openstack setup?14:06
kbringardor is that the default network that libvirt sets up when it's installed?14:06
mpatelno I didn't configured anything for virbr0 not sure how it got deployed14:07
mpatelI believe it's default14:07
kbringardtry running14:07
kbringardvirsh net-destroy default14:08
mpatelalso I see some SNT and DNAT rules in iptables14:08
*** errr_ is now known as errr14:08
*** errr has joined #openstack14:08
mpatelok that command removed the virbr0 interface14:09
kbringardok cool, so then do14:09
kbringardvirsh net-undefine default14:09
kbringardthat should keep it from starting at boot14:10
kbringardseems that was interfering with your OpenStack setup for some reason14:10
kbringardI think by default it's on
fulcHello, I just managed to install a dualnode configuration using StackOps. The smart installer is at http://address:8888/, but when I try using the https://address:8773/ to actually have it running, i get this error (Error code: ssl_error_rx_record_too_long)14:10
*** ldlework has joined #openstack14:11
kbringardfulc: are you using FireFox 3?14:12
kbringardor 4?14:12
mpatelis there something I need to make sure while installing openstack to avoid those issues14:12
fulcit says 514:12
kbringardsometimes Firefox gives that error when SSL is running on a non standard port14:12
kbringardfulc: for esses and gees maybe try a different browser?14:12
fulcooh, so what would you recommend for ubuntu?14:12
kbringardchrome maybe?14:13
fulci have always used firefox14:13
kbringardyea, I dunno... I'm one of those dirty Mac users14:13
kbringardyea, I use Safari and Firefox (but I've never used stackops, so I'm not sure about that specific issue)14:13
kbringardbut I have seen that exact error when trying to login to the remote console on a blade enclosure14:14
kbringardin FF14:14
kbringardI'd probably try Chrome14:14
kbringardmpatel: in theory it shouldn't conflict, but it seems something about your setup is causing it to14:14
fulcyeah, i dont think this error is connected to openstack14:14
kbringardmpatel: when I install compute nodes, part of my install script is to run those virsh commands14:14
*** Jedicus2 has joined #openstack14:15
kbringardso I guess just make removing them part of your deploy procedure14:15
*** cweidenk1ller is now known as cweidnekeller14:15
mpatelkbringard:  is your script is available to public14:16
huslagehow does openstack do locality with volumes? does it try to put a machine and its volume on the same machine if i'm running nova-volume on multiple machines?14:16
kbringardmpatel: not really, but it's just a few apt-get commands14:16
*** cweidnekeller is now known as cweidenkeller14:16
mpatelkbringard: what are those few commands, so I can try14:17
kbringardhuslage: good question... I'm not sure... I've only ever run one instance of nova-volume14:17
huslagekbringard: there is so much to do :)14:18
kbringardin our environment we've opted to use mostly NFS exports14:18
*** cweidenkeller has quit IRC14:18
kbringardso I have not messed with nova-volume a ton, beyond just the basics14:18
mpatelkbringard: I also see some SNAT and DNAT rules with MASQUERADE network, so what are those for14:18
*** circloud has quit IRC14:19
kbringardmpatel: that is part of the default network that we destroyed14:19
*** jkoelker has joined #openstack14:19
kbringardin theory you should be able to flush the iptables and then restart nova-network to get rid of them14:19
kbringardbut, ymmv14:19
*** circloud has joined #openstack14:19
jatsrtwhat is "agent" in nova-manage?14:19
jatsrtthat's new in the past few week, I have so much catching up to do14:20
*** cweidenkeller has joined #openstack14:20
kbringardjatsrt: not sure... based on the context I'd guess it has to do with the new scheduler and zone stuff14:21
jatsrt" arguments: os architecture version url md5hash"14:21
jatsrtI'm intrigued14:21
fulcugh, some other error appears in chrome14:21
fulcerror 10714:22
fulcssl again14:22
kbringardsounds like something is wonky in the SSL setup in the stackops thing14:22
kbringardsomeone suggested disable SSl2 and enable SSL2?14:23
kbringarddisable SSL314:23
fulclemme try14:23
kbringardthat's in an old (3/20/2010) chrome support thread about the same thing14:23
kbringardif that works, it may work in FF as well14:24
*** amccabe has joined #openstack14:24
fulcError 112 (net::ERR_NO_SSL_VERSIONS_ENABLED)14:24
kbringarddude, I dunno14:24
fulcthis is annoying14:24
kbringardI've never used stackops, I have no idea what they're up to14:24
kbringardis that that dell thing?14:25
*** jonkelly has joined #openstack14:25
fulci dont think so14:25
kbringardno, I guess not (sorry, just not googling it)14:26
*** icarus901 has quit IRC14:26
fulcit's just a distro with openstack already inside, so you dont bother getting all the dependencies right14:26
kbringardyea, makes sense14:26
*** alexn65 has joined #openstack14:27
nagyzkbringard, so, I've started up a new instance, and now having a look at tcpdump...14:27
nagyzI see the DHCP requests going out on br10014:28
nagyzbut no reply14:28
*** bcwaldon has joined #openstack14:28
nagyzand I don't see them on the first compute node14:28
kbringarddo you see them on the network controller?14:28
nagyzfirst compute node = network controller14:28
nagyzbut no, I can't say I do14:28
kbringardare your physical switch ports configured correctly for the VLANs?14:29
nagyzwell, it's a virtual switch inside vmware...14:29
kbringardI had that issue... one of my 16 blades' ports wasn't configured14:29
nagyzvlan should work OOB there14:29
kbringardyea, I'd think so14:30
nagyzwhat could I try?14:30
kbringardso, another thing jatsrt just figured out14:31
*** cp16net_ has joined #openstack14:31
kbringardit looks like bridge_interface needs to be set in the DB14:31
kbringardif you're running the latest trunk14:31
nagyzno, it's the bundled version in natty14:31
kbringardit looks like it's replaced vlan_interface14:31
jatsrtupdate bridge_interface14:31
nagyzbut it's not trunk. :)14:31
jatsrteasier that deleting and recreating all of your networks14:31
nagyzI use the same nova.conf in both compute nodes, just changed myip14:31
kbringardnagyz: ah... well, if you see bridge_interface in the DB i'd make sure it's populated14:32
nagyzeven on cactus?14:32
kbringardI'm not sure when it was introduced14:32
kbringardno idea14:32
nagyzI had a look at the network table, and it has a "bridge" field14:32
jatsrthas anyone tried to pause/suspend VMs14:32
kbringardjatsrt: I have not14:32
jatsrtnagyz: not the same issue then14:33
kbringardwe mostly use the ec2 api14:33
jatsrtrecently they added a bridge_interface field too14:33
nagyzis there an easy way to check if both machines see each other thru the vlan?14:33
nagyzjust for testing can I add a random IP to br100?14:33
alexn65hi! can someone tell me about network, bridging and public IPs? For now in flat network mode instance have private net ip and do not see inet14:33
nagyztwo from the same subnet of course on the two hosts14:33
jatsrtkbringard: just thinking if I have to reboot a compute node for maintenance and don't want to live migrate I could suspend unsuspend them14:33
nagyzwell, br100 on the host that has nova-network already has
kbringardjatsrt: yea, it seems like a good feature, I've just not played with it14:34
kbringardnagyz: I'd bring up a new interface14:34
mpatelkbringard: now can't reach through the public IP on eth0 interface after restarting the nova-network14:34
kbringardlike, eth1:1 or something14:34
nagyzthat won't be part of the vlan14:34
kbringardif eth1 is tagged14:34
kbringardit should work14:34
kbringardthat's how I test14:35
nagyzI did add to eth1:1 on the second compute node14:35
nagyzand to the first compute node on eth1:114:35
nagyzand they can ping each other14:36
kbringardand if you cat /proc/net/vlan/config you see that eth1 is tagged with the correct vlan on both nodes?14:36
kbringardah, then that would indicate that the vlan tagging is working properly14:36
nagyzthey both has id 10014:36
jaypipeskbringard: you're up early today...14:37
kbringardjaypipes: hah14:37
kbringardI usually start work at 7 Mountain14:37
*** [1]RickB17 has joined #openstack14:37
jaypipeskbringard: either that or never went to sleep ;)14:37
kbringardbut I keep my mouth shut until I've had more than 3 or 4 cups of coffee ;-)14:37
kbringardthis morning I made the mistake of talking too soon ;-)14:37
nagyzbut if I add an ip from the range to br100 on the second node, should that be able to ping on the first node?14:37
kbringardin theory14:38
nagyzlets see..14:38
huslagei hate computers14:38
nagyzok, added the IP14:38
nagyznothing happens14:38
kbringardI'd also try adding like (or some unused address) to eth1:1 and see if that can ping
nagyzI don't even see the ping coming from cn114:38
nagyzlets see14:38
*** cweidenkeller has quit IRC14:39
nagyzno, it cant14:39
*** cweidenkeller has joined #openstack14:39
nagyzand I don't even see that ping coming on br100 on the first node14:39
alexn65does gateway on br100 need special conf? iptables or so?14:39
kbringardhey jaypipes, do you know what this AgentBuildCommands stuff in nova-manage is all about?14:40
*** vladimir3p has joined #openstack14:40
kbringardI'm intrigued, because it looks like it may be intended to build out nodes14:41
jaypipeskbringard: hmm, I'm actually not sure about that.14:41
jaypipesdabo: you know about that one?14:41
kbringardit's not really a big deal, mostly just curious14:41
jaypipesin unrelated news, I'm still getting used to the new Unity desktop in 11.04... upgraded last night...14:42
*** dysinger has joined #openstack14:42
kbringard<--- running OSX14:42
kbringardso I'll be getting used to the new Lion interface soon ;-)14:42
*** 14WABGEOF has joined #openstack14:43
jaypipeskbringard: indeed14:43
jatsrtkbringard: already have Lion GM14:44
*** alexn65 has left #openstack14:44
huslagekbringard: Lion's pretty good14:45
huslageexcept that my machine randomly tuns off14:45
huslagewhich is kind of annoying14:45
jatsrthuslage: new GM release seems more stable now14:45
kbringardhah, that seems counter productive14:45
huslagejatsrt: that's the one i'm running14:46
kbringardI saw that it was available in my dev downloads14:46
huslageon a 2010 Core i7 MBP14:46
huslageit's not THAT old of a machine lol14:46
kbringardbut I was busy drinking over the weekend :-p14:46
jatsrtsame as me!14:46
huslagekbringard: good idea14:46
*** rchavik has quit IRC14:46
huslageanyone have any idea why the dashboard is giving me fits?14:48
huslageCould not import dashboard.views. Error was: No module named routes14:48
huslagemy django fu is pitiful14:48
*** chetan has quit IRC14:49
kbringardcan't be worse than mine14:49
huslagethis stuff needs to GET TOGETHER14:49
huslagetoo much potential and not enough coordination14:50
doude_Hi, who use Nova in a multizone mode ?14:55
deshantmFor development purposes, is it OK/recommended to use Ubuntu 11.04 or Ubuntu unstable as the base? (for context we are porting Xen Cloud Platform to Ubuntu)14:56
kbringardI don't know that I'd say it's recommended... but I think if you want to use Xen as your hypervisor on Ubuntu you'll have to run it on 11.1014:56
*** patcoll has joined #openstack14:57
kbringardsince modern versions of Xen aren't really supported before that14:57
bastichelaarstill nobody here who uses lxc as hypervisor?14:58
deshantmkbringard: thanks, yeah that's what we are going to target (as well as debian unstable)14:58
kbringardI've never tried it, but soren tells me it works14:58
*** Shentonfreude has joined #openstack14:59
*** dragondm has joined #openstack15:03
*** huslage has quit IRC15:07
*** reidrac has quit IRC15:12
*** ccc11 has quit IRC15:14
doude_Who can help me to use the multizone mode ?15:18
*** foxtrotdelta has quit IRC15:18
*** rupakg has quit IRC15:18
*** rupakg has joined #openstack15:18
nagyzanyone know how can I set up a vsphere vswitch so it won't eat my vlan tagged packets?15:20
*** med_out is now known as medberry15:21
*** 14WABGEOF has quit IRC15:23
*** mpatel has quit IRC15:23
*** huslage has joined #openstack15:25
*** ccc11 has joined #openstack15:27
kbringardsmoser: not really related to OpenStack directly, but if you have a few moments, I have some questions about cloud-init15:27
smoserkbringard, i'd be happy to ask, but i'm actually on holiday now. will return Monday.15:28
kbringardoh, sorry, in that case I won't bug ya15:28
kbringardenjoy your holiday!15:28
*** huslage has quit IRC15:29
Jbainnagyz: are you wanting to pass the tagged packets to a vm?15:30
*** dobber has quit IRC15:30
nagyzI want to set the switch to trunk mode15:30
nagyzbut I've found it15:30
nagyzhad to set the vlan id to 409515:31
*** mancdaz has quit IRC15:31
nagyzlets see if it works now15:32
*** Ephur has joined #openstack15:34
doude_In multizone mode, no host filter is applied. I set flag 'default_host_filter' to 'nova.scheduler.host_filter.HostFilterScheduler', but the filter doesn't return any host15:35
*** mancdaz has joined #openstack15:35
*** KAM has left #openstack15:38
heckjdeshantm: I'd recommend going with one of the LTS versions - which means 11.0415:43
nagyzso what would I need for live migration?15:47
nagyznova-manage vm live_migration doesnt work15:48
nagyzI have the images on a shared nfs15:48
nagyzand I'm using glance15:48
jatsrtnagyz: what's not working about it?15:48
kbringardare the images on shared nfs, or the instances directory?15:48
jatsrttook me a week to get all the details right15:48
*** ccc11 has quit IRC15:48
nagyzwell, I don't see any errors15:48
nagyzbut the machine stays on the original compute node15:48
kbringardI don't think the images has to be on shared storage, but the instances directory does15:48
jatsrtinstances directory15:48
nagyzI only have the instance directory shared between the nodes15:48
nagyzinstances, yes.15:49
jatsrtI also put a note on one of the wiki's about a libvirt param that needed to be set15:49
nagyzI haven't seen that15:49
kbringardah, yea, I remember that now...15:49
jatsrtthe good news Is I can 100% say it does work15:49
nagyzthis is the scheduler log15:49
nagyzwhat param?15:49
jatsrtmake sure you are looking at your compute log too15:50
jatsrton bot source and destination15:50
nagyzno errors per say15:50
nagyzoh, I found it15:50
jatsrtdon't remember the specific param, but you need to hard set the user15:50
jatsrtyou need to make sure you modify all the files listed in the wiki15:51
*** medberry is now known as med_out15:51
jatsrtthat sets it up so it can allow remote communication15:51
nagyzdoing that now, sec15:51
jatsrtbetween libvirt processes15:51
jatsrtalso, are you sure your permissions on your nfs mounts are all straight15:52
jatsrtfinally, make sure all of your compute nodes are the same architecture15:52
jatsrtbut that's for later15:52
*** marrusl has quit IRC15:53
nagyzI have the same uids for the users on both cns15:54
nagyzok, set up libvirt15:54
nagyzlets see15:54
nagyzit sais it migrated it :)15:55
kbringardnagyz: sorry, I forgot about the libvirt stuff that had to be set15:55
nagyzand it did15:55
nagyzit's working :)15:56
jatsrtcongrats :-)15:56
kbringardlike I said, it's been awhile since I used it :-)15:56
nagyzif you're ever in Switzerland, I'm buying a beer ;)15:56
nagyz(or I'll be in Chicago in oct, if everything goes as planned...)15:56
*** lborda has joined #openstack15:58
*** KAM has joined #openstack15:58
*** dgags has joined #openstack16:01
*** Razique has joined #openstack16:03
*** marrusl has joined #openstack16:05
RaziqueHi all16:08
MarkusWPHRorgHi Razique16:08
Raziquesomething went fubar with my san connection, now, im' unable to attach volumes to instances16:08
MarkusWPHRorgMorning all (YMMV depending on time zone)16:08
Raziquethe attachement goes fines, but the instance doesn't see it16:08
Raziquewhile nova considers it as attached, and I see it on the compute node16:08
Raziqueany clue ?16:08
*** cp16net_ has quit IRC16:09
*** koolhead17 is now known as koolhead11|Afk16:09
MarkusWPHRorgRazique: what kind of HBA are you using?16:09
*** MarcMorata has quit IRC16:10
MarkusWPHRorgor are you using an HBA/FC switch?16:10
RaziqueMarkusWPHRorg: it's an HP SAN connected with iscsi16:10
*** maplebed has joined #openstack16:10
MarkusWPHRorgover fiber channel or ethernet?16:11
*** willaerk has quit IRC16:12
*** maplebed has quit IRC16:12
*** maplebed has joined #openstack16:13
*** bcwaldon has quit IRC16:14
*** mdomsch has joined #openstack16:16
nagyzbe back tomorrow16:16
nagyzhave a nice day all16:16
*** nagyz has quit IRC16:16
doude_Who can help me to use the multizone mode ?16:17
doude_In multizone mode, no host filter is applied. I set flag 'default_host_filter' to 'nova.scheduler.host_filter.HostFilterScheduler', but the filter doesn't return any host16:17
*** j05h has joined #openstack16:17
*** Eyk is now known as Eyk^off16:18
Raziqueit's like my instance's libvirt.xml doesn't contain the link to that volume16:18
*** joearnold has joined #openstack16:18
Raziquewhen I try to dettach it :
Raziquecan I "purge" the attachments ?16:19
*** obino has quit IRC16:20
*** peads has joined #openstack16:20
dabojaypipes: Just got back, and saw your question about AgentBuildCommands. Ask Johannes Erdfelt or comstud about that.16:22
*** fulc has quit IRC16:23
*** cp16net_ has joined #openstack16:24
*** huslage has joined #openstack16:24
*** ziyadb has quit IRC16:25
*** joearnold has quit IRC16:26
*** joearnold has joined #openstack16:27
*** HouseAway has quit IRC16:27
*** cp16net_ has quit IRC16:27
*** HouseAway has joined #openstack16:27
kbringarddabo: cool, thanks16:27
kbringard(jay was asking for me :-))16:27
*** huslage has quit IRC16:28
Raziquecould anyone help me ? :D kinda f****d up here :/16:29
MarkusWPHRorgRazique: does HP run a linux-base iscsi16:30
Raziquethink so yes16:30
*** zenmatt has quit IRC16:31
*** zenmatt has joined #openstack16:31
Raziquewhen you attach a volume, the libvirt on the compute node adds line into the libvirt.xml right ?16:32
Raziquesome <disk> entries I guess16:32
*** ziyadb has joined #openstack16:32
*** Capashen has quit IRC16:33
*** jtran has joined #openstack16:33
*** huslage has joined #openstack16:37
huslageoh wow. internets16:38
Raziqueactually, I figured that I just need to "reset" the attachment16:40
Raziqueand everything should be ok by now ; i just created a volume and successfully dettached and reattached it16:40
comstudkbringard: nova maintains a list of the guest agent versions by OS/arch and so forth.  Those AgentBuildCommands are there so you can update the table with what the current version of the agent should be... and the url and md5sum of the binary pkg.16:40
*** jdurgin has joined #openstack16:40
comstudkbringard: This is so when nova fires up a build, it can check the version of the guest agent and auto-update it to a newer version, if one exists in the table.16:40
*** KAM has left #openstack16:41
huslagecomstud: agent?16:41
*** ziyadb_ has joined #openstack16:41
*** ziyadb_ has joined #openstack16:41
*** PiotrSikora has quit IRC16:41
comstudThe guest agent is xenserver-only right now and can be found at lp:openstack-guest-agents16:41
kbringardcomstud: ah, nice. so it's useful in managing your entire cluster and making sure everything is in sync16:41
comstudhuslage: auto-configures a VM16:41
huslagelike cloud-init?16:41
huslageor more like puppet or chef?16:41
comstudnot familiar with cloud-init.. not really puppet or chef, either16:42
comstudA guest agent is needed with xenserver so it can configure the IP adderss16:42
kbringardcloud-init is the stuff that ubuntu wrote for querying the meta-data API for ssh keys and stuff16:42
comstudaddress.. set the root password...16:42
comstudand things like that.16:42
comstudOk, yeah, similar to that, then.16:42
huslageyeah kbringard it's pretty spiffy16:43
kbringardthat sounds spiffy... cloud-init is nice but it's a beeyatch to get running on anything other than Ubuntu :-)16:43
comstudexcept the agent accepts commands from nova.. so things are pushed instead of pulled, I guess you could say.16:43
huslagewhy is it xen only?16:43
comstudhuslage: it was originally internal code at rackspace.. we run xen16:44
comstudit communicates via xenstore16:44
comstudbut it's pluggable now..16:44
kbringardand you guys are using Cent there, yea?16:44
comstudyou can swap that piece out for another communication layer.16:44
kbringardor RH16:44
*** ziyadb has quit IRC16:44
*** obino has joined #openstack16:44
comstudkbringard: not for cloud16:44
comstudother areas of the biz use RH i _think_16:44
kbringardif I may ask, what OS are you running Xen on?16:45
huslagecomstud just works there16:45
comstudkbringard: I'm not sure what Ops is using.16:45
comstudWe're using debian in dev.16:45
*** PiotrSikora has joined #openstack16:45
kbringardah, OK16:45
kbringardmakes sense16:45
comstudsome folks are using ubuntu16:45
kbringardI'm sort of waiting for 11.10 before messing much with Xen16:47
kbringardI've had limited success getting OpenStack to run on anything but Ubuntu, so I just don't want to mess with it, heh16:48
comstudsorry, I should clarify...16:48
*** jatsrt has quit IRC16:48
comstudI was thinking about our dev environments...16:48
comstudWe're running Citrix XenServer16:48
comstud(which appears to be centos based)16:49
kbringardah, OK16:49
kbringardthat makes sense16:49
*** MarkAtwood has joined #openstack16:51
*** zul has quit IRC16:52
*** mihgen_ has joined #openstack16:53
*** bcwaldon has joined #openstack16:55
*** zul has joined #openstack16:55
*** lorin1 has joined #openstack16:57
*** bastichelaar has quit IRC16:57
*** zenmatt has quit IRC16:58
*** berto- has joined #openstack16:59
kbringardsoren: you around?17:02
*** skraps has joined #openstack17:02
*** mdomsch has quit IRC17:02
*** berto- has quit IRC17:02
*** stephan` has joined #openstack17:05
*** pguth66 has joined #openstack17:06
*** berto- has joined #openstack17:06
huslagein flatDHCP, can the nodes all talk to one another via the internal IP?17:07
kbringardhuslage: it depends17:07
huslagei want to start an internal-only mysql server, for instance17:08
*** ohnoimdead has joined #openstack17:08
kbringardwell... there is a flag --allow_project_net_traffic17:08
kbringardI think it's true by default17:08
kbringardoh wait17:08
kbringardthat's vlan mode17:08
kbringardI think17:08
kbringardsorry, ignore me17:08
huslagei can't do vlan mode yet. waiting for the Real Hardware17:08
kbringard<--- not enough coffee17:08
kbringardI don't know much about flatDHCP, sorry17:09
*** deshantm_laptop has joined #openstack17:09
huslageok…how do i return an address to the floating pool?17:10
huslageit seems to never reuse17:10
*** irahgel has quit IRC17:11
*** mdomsch has joined #openstack17:12
*** erik-s has quit IRC17:13
*** erik-s has joined #openstack17:14
*** kd926 has joined #openstack17:15
kd926I'm trying to start nova on gentoo and I get a trace when doing nova-manage db sync:17:15
kd926TypeError: immutabledict object is immutable17:15
*** CatKiller has quit IRC17:17
MarkusWPHRorgDoesn anyone here know anything about running Openstack with rPath?17:19
MarkusWPHRorgIf so, please PM me....17:19
*** AhmedSoliman has joined #openstack17:20
*** joearnold has quit IRC17:20
*** joearnold has joined #openstack17:20
*** dendro-afk is now known as dendrobates17:20
*** ziyadb_ has quit IRC17:21
*** po has quit IRC17:21
*** AhmedSoliman has quit IRC17:22
*** AhmedSoliman has joined #openstack17:23
*** CatKiller has joined #openstack17:23
*** Ryan_Lane has joined #openstack17:24
kd926MarkusWPHRorg: can rPath export qemu compatible images?17:24
*** jheiss has joined #openstack17:25
*** berto- has quit IRC17:26
*** llang629 has joined #openstack17:26
*** llang629 has left #openstack17:26
MarkusWPHRorgkd926: Sorry I'm very new to rPath17:27
MarkusWPHRorgVERY new. But the outfit I work for ais a customer of theirs17:27
MarkusWPHRorgso I was hoping some folks from there might be hanging out here so we could chat17:28
kd926the outfit wouldn't happen to have a green logo would it17:28
MarkusWPHRorgDon't think so17:29
kd926ah alright, just curious17:29
MarkusWPHRorgWho has a green logo?17:29
kd926so I am vaguely familiar with rPath17:29
*** jaypipes has quit IRC17:30
kd926do you have the infrastructure running somewhere at your outfit?17:30
*** MarcMorata has joined #openstack17:30
MarkusWPHRorgnot sure17:30
kd926you said they are a customer of rPath's product though?17:30
MarkusWPHRorgYeah, but I'm not sure if we have it functional yet17:31
MarkusWPHRorgand as far as Openstack is concerned, that's something I'm looking into17:31
*** jaypipes has joined #openstack17:31
*** berto- has joined #openstack17:31
*** CatKiller has quit IRC17:37
*** GeoDud has joined #openstack17:40
*** huslage has quit IRC17:43
*** katkee has quit IRC17:44
*** CatKiller has joined #openstack17:49
*** CatKiller has quit IRC17:50
*** CatKiller has joined #openstack17:51
*** circloud has quit IRC17:53
*** nacx has quit IRC17:53
*** negronjl_ has quit IRC17:54
*** ryker has quit IRC17:56
*** ryker has joined #openstack17:56
*** zul has quit IRC17:59
*** evil_e_ is now known as evil_e18:00
*** stephan` has left #openstack18:02
*** mgius has joined #openstack18:03
*** mdomsch has quit IRC18:07
*** jtran has quit IRC18:10
*** obino1 has joined #openstack18:14
*** AhmedSoliman has quit IRC18:16
*** obino has quit IRC18:17
*** j05h has quit IRC18:18
*** j05h has joined #openstack18:18
*** zul has joined #openstack18:21
*** RickB17 has quit IRC18:27
*** [1]RickB17 is now known as RickB1718:27
*** joearnold has quit IRC18:30
*** mihgen_ has quit IRC18:32
*** huslage has joined #openstack18:32
*** Glace has joined #openstack18:35
*** Glace is now known as Guest2852218:35
*** Guest28522 is now known as Glacee18:35
*** bastichelaar has joined #openstack18:36
*** kashyap has quit IRC18:37
*** med_out is now known as medberry18:41
*** skraps has quit IRC18:44
*** ctennis has quit IRC18:46
RaziqueHi, I need to run a fsck on a lvm volume but nova uses it18:47
Raziquehow can I disable it ? lvchange -an doesn't work18:48
*** MarcMorata has quit IRC18:54
*** zul has quit IRC18:55
*** NeCRoManTe1 has joined #openstack19:02
*** NeCRoManTe has quit IRC19:02
*** MarkusWPHRorg has quit IRC19:07
*** royh has joined #openstack19:09
*** RoAkSoAx has quit IRC19:09
*** RoAkSoAx has joined #openstack19:09
*** thinkscientist has joined #openstack19:11
*** katkee has joined #openstack19:12
*** zul has joined #openstack19:12
vishyRazique: is it attached to an instance?19:14
RaziqueI've stopped the whole nova cloud19:14
Raziquestopeed iscsi19:14
Raziquebut still unable to run a lvchange -an19:14
vishysomething must still have it open19:15
vishytry an lsof on it19:15
*** mrmartin has joined #openstack19:16
*** obino has joined #openstack19:20
*** thinkscientist has quit IRC19:21
*** obino1 has quit IRC19:24
*** 92AADFGZE has joined #openstack19:27
*** katkee has quit IRC19:27
*** morfeas has joined #openstack19:28
*** joearnold has joined #openstack19:29
*** katkee has joined #openstack19:29
*** ryker has quit IRC19:32
bastichelaaranyone here using LXC?19:36
*** ryker has joined #openstack19:38
uvirtbotNew bug: #806647 in nova "NBD device doesn't get disconnected after terminating LXC instance" [Undecided,New]
*** obino has quit IRC19:42
*** obino has joined #openstack19:43
*** syah has joined #openstack19:48
*** ctennis has joined #openstack19:53
*** mies has quit IRC19:55
uvirtbotNew bug: #806653 in nova "expose utils.usage_from_instance in nova api" [Undecided,New]
*** zul has quit IRC19:59
vishybastichelaar: is that your bug/patch?20:01
devcamcarvishy: reading your ha-network bits20:03
devcamcarvishy: sweet stuff20:03
devcamcarvishy: is vlan support in scope?20:03
vishydevcamcar: vlan mode should work but it is costly20:03
devcamcarvishy: how so?20:04
vishybastichelaar: thanks for finding that issue, i think it needs to be reworked a little bit but that is close to the right fix20:04
devcamcarvishy: does br100 go away?20:04
vishyin vlan mode, it will create all vlans on all hosts20:04
vishyand every host needs an ip in every project network, so it burns a lot of ips20:05
vishynot that big of a deal for a 12 node cluster20:05
bastichelaarvishy: thanks. I'm not a good programmer, but it seems that the LXC stuff is really basic and not used a lot, right?20:06
vishybut for a 255+ that is killing a /24 for every project20:06
WormManI had no problem not giving my compute nodes IPs in the added network20:06
WormManbut I didn't try VLAN20:06
*** jtran has joined #openstack20:06
WormManer in the project network20:06
vishybastichelaar: most people using LXC are not using cow images i think20:06
vishyWormMan: true in the current code, we're talking about the ha-net branch20:06
vishyusing multi_host networks20:07
bastichelaar vishy: what are they using then? I'm quite new in this area...20:08
*** tonycampbell has joined #openstack20:08
heckjvishy: where do the floating IPs get allocated in the ha-net branch? Spread around the various node, or collocated with the VMs?20:09
vishyfloating ips are allocated to the host that the vm is on20:10
vishyand natted to the fixed_ip20:10
vishybastichelaar: --nouse_cow_images20:10
*** tonycampbell is now known as tcampbell20:10
vishywill use losetup mounts instead of nbd20:10
vishyshould be a bit faster20:10
bastichelaarvishy: thanks, I just discovered that option :)20:10
vishyin any case your bug should be fixed so thanks for reporting it20:11
bastichelaarno problem, hope to find more bugs ;)20:11
*** msivanes has quit IRC20:12
bastichelaarbut it seems that the lxc implementation in libvirt is quite poor20:12
*** msivanes has joined #openstack20:13
*** willaerk has joined #openstack20:13
*** dprince has quit IRC20:26
*** jtran has quit IRC20:28
*** winston-d has quit IRC20:37
*** bpaluch has joined #openstack20:41
*** HouseAway is now known as AimanA20:43
*** dendrobates is now known as dendro-afk20:51
*** mies has joined #openstack20:54
*** msivanes has quit IRC20:59
*** lorin1 has quit IRC21:04
vishybastichelaar: is it missing some things that you need?21:04
*** TREllis has joined #openstack21:09
*** aliguori has quit IRC21:12
*** aliguori has joined #openstack21:12
TREllisdoes nova bind a metadata service IP to an interface?21:13
*** mrmartin has quit IRC21:18
*** huslage has quit IRC21:18
bastichelaarvishy: not really, but I'm struggling with a bug that causes libvirt_lxc to generate 100% cpu load when I start an instance21:20
vishyhmm that is nasty21:20
vishydo you have the newest libvirt?21:20
bastichelaarthis is the output of the log file21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaar23:02:11.837: 3722: error : lxcFdForward:287 : read of fd 7 failed: Input/output error21:20
bastichelaarsorry for spamming, should have used pastebin :)21:20
WormManahh, the classic read from a closed pipe bug that every software product has had at least once :)21:20
bastichelaarand it generates this message every millisecond, so the disk is filling up21:21
bastichelaarbut I cant get my head aroung this bug21:21
bastichelaaror configuration error21:21
bastichelaardont know yet21:21
WormManrun it under strace and see what's going on? (if you're trying to troubleshoot it) may be a bit too low level though21:22
*** NeCRoManTe1 has quit IRC21:22
bastichelaarok, I will try to create a wrapper around libvirt_lxc and output the strace in a file21:22
*** NeCRoManTe has joined #openstack21:23
*** _vinay has joined #openstack21:26
TREllishmmm can't ssh to instance and notice it has issue contacting metadata service, virsh console shows "error: internal error character device (null) is not using a PTY" on the instance, any ideas?21:35
*** kbringard has quit IRC21:35
*** patcoll has quit IRC21:37
*** titaniumrain has joined #openstack21:43
titaniumrainanyone there?21:44
*** bcwaldon has quit IRC21:45
*** dendro-afk is now known as dendrobates21:46
*** bcwaldon has joined #openstack21:47
_vinayHi, I am having trouble injecting  data into an instance21:48
_vinaylibvir: QEMU error : Domain not found: no domain with matching name 'instance-00000001'21:49
_vinay libvir: Network Filter error : Network filter not found: no nwfilter with matching name 'nova-instance-instance-00000001-secgroup21:49
_vinayany idea how to go about debugging this21:49
_vinayfrom compute logs I see following error also21:51
_vinayCommand: sudo qemu-nbd -c /dev/nbd15 /nova/..//instances/instance-00000002/disk21:51
_vinayStderr: "qemu-nbd: Could not access '/dev/nbd15': No such file or directory\n")21:51
*** bcwaldon has quit IRC21:52
*** matiu has joined #openstack21:53
*** Deirz- has quit IRC22:00
bastichelaarIm not the expert here, but it looks like libvirt didnt start the instance22:00
bastichelaarlook in the logs before the entry you posted22:00
*** londo_ has joined #openstack22:01
*** ldlework has quit IRC22:09
ohnoimdeadi'm awesome22:22
*** berto- has quit IRC22:22
devcamcarmy hero22:23
* ohnoimdead is a pro22:23
*** berto- has joined #openstack22:24
*** dendrobates is now known as dendro-afk22:26
*** bastichelaar has quit IRC22:26
*** nati has joined #openstack22:28
*** amccabe has quit IRC22:30
*** joearnold has quit IRC22:32
*** 92AADFGZE has quit IRC22:32
vishyvirsh console doesn't work normally on instances22:32
*** markvoelker has quit IRC22:32
*** lool has quit IRC22:32
*** jonkelly has quit IRC22:33
*** jaypipes has quit IRC22:33
*** jkoelker has quit IRC22:34
*** Shentonfreude has quit IRC22:35
*** dsockwell has quit IRC22:36
*** lool has joined #openstack22:36
*** dsockwell has joined #openstack22:37
*** joearnold has joined #openstack22:38
*** katkee has quit IRC22:44
*** matiu has quit IRC22:45
*** jaypipes has joined #openstack22:45
*** matiu has joined #openstack22:45
*** miclorb has joined #openstack22:46
*** AhmedSoliman has joined #openstack22:47
*** lvaughn has quit IRC22:51
*** jaypipes has quit IRC22:51
*** lorin1 has joined #openstack22:52
*** tcampbell has quit IRC22:54
*** NeCRoManTe has quit IRC22:55
*** technicool has joined #openstack22:57
*** nati has quit IRC22:58
*** nati has joined #openstack22:58
*** joearnold has quit IRC22:58
*** joearnold has joined #openstack22:58
*** jaypipes has joined #openstack23:01
*** matiu__ has joined #openstack23:03
*** matiu__ has quit IRC23:03
*** dgags has quit IRC23:03
*** matiu has quit IRC23:05
*** willaerk has quit IRC23:08
*** hggdh has quit IRC23:11
*** ewindisch has joined #openstack23:15
*** hggdh has joined #openstack23:16
*** Vasichkin has quit IRC23:19
*** zedas has joined #openstack23:20
*** AhmedSoliman has quit IRC23:21
*** nati has quit IRC23:24
*** mihgen has joined #openstack23:25
*** medberry is now known as med_out23:26
*** deshantm_laptop has quit IRC23:26
zedashey, what part of the code figures out it needs to load the config file: /etc/swift/account-server/4.conf23:29
*** huslage has joined #openstack23:31
*** Ephur has quit IRC23:33
*** mihgen has quit IRC23:34
notmynamezedas: swift-init calls swift.common.manager. finds the conf files (I'm pretty sure this is where it gets found and called)23:36
*** joearnold has quit IRC23:36
*** joearnold has joined #openstack23:36
*** joearnold has quit IRC23:40
notmynamezedas: so `swift-init account start` ends up finding and starting all (4) account servers in the all-in-one23:40
notmyname(called from Server.launch() in swift.common.manager)23:41
zedasnotmyname: thanks, and then that's doing a search of the tree and looking for a specific number of config files that comes from.....?23:45
notmynamezedas: unless the number is given it finds (and therefore launches) an instance for every conf file. so 4 conf files == 4 servers launched. the number can get passed in by swift-init (-c or --config-num)23:49
*** adjohn has joined #openstack23:50
*** tcampbell has joined #openstack23:53
zedasnotmyname: ok great, thanks for the guidance23:57

Generated by 2.14.0 by Marius Gedminas - find it at!