Monday, 2011-06-13

*** rchavik has quit IRC00:04
*** rchavik has joined #openstack00:05
*** dragondm has quit IRC00:28
*** Binbin has joined #openstack00:38
*** cloudgroups has joined #openstack00:39
*** cloudgroups has left #openstack00:40
*** obino has quit IRC00:47
*** thickskin has quit IRC00:50
*** yui has joined #openstack00:51
*** miclorb_ has quit IRC00:53
*** thickskin has joined #openstack00:54
*** obino has joined #openstack00:57
*** obino has quit IRC01:04
Ryan_LaneI'm getting 500 errors from my metadata service. basically, it is doing requests, and the api server is throwing 500s saying "NotFound: No floating ip for address <address of nova-network's host network device>"01:10
Ryan_Lanethe metadata service uses the client ip as the remote_address, but when doing DNAT, that address is the address of the network server, not the instance01:11
Ryan_Laneanyone have any ideas?01:12
*** obino has joined #openstack01:23
*** katkee has quit IRC01:24
*** chilts has joined #openstack01:35
*** adjohn has joined #openstack01:39
*** dendro-afk is now known as dendrobates01:44
*** KnightHa1ker has joined #openstack01:49
*** KnightHa1ker has left #openstack01:49
*** Eyk^off has joined #openstack01:51
*** KnightHa1ker has joined #openstack01:57
*** winston-d has joined #openstack01:59
*** KnightHa1ker has joined #openstack02:00
*** KnightHacker has quit IRC02:03
*** KnightHa1ker has quit IRC02:05
*** KnightHacker has joined #openstack02:05
*** KnightHacker has left #openstack02:05
winston-dBK_man : hi02:06
*** Ephur has joined #openstack02:08
*** Ephur has quit IRC02:16
*** dendrobates is now known as dendro-afk02:24
*** obino has quit IRC02:31
*** Ephur has joined #openstack02:31
uvirtbotNew bug: #796404 in swift "Cannot install swift-account on the same machine with swift-proxy" [Undecided,New]
*** gaveen has joined #openstack02:43
*** gaveen has joined #openstack02:43
*** osier has joined #openstack02:50
*** ton_katsu has joined #openstack02:55
*** zns has quit IRC03:04
HugoKuo__How to do Instance network HA ?  What if a nova-network host dead .....03:11
*** joearnold has joined #openstack03:13
*** Binbin has quit IRC03:17
*** obino has joined #openstack03:29
HugoKuo__in flatDHCP , is that possible to change instance's gateway ....03:34
*** zns has joined #openstack03:34
*** bluetux has joined #openstack03:36
*** AimanA is now known as HouseAway03:45
*** jedi4ever has joined #openstack03:46
*** csgeek has joined #openstack03:48
*** jedi4ever has quit IRC03:55
*** kashyap has joined #openstack04:09
*** joearnold has quit IRC04:19
*** Ephur has quit IRC04:32
*** f00b0t has joined #openstack04:39
f00b0tHas anyone here tried to deploy swift on FreeBSD?04:39
*** zns has quit IRC04:49
*** Ephur has joined #openstack04:58
*** shehjart has joined #openstack05:00
*** kennethkalmer has quit IRC05:01
*** ctennis has quit IRC05:07
OutBackDingof00b0t: hrmmm now i wonder05:09
OutBackDingointeresting concept05:10
vishyRyan_Lane: yeah your packets are probably being snatted05:19
vishyRyan_Lane: it SNATs outbound traffic to the nova-network ip.  If your api is on a diffeernt machine, you need to add an ACCEPT before the snat rule to make sure that it doesn't get SNATTED to the network host's ip05:23
*** hagarth has joined #openstack05:24
*** hggdh has quit IRC05:27
*** Binbin has joined #openstack05:31
*** miclorb has joined #openstack05:49
*** pygi has quit IRC06:17
*** guigui1 has joined #openstack06:19
*** mgoldmann has joined #openstack06:27
*** koolhead17 has joined #openstack06:32
*** Ephur has quit IRC06:35
*** wilmoore has joined #openstack06:43
*** kennethkalmer has joined #openstack06:47
*** obino has quit IRC06:49
*** f00b0t has quit IRC07:08
*** yui has joined #openstack07:10
*** reidrac has joined #openstack07:16
*** miclorb has quit IRC07:18
*** pygi has joined #openstack07:23
*** Razique has joined #openstack07:23
*** burris has quit IRC07:25
*** Ryan_Lane has quit IRC07:31
*** salmanearn has joined #openstack07:36
salmanearnhi guys i am getting some problem in configuring openstack compute ,can i ask here07:38
salmanearni am a newbe to openstack please suggest07:40
*** dobber has joined #openstack07:47
*** byeager has quit IRC07:48
*** byeager has joined #openstack07:48
Raziquesalmanearn: just ask :)07:49
salmanearni am following
salmanearni am, stuck at  Setting up a volume group07:51
salmanearnplease do you know what does it mean replace    /dev/loop0 below with whatever losetup returns07:55
Raziquethe step consists in creating a file seen as a disk by using /dev/loopX07:57
Raziqueyou could have some loops already taken, so choose an empty slot07:57
salmanearnnova-volumes is the default for the --volume_group flag vgcreate nova-volumes /dev/loop007:58
salmanearnnova-volumes: command not found07:59
salmanearnthis is the out put i am getting07:59
salmanearnany idea07:59
salmanearnNo physical volume label read from /dev/loop2   Physical volume "/dev/loop2" successfully created   Volume group "nova-volumes" successfully created08:00
salmanearnNow i am moving to Setting up your user08:01
salmanearn<Razique> Thanks for your help will update soon after lunch :)08:03
Raziquesalmanearn: :)08:04
Raziquenova-volumes is not a command, it's the name of the VG :)08:04
*** kennethkalmer has quit IRC08:05
*** wilmoore has quit IRC08:06
*** kennethkalmer has joined #openstack08:09
*** allsystemsarego has joined #openstack08:14
*** allsystemsarego has joined #openstack08:14
*** adjohn has quit IRC08:15
*** gohko has joined #openstack08:19
*** hggdh has joined #openstack08:19
*** mcclurmc_home has joined #openstack08:33
*** llang629_ has joined #openstack08:34
*** MarkAtwood has joined #openstack08:39
*** cuzoka has joined #openstack08:53
*** MarkAtwood has quit IRC08:56
*** bluetux has quit IRC08:57
*** mcclurmc_home has quit IRC08:58
salmanearn<Razique> . novarc09:08
salmanearnits not working09:08
Raziquesalmanearn: what do u mean ?09:08
salmanearnUnzip and source credentials09:09
salmanearn<Razique>You'll see these files extract:09:09
salmanearn<Razique>Then type or copy/paste the following to source the novarc file in your current working directory.09:10
salmanearn<Razique> . novarc09:10
salmanearn<Razique> :/$ sudo . novarc sudo: : command not found09:14
Raziquewhen you unzipped, did u see the files ?09:20
salmanearnthe same out put09:20
Raziquenow as root09:20
Raziquedid u sourced the novarc ?09:21
salmanearnpls eloberate how to source09:21
Raziqueas root, you type ". novarc"09:21
salmanearnhmm sudo . novarc09:22
Raziqueit won't work09:22
salmanearncommand not found09:22
Raziquejust . novarc09:22
Raziquebut as root09:22
Raziquesudo -s first09:22
salmanearnpls 2 min i let you know09:23
salmanearnthanks a lot for your help09:24
salmanearnwhat work09:25
salmanearn<Razique> there is one more thing which i was not able to sort out earlier09:27
Raziqueyup ?09:27
salmanearncd /opt/nova-2011.1/CA sudo ./genrootca.sh09:27
salmanearnDownload credentials for your user/project09:28
Raziquedoes /opt/nova.. exists ?09:28
salmanearnlet me chack09:29
salmanearnno its empty09:30
Raziqueok so you have to figure out where the script is09:32
Raziquefind / -type -f -name ""09:32
Raziquefind / -type f -name ""09:32
salmanearnits showing the path of a downloaded nova script09:35
salmanearn<Razique> but its didnt use that script09:35
Raziqueseems the right one09:36
Raziqueactually, the guy who wrote the tuto extracted the file into /opt09:36
Raziquethat just that :)09:36
salmanearnif i do that it should be fine i guess09:37
HugoKuo__while I add a child zone09:39
HugoKuo__what's the username and password ?09:39
HugoKuo__OS ?09:39
Raziquewhat do you mean by a child zone ?09:39
HugoKuo__or another nova cluster's cloud admin ?09:39
Raziqueyou mean a project ?09:39
HugoKuo__no , multizones09:40
salmanearnThanks <Razique>09:40
HugoKuo__a feature of NOVA09:40
*** HugoKuo__ has quit IRC09:42
*** HugoKuo has joined #openstack09:42
RaziqueHugoKuo: sorry, I thought I was talking to salmanearn :p09:42
HugoKuoRazique : it's ok :>09:43
*** irahgel has joined #openstack09:43
*** gaveen has quit IRC09:45
RaziqueHugoKuo: sorry I don't know multizones well09:45
*** Binbin has quit IRC09:50
*** salmanearn has quit IRC09:56
*** philippkueng has joined #openstack09:57
*** salmanearn has joined #openstack10:02
*** dirkx has joined #openstack10:03
*** carlp has quit IRC10:04
*** katkee has joined #openstack10:06
*** gaveen has joined #openstack10:08
*** Guest26577 is now known as Adri200010:08
*** Adri2000 has joined #openstack10:08
*** yui has quit IRC10:19
*** kd9261 has quit IRC10:22
*** gaveen has quit IRC10:24
*** kennethkalmer has quit IRC10:24
HugoKuoanyone familiar with Zones feature?10:25
*** dirkx has quit IRC10:33
*** osier has quit IRC10:36
*** gaveen has joined #openstack10:37
*** zul has quit IRC10:45
*** salmanearn has quit IRC11:09
*** rchavik has quit IRC11:10
*** kim0_ is now known as kim011:11
*** kashyap has quit IRC11:11
*** kim0 has quit IRC11:11
*** kim0 has joined #openstack11:11
*** alexander_ has joined #openstack11:15
alexander_i'm an openstack beginner, i am here to fix that :)11:16
alexander_on the compute node, is it required to have nova-network? most of config examples out there present in nova.conf the line ' --dhcpbridge=/usr/bin/nova-dhcpbridge11:18
alexander_ '11:18
*** kennethkalmer has joined #openstack11:22
*** kennethkalmer has quit IRC11:25
*** kennethkalmer has joined #openstack11:27
*** pmullaney has joined #openstack11:29
*** kashyap has joined #openstack11:30
*** Adri2000 has quit IRC11:34
*** dovetaildan has quit IRC11:35
*** dovetaildan has joined #openstack11:39
*** med_out is now known as med11:39
*** med is now known as medberry11:40
*** jfluhmann has quit IRC11:42
HugoKuo     a issue about Zones11:49
*** Razique has quit IRC11:55
*** Razique has joined #openstack11:58
*** irahgel has quit IRC12:03
*** dovetaildan has quit IRC12:04
*** dovetaildan has joined #openstack12:06
*** irahgel has joined #openstack12:20
*** ziyadb has joined #openstack12:21
ziyadbcan nova do SAN? iirc swift requires DAS?12:21
*** ryker has joined #openstack12:22
*** dprince has joined #openstack12:25
*** hagarth has left #openstack12:29
*** ziyadb_ has joined #openstack12:29
*** ziyadb has quit IRC12:31
*** burris has joined #openstack12:33
*** PeteDaGuru has joined #openstack12:35
*** kennethkalmer has quit IRC12:35
*** philippkueng has joined #openstack12:37
*** philippkueng has quit IRC12:44
*** iammartian has joined #openstack12:46
*** iammartian has left #openstack12:46
*** dprince has quit IRC12:49
*** philippkueng has joined #openstack12:50
*** ziyadb_ has quit IRC12:50
*** ton_katsu has quit IRC12:52
*** ziyadb has joined #openstack12:53
*** ziyadb has joined #openstack12:53
*** osier has joined #openstack12:53
*** ameade has joined #openstack12:54
*** ziyadb_ has joined #openstack12:57
*** philippkueng has quit IRC12:59
*** ziyadb has quit IRC13:00
*** allsystemsarego has quit IRC13:03
*** philippkueng has joined #openstack13:05
ziyadb_would it make sense to use a storage array and provision it as required for nova and swift?13:10
notmynameziyadb_: (I can't speak for nova.) that will work for swift. all swift needs is storage volumes. however, that probably won't be optimal. your storage in swift should be focused on $/GB (and secondarily on IOps)13:12
ziyadb_so the optimal solution would be what? DAS SATA drives?13:12
ziyadb_get a cluster of servers for DAS purposes? wouldn't getting a few single servers (to make the storage requests) and a storage array be more efficient?13:13
ziyadb_since the computing power of the cluster would be useless?13:14
notmynameit's not useless since there are swift processes running on all those CPUs. it's just that you get that for "free" when you optimize for the storage13:15
ziyadb_hmm, good point.13:16
ziyadb_can you help me shape an idea of what kind of storage is required for nova? I'm not going to quote you on it :)13:16
notmynamethere are ways to optimize a large cluster that may mean you pick different hardware for different parts, but the biggest single cost in a swift cluster is hard drives13:17
*** philippkueng has quit IRC13:17
notmynameso you should make those as cheap as possible13:17
*** hadrian has joined #openstack13:17
notmynameIMO, there's no point arguing if you need some extra RAM or CPU if it's only 5% of the cluster cost13:17
notmynamewhen drives are a _much_ bigger percentage of the total13:17
ziyadb_yeah, noted.13:18
*** imsplitbit has joined #openstack13:18
ziyadb_now I Just need to know what nova's optimal requirements are. I can't find this info anywhere.13:18
notmynameso really, it depends on what you're looking for. if you need a poc and have a SAN lying around, go for it. if you want a huge production cluster, we can give more guidance for more optimal installs13:19
ziyadb_yea, we're rolling out a production install. just assessing whether getting  SAN for nova/swift would be a better idea, but apparently not.13:21
ziyadb_might still get one if nova requires it, assuming of course, that it has a better price:performance numbers than DAS.13:21
notmynameif you can share, what's the order of magnitude of scale for the swift cluster? will it be public or private?13:22
*** irahgel has quit IRC13:22
*** philippkueng has joined #openstack13:22
ziyadb_public for the most part, not big, main purpose is to offer object storage for business customers (we're a CLEC with IaaS). so we'll probably start around 30-50TB.13:23
*** dgags has joined #openstack13:24
notmynamegreat. if you are able, keep us informed on how the deploy is going. :-)13:26
ziyadb_heh, will do.13:28
*** markvoelker has joined #openstack13:29
kpeppleziyadb_: nova-volume only supports a few combinations of storage, so you might want to check source code and blogs on that ...13:30
*** dendro-afk is now known as dendrobates13:30
*** amccabe has joined #openstack13:30
*** lborda has joined #openstack13:31
*** philippkueng has quit IRC13:31
*** Adri2000 has joined #openstack13:33
*** Adri2000 is now known as Guest8761913:33
*** jfluhmann has joined #openstack13:34
*** philippkueng has joined #openstack13:37
*** jfluhmann has quit IRC13:37
*** irahgel has joined #openstack13:37
*** zul has joined #openstack13:40
*** jfluhmann has joined #openstack13:42
ziyadb_kpepple ah, I see! I will look into it. IF anyone else has more information on what types of storage are ideal for nova, feel free to speak up :)13:45
*** gondoi has joined #openstack13:46
*** Guest87619 is now known as Adri200013:48
*** Adri2000 has joined #openstack13:48
*** philippkueng has quit IRC13:48
kpeppleziyadb_: check out and for some details of what people are using ... i have heard others using DRDB or Nexenta13:49
RaziqueHi all13:52
Raziqueevery time I try to launch a custom ami, I've this error
*** philippkueng has joined #openstack13:54
kpeppleRazique: what does the error code at the end of line 8 say --- can't read French "Erreur d'entr\xc3\xa9e/sorti"13:55
Raziquehi kpepple it means Input/outpur error13:56
kpeppleRazique: what OS are you on ?13:56
uvirtbotNew bug: #796619 in nova "Revoke cert by user and project broken" [Undecided,In progress]
Raziquethe custom AMI is a debian squeeze13:56
kpeppleRazique: are you mod-probing nbd in your custom AMI ? otherwise, you will see something like
Raziquekpepple: here is the content of my rc.local file
RaziqueI think it's what you were suggesting about mod-probing ?13:59
*** jatsrt has joined #openstack14:00
kpeppleRazique: not a debian expert, but shouldn't you be mod-probing nbd in line 3 or 4 ?14:00
Raziquewhat ndb means ?14:01
Raziquenbd sorry14:01
ziyadb_kpepple awesome, thanks!14:02
kpeppleRazique: network block device ... it allows you to write to the instance image (
kpeppleziyadb_: np .. would love to know how you go on your install, am working with a number of installations facing similar choices14:03
*** philippkueng has quit IRC14:03
*** alexander__ has joined #openstack14:03
*** jkoelker has joined #openstack14:04
Raziquekpepple: in all the doc I've been accros, for UEC or openstack14:04
Raziquethey all use this method for key injection, never seen the nbd method actually14:04
ziyadb_kpepple yeah, I'll let #openstack know all about it :) I also think a book might be in the cards.14:04
kpeppleRazique: hmmm ... can't say i'm an expert at the images, but it's strange that you are retrieving the key from the 169 network but can't write it to your filesystem14:07
Raziquekpepple: actually, the nova setup contains a prerouting rule that catches all the hit to that IP and redirect the the nova-api server14:08
alexander__hello, 2 minutes of your time if it's possible :)14:09
kpeppleRazique: yeah, usually when you see an issue during startup like this it is a networking issue ... but here it is not14:09
Raziqueoh, here is my libvirt.xml for that domain14:09
Raziquelooks like there is a network definition issue ,14:10
alexander__on the compute noe, where only nova-compute is installed, the line --dhcpbridge=/usr/bin/nova-dhcpbridge14:10
alexander__ is valid?14:10
alexander__in nova.conf14:10
Raziqueyes, because I'm able to run images that I get from uec.canonical.com14:11
Raziqueinstantiation works pretty well14:11
*** fabiand__ has joined #openstack14:11
*** murkk has joined #openstack14:11
kpeppleRazique: have you tried adding the mod-probe nbd line into your debian image ?14:11
Raziquekpepple: not yet, would you give me a link about what I should add ?14:12
*** murkk has quit IRC14:12
*** murkk has joined #openstack14:12
*** philippk_ has joined #openstack14:13
*** dendrobates is now known as dendro-afk14:13
kpeppleRazique: i think it is just 'modprobe nbd' after line 2 in your file14:14
Raziquethanks, 'ill give a try :)14:15
kpepplealexander_: i don't think dhcpbridge does anything on nova-compute ... it just handles db updates from the dhcp server14:15
alexander__i understand, thanks14:16
kpepplealexander_: as i read the code, it is just checking the dhcp lease file and updating the nova db when things change ... so this would really only be applicable on the nova-network nodes14:17
*** smoser` is now known as smoser14:17
alexander__i saw that line in most of the config examples, and i was confused as nova-dhcpbridge14:18
alexander__ wasn't in /usr/bin14:18
*** Shentonfreude has joined #openstack14:18
kpepplealexander_: it should be installed wherever openstack installs it ... probably with the other nova-* binaries14:19
alexander__ok, i understand14:19
alexander__that line should go only on the controller, userd by nova-compute network *14:20
Raziquekpepple: It did not change anything14:21
*** philippk_ has quit IRC14:21
kpepplealexander_: it can probably go on every machine, but it will only be used on the nova-network machine ... i use pretty much the same nova.conf on all my machines14:21
RaziqueI used that link in order to create my custom AMI :
*** alexander__ is now known as Staetu14:22
Razique(without the whole initrd and vmlinuz extraction)14:22
*** llang629_ has left #openstack14:22
Staetui understand now, thank you14:23
kpeppleRazique: really ? hmmm ... not sure what else causes an input/output error14:23
ziyadb_hmm, looks like an iscsi SAN is a viable option for nova.14:23
Raziquemm when I use kvm to create the image, I specify the network driver -net nic,vlan=0,model=virtio,macaddr=00:16:3e:de:ad:0114:24
Raziqueso maybe forcing virtio usage causes the issue14:24
Raziquemmm not really, I compared my libvirt.xml to a working one14:25
jatsrtRazique: Still working on it?14:25
jatsrthavn't been around for a week14:26
jatsrthow's it going14:26
Raziquejatsrt: na, the cloud works perfectly fine now :)14:26
RaziqueI'm trying to create custom AMI's now :)14:26
Raziquebut I can''t go through my debian ami14:26
StaetuIs it possible to deploy openstack on a pc with 1 lan card using the vlan mode? The dhcp mode will start ok14:26
jatsrtRazique: Ahh, I've only modified ubuntu images at this point14:27
Raziquejatsrt: maybe it's the same procedure for both OS N14:27
jatsrtAre you modifying a running AMI?14:27
Raziquehere is the link I'm working on :
kpeppleRazique: did you tweak the fstab like they say ? i am wondering if there isn't a filesystem there due to a resize ...14:28
Raziqueno I'm starting from scratch14:28
jatsrthmm, I've taken the stick UEC images and modified them without a problem14:28
*** hggdh has quit IRC14:28
jatsrtThey are pretty bare to begine with and with that and cloudinit gets a clean install14:28
*** philippkueng has joined #openstack14:29
Raziquekpepple: I have not used the tune2fs, but my fstab uses sda114:29
Raziqueoh, maybe that's the issue14:29
*** alandman has joined #openstack14:29
Raziquenova creates disks called vdX, not sdX14:29
*** hggdh has joined #openstack14:30
Raziquei'll try to modify the fstab and using vda1 instead of sda114:30
jatsrtRazique: do you have a paste of the console output when it boots14:31
Raziquethe console.log is empty at that stage unfortunately14:33
*** ziyadb_ has quit IRC14:34
*** ziyadb has joined #openstack14:34
*** ziyadb has joined #openstack14:34
Raziquei'm updating the images repo with the new version ^^14:35
Raziquemm no success14:36
*** kennethkalmer has joined #openstack14:36
*** kennethk_ has joined #openstack14:37
Raziquei'll enable the verbose mode14:37
Raziquemaybe i'll have more traces14:37
*** jonkelly has joined #openstack14:38
Raziquemmm not better...14:40
*** kennethkalmer has quit IRC14:41
*** lvaughn has joined #openstack14:41
*** ziyadb_ has joined #openstack14:42
*** ziyadb has quit IRC14:42
*** Refresher has joined #openstack14:45
*** mldennis has joined #openstack14:46
RefresherBeen looking at OpenStack storage for Camb Uni.  Can anyone by any chance tell me if there is a way of using, say, Cloud storage as a back end to OpenStack?  The idea is to have an "elastic storage cloud" where the elasticity is provided by someone else.  :-)14:46
*** mldennis has left #openstack14:48
*** ziyadb_ has quit IRC14:49
*** ziyadb has joined #openstack14:49
*** dragondm has joined #openstack14:50
*** primeministerp has quit IRC14:51
*** philippkueng has quit IRC14:57
*** MarcMorata has joined #openstack14:58
*** primeministerp has joined #openstack14:59
*** staetualex has joined #openstack15:01
*** hggdh has quit IRC15:03
*** ziyadb_ has joined #openstack15:03
RaziqueRefresher: sorry, but I don't understand at all :D15:04
Raziquecould you please elaborate ?15:04
RefresherOpenstack appears to be wonderful at having a number of storage nodes which can all store objects which can be presented to users.15:05
RefresherThis is great.  What I want to know is if I can add, not a storage node (as in a computer with hard drives) but a link to, say, Amazon S3.15:05
RefresherThis means that if I get a sudden burst of storage needs I can overflow to S3, until such time as I can buy more storage15:06
*** alexander__ has joined #openstack15:06
ziyadb_so what happens to a customer's VM when a fault zone fails? (I am assuming that the vms are not done in parallel, correct me if I'm mistaken)15:06
*** ziyadb has quit IRC15:06
*** hggdh has joined #openstack15:06
RefresherIf that was to me, I am not planning on using the storage for VMs15:06
*** alexander__ has quit IRC15:08
*** alexander_ has quit IRC15:08
*** Staetu has quit IRC15:09
*** heckj has joined #openstack15:10
RaziqueRefresher: I think the project Swift allows you to add an S3 node15:11
Razique(but I'm not sure at 100%)15:11
*** reidrac has quit IRC15:12
Refresherok, that is something I can try and find.  If anyone else wakes up before I go, I would appreciate other people chipping in.15:12
RefresherIt makes OpenStack go from appealing to convincing.15:12
notmynameRefresher:  Razique: no. there is no such thing as "burstable storage" in swift at this time15:12
notmynamebut it's something I've been thinking about a little bit. it's got some cool possibilities15:13
*** ziyadb_ has quit IRC15:13
Raziquenotmyname: ok, thanks for the correction15:13
notmynameRefresher: we've just about finished a feature that allows a container to be synced between 2 swift clusters. it's not what you are referring to, but it's the closest thing we have to that now15:13
RefresherIt appears to be something which iRODS supports, but that has its own set of problems.15:14
Refreshernotmyname - unless a commercial supplier out there is going to be able to offer this to me, that is unlikely to be too useful.15:14
*** guigui1 has quit IRC15:14
*** ziyadb has joined #openstack15:14
*** ziyadb has joined #openstack15:14
Refreshersorry - I mean not that useful to _my_ use case.15:15
*** mdomsch has joined #openstack15:15
notmynameit's something I'd like to see (although I don't know if it would be s3, per se). currently, each swift cluster is autonomous. you can expand it very easily with 0 downtime (we constantly do it at rackspace). but I understand if you have different requirements15:16
notmynameRefresher: Razique: you may have been thinking about glance. it can store VMs in both swift and s3 (and locally and ....)15:18
ziyadbnotmyname is that in reply to my question?15:19
*** obino has joined #openstack15:19
notmynameziyadb: if it answers your question, then yes ;-)15:19
Raziquenotmyname: yah, you right I mixed both projects on the S3 part, sorry :)15:19
notmynameziyadb: actually, I don't think I saw your question15:20
*** ziyadb_ has joined #openstack15:20
ziyadb_notmyname I presume data is written redundantly to aforementioned "autonomous" clusters in a fashion similar to RAID?15:21
ziyadb_to safeguard against cluster failures that is.15:21
RefresherS3 was merely meant as an example - I don't particularly care who it is, so much as that it be possible with a provider that I can use.15:21
RefresherMy plan is to offer large scale but slow storage over a RESTful interface.15:22
notmynameziyadb: you mean across clusters? no. well, I supposed it /could/ be, but that is completely up to the client to manage. the swift cluster doesn't know about anything but itself (except for the limited case of the currently-in-dev container sync feature)15:22
RefresherThe expandability of OpenStack storage is one of its appeals though.  Just sometimes it takes over a month to get new hardware, and we don't always get that much notice of people wanting a service.15:23
*** ziyadb has quit IRC15:23
ziyadb_notmyname well not across *clusters* more of across fault zones, or physical servers, to prevent single points of failure.15:23
notmynameRefresher: it could be done with a swift cluster and fall over to Rackspace Cloud Files or Internap's storage (since they are both swift clusters). that would allow you to use the same API. you'd just have to manage the failover with your auth system or your client15:23
*** dobber has quit IRC15:24
notmynameziyadb_: oh. yeah, totally. swift stores <replica count> copies (default to 3) throughout the cluster. it also has the concept of availability zones to allow the deployer to control the level to which each copy is isolated15:24
ziyadb_simple enough, now if I could get a similar answer with regards to nova.15:25
notmynameziyadb_: currently, swift will only store one copy per zone15:25
*** hggdh has quit IRC15:25
ziyadb_it is quite trivial, but I have not gotten an answer so far. nova provisions VM instances on a single physical host? well, what happens when it fails?15:25
ziyadb_when the physical host fails to be specific.15:26
RefresherOoh.  That means I need to look at Rackspace and Internap's offerings.  I need at the least something which will reside in the EU.  I also have reason to believe that Eduserv may be about to offer an OpenStack offering which would be helpful.15:27
notmynameI /think/ in that case you're SOL. you need to handle HA at a higher layer than nova15:27
notmynameRefresher: I think nephoscale also has a public swift cluster. I don't know any details about internap or nephoscale, but rackspace does have a london cluster15:28
ziyadb_and what are you hinting at by "a higher layer than nova"? It's an application :) I don't think you get any higher than that.15:28
notmynameziyadb_: out of luck15:28
ziyadb_BTW, thanks for taking the time to answer my/our questions, much appreciated.15:29
notmynameziyadb_: I see nova as the infrastructure controller. the "higher layer" is the application running on the instances provided by nova15:29
*** dolph has joined #openstack15:30
*** hggdh has joined #openstack15:30
ziyadb_ohh..right! understood.15:30
Refreshercurrently "live chatting" to a salesdroid at rackspace.  :-)15:31
RefresherNot that Farouk seems to be awake.  :-(15:32
*** obino has quit IRC15:32
RefresherI take that back, he replied as soon as I was mean about him.  :-)15:33
RefresherBrilliant.  Thanks for your help, esp notmyname.  :-)15:40
*** Refresher has left #openstack15:40
uvirtbotNew bug: #796660 in nova "xenserver migration 'list index out of range' when SR isn't named 'slices'" [Undecided,In progress]
*** Guest44298 has quit IRC15:42
*** dragondm has quit IRC15:52
*** dragondm has joined #openstack15:53
*** odyi has joined #openstack15:59
*** sophiap has joined #openstack15:59
*** odyi has quit IRC15:59
*** odyi has joined #openstack15:59
*** llang629 has joined #openstack16:03
*** obino has joined #openstack16:04
*** llang629 has left #openstack16:05
*** kbringard has joined #openstack16:08
*** Razique has quit IRC16:08
*** joearnold has joined #openstack16:09
*** vladimir3p has joined #openstack16:11
*** maplebed has joined #openstack16:21
*** osier has quit IRC16:23
*** katkee has quit IRC16:24
*** joearnold has quit IRC16:24
*** dolph has quit IRC16:25
*** briarrose has joined #openstack16:28
*** dolph has joined #openstack16:29
*** briarrose has quit IRC16:31
*** mgoldmann is now known as mgoldmann|away16:31
*** cuzoka has quit IRC16:31
*** kashyap has quit IRC16:33
*** joearnold has joined #openstack16:33
*** irahgel has left #openstack16:36
*** bcwaldon has joined #openstack16:42
*** ziyadb has joined #openstack16:48
*** ziyadb has joined #openstack16:48
*** ziyadb_ has quit IRC16:50
*** kashyap has joined #openstack16:54
*** Ryan_Lane has joined #openstack16:59
uvirtbotNew bug: #796709 in nova "OSAPI: delete server instance response code does not match spec" [Undecided,New]
*** zul has quit IRC17:02
*** jdurgin has joined #openstack17:02
*** mgius has joined #openstack17:05
*** hggdh has quit IRC17:12
*** allsystemsarego has joined #openstack17:16
*** allsystemsarego has joined #openstack17:16
*** aliguori has quit IRC17:16
*** ziyadb has quit IRC17:18
*** ziyadb has joined #openstack17:19
*** dendro-afk is now known as dendrobates17:21
*** heckj has quit IRC17:24
*** GeoDud has joined #openstack17:25
*** jatsrt has quit IRC17:29
*** blamar has quit IRC17:31
*** maplebed has quit IRC17:31
*** heckj has joined #openstack17:33
*** pguth66 has joined #openstack17:33
*** johnpur has joined #openstack17:33
*** ChanServ sets mode: +v johnpur17:33
*** maplebed has joined #openstack17:35
*** medberry is now known as med_out17:41
*** lborda has quit IRC17:45
*** lborda has joined #openstack17:45
*** dolph has quit IRC17:45
*** aliguori has joined #openstack17:48
*** dgags has quit IRC17:49
*** zul has joined #openstack17:50
*** katkee has joined #openstack17:52
*** MotoMilind has joined #openstack17:53
*** obino has quit IRC17:54
*** obino has joined #openstack17:54
*** imsplitbit has quit IRC17:56
*** ziyadb has quit IRC17:57
*** dgags has joined #openstack17:57
*** ziyadb has joined #openstack17:57
*** ziyadb has joined #openstack17:57
*** sophiap has quit IRC17:58
*** blamar has joined #openstack17:59
*** pguth66 has quit IRC18:02
*** pguth66 has joined #openstack18:02
*** ziyadb_ has joined #openstack18:06
*** obino has quit IRC18:07
Ryan_Lanevishy: is there some way for nova to set that rule for me?18:07
Ryan_Lanevishy: can you also give me an iptables rule example of what you mean?18:07
vishyRyan_Lane: no but we shoul probably add it :)18:07
Ryan_LaneI guess it's assumed most people run the api and network on the same host?18:07
vishyi guess :)18:08
*** obino has joined #openstack18:08
Ryan_LaneI did previously too, but network decisions made it easier to have it separate18:08
*** aliguori has quit IRC18:08
Ryan_Lanevishy: btw, you going to be at velocity?18:08
*** ziyadb has quit IRC18:09
vishywe have custom rules for dmz which must be fixing things.  Not quite sure why we didn't have to add that rule specifically18:09
*** koolhead17 has joined #openstack18:10
Ryan_Lanethis rule goes on the network node, in the prerouting chain?18:10
uvirtbotNew bug: #796742 in nova "OSAPI 1.1: Create server instance response code does not match spec" [Undecided,New]
vishyRyan_Lane: if you set --dmz_cidr to a range that includes your network host it should fix it18:12
Ryan_Laneah. great18:12
vishyi.e. create it automatically18:12
vishyRyan_Lane: or you can create it manually with iptables -t nat -A nova-network-POSTROUTING -s <fixed range> -d <nova-api-ip>/32 -j ACCEPT18:14
vishyRyan_Lane: gl I'm off to bed18:14
Ryan_Lanevishy: thanks for the help18:14
vishysorry the --dmz_cidr needs to include your api-hosts (as in the same address that you are forwarding 169.254 to18:15
Ryan_Laneah. ok.18:15
*** gaveen has quit IRC18:16
vishyit should probably create a rule for the api host in addition to the --dmz_cidr18:16
vishyif that works, could you file a bug for it?18:16
vishyif we are creating the forwarding rule we should probably create the accept as well18:16
* Ryan_Lane nods18:16
Ryan_Lanewill do18:16
vishycool, night!18:16
*** aliguori has joined #openstack18:25
*** vernhart has joined #openstack18:26
*** MarkAtwood has joined #openstack18:26
*** clauden has joined #openstack18:26
*** ziyadb_ has quit IRC18:27
*** carlp has joined #openstack18:28
*** KnightHacker has joined #openstack18:28
*** KnightHacker has left #openstack18:29
*** dolph has joined #openstack18:31
*** ziyadb has joined #openstack18:32
*** ziyadb has quit IRC18:41
*** briarrose has joined #openstack18:41
*** joearnold has quit IRC18:42
*** briarrose has quit IRC18:43
*** ziyadb has joined #openstack18:44
*** ziyadb has joined #openstack18:44
*** clauden has quit IRC18:45
*** clauden_ has joined #openstack18:46
*** clauden___ has joined #openstack18:47
*** clauden_ has quit IRC18:47
kbringardhmmm, looks like the glance_host and glance_port directives in the nova.conf are no longer valid (intentionally or otherwise)?18:47
kbringardI put an auth proxy in front of glance, and locked it down to no longer listen on
*** clauden___ has quit IRC18:47
*** clauden has joined #openstack18:48
kbringardit only listens on an internal IP18:48
kbringardand euca-describe-images works when the proxy is running, but it doesn't when it's not18:48
*** clauden has quit IRC18:48
kbringardeven though glance_host and glance_port are set to the internal in the nova.conf18:48
kbringardI grepped through the code and didn't see anywhere where those flags are referenced18:48
*** carlp has quit IRC18:49
*** clauden has joined #openstack18:49
*** carlp has joined #openstack18:53
*** clauden has quit IRC18:53
*** zul has quit IRC18:58
*** zul has joined #openstack18:58
*** kashyap has quit IRC18:59
*** ewindisch has joined #openstack19:00
*** sebastianstadil has joined #openstack19:03
*** brd_from_italy has joined #openstack19:05
*** timrc has quit IRC19:10
*** dolph has quit IRC19:12
*** mgoldmann|away is now known as mgoldmann19:13
*** timrc has joined #openstack19:14
*** dolph has joined #openstack19:15
*** alandman has quit IRC19:20
*** vernhart has quit IRC19:21
*** med_out is now known as medberry19:23
*** ziyadb has quit IRC19:31
*** theron has joined #openstack19:34
*** zul has quit IRC19:35
uvirtbotNew bug: #796785 in nova "Remove clocksource=jiffies from xenserver configurations" [Undecided,New]
*** vladimir3p has quit IRC19:37
*** coyled has left #openstack19:39
*** nelson____ has joined #openstack19:40
*** 77CAAPWAA has joined #openstack19:42
*** fabiand__ has quit IRC19:49
*** zul has joined #openstack19:52
*** clauden_ has joined #openstack19:52
*** ameade has quit IRC19:55
*** kbringard has quit IRC20:01
*** kbringard has joined #openstack20:02
*** joearnold has joined #openstack20:02
*** llang629_ has joined #openstack20:04
*** sebastianstadil_ has joined #openstack20:05
*** sebastianstadil has quit IRC20:05
*** sebastianstadil_ is now known as sebastianstadil20:05
*** bcwaldon has quit IRC20:07
*** zul has quit IRC20:10
*** danishman has joined #openstack20:10
*** kbringard_ has joined #openstack20:11
*** kbringard has quit IRC20:11
*** kbringard_ is now known as kbringard20:11
*** zns has joined #openstack20:14
*** theron has quit IRC20:18
staetualexare there any limitations/things to consider when running opensack on machines with only one nic(virtual interfaces bounded to separate vlans)20:21
staetualexvlan mode20:21
*** mgoldmann has quit IRC20:22
*** sebastianstadil has quit IRC20:23
*** miclorb_ has joined #openstack20:23
*** hydra_ has joined #openstack20:23
*** vladimir3p has joined #openstack20:23
*** GeoDud has quit IRC20:27
*** dolph has quit IRC20:27
*** koolhead17 has quit IRC20:36
*** koolhead17 has joined #openstack20:36
*** fabiokung has quit IRC20:41
uvirtbotNew bug: #796813 in nova "vmwareapi does not support distributed vswitch" [Undecided,New]
*** fabiokung has joined #openstack20:44
*** bkkrw has joined #openstack20:47
*** nphase has quit IRC20:48
*** Ryan_Lane has quit IRC20:48
*** Ryan_Lane has joined #openstack20:48
*** allsystemsarego has quit IRC20:51
uvirtbotNew bug: #796820 in nova "Race condition causes unit test to occasionally fail" [Undecided,New]
*** mdomsch has quit IRC20:52
medberryThe SAIO emulates a four node cluster, but isn't it also emulating a proxy as well (so really four nodes plus a server)?20:53
notmynamemedberry: yes20:55
notmynamebut it's 4 storage nodes20:55
medberrynod, tx.20:56
*** jsm has joined #openstack20:57
*** Shentonfreude has quit IRC20:58
jsmi've been reading documentation trying to find example's of openstack's UI so I can show some folks.  If it's available where should I look?20:58
*** zns has quit IRC20:59
*** zns has joined #openstack21:02
*** hydra_ has quit IRC21:02
*** kennethk_ has quit IRC21:03
*** ccustine has joined #openstack21:10
devcamcarjsm: there is a simple django based ui at
*** gondoi has quit IRC21:16
*** danishman has quit IRC21:17
*** HouseAway is now known as AimanA21:18
*** GeoDud has joined #openstack21:21
uvirtbotNew bug: #796834 in nova "ESX(i) : Incorrect machine id parameters in vmx for" [Undecided,New]
annegentlejsm: there's a video demonstration of that Dashboard at
*** matiu has joined #openstack21:26
jsmdevcamcar & annegentle, thanks!21:27
*** llang629_ has quit IRC21:29
*** aliguori has quit IRC21:31
*** aliguori has joined #openstack21:31
*** KnightHacker has joined #openstack21:31
*** KnightHacker has left #openstack21:32
*** llang629_ has joined #openstack21:39
*** llang629 has joined #openstack21:40
*** llang629_ has quit IRC21:44
*** markvoelker has quit IRC21:48
*** zns has quit IRC21:48
*** zns has joined #openstack21:53
*** kbringard has quit IRC21:53
*** jonkelly has quit IRC21:59
staetualexyou should try hybrid fox22:10
*** llang629 has left #openstack22:11
*** matiu has quit IRC22:17
*** matiu has joined #openstack22:18
*** matiu has joined #openstack22:18
*** vernhart has joined #openstack22:18
*** jfluhmann has quit IRC22:23
*** dendrobates is now known as dendro-afk22:28
*** jsm has quit IRC22:32
*** dendro-afk is now known as dendrobates22:32
*** _vinay has joined #openstack22:33
_vinayI am trying to associate a IP address to a VM created in single node nova installation22:34
_vinayso that I can access it from another host22:34
_vinayI have in my nova.conf --public_interface=eth0  and --vlan_interface=eth022:35
_vinayis that going to be a problem?22:35
*** brd_from_italy has quit IRC22:35
_vinaythe admin manual says I should have --public_interface=vlan100 (7.3.6 Associating a public IP address)22:36
_vinaydoes it mean change to vlan100 for both the configs??? or only for --public_interface22:37
_vinayand keep the --vlan_interface=eth0 ???22:37
_vinayI am a bit confused at that22:37
staetualexwhat the euca-describe-availability-verbose and log files are saying?22:39
*** johnpur has quit IRC22:39
_vinaystaetualex    is that for me??22:40
staetualexeuca-describe-availabilit-zones verbose "22:40
*** PW_ has joined #openstack22:40
_vinaywhich log files should I look at ??22:42
*** vladimir3p has quit IRC22:42
*** katkee_ has joined #openstack22:42
staetualexnova-network there22:43
_vinayI dont have anything there ( /var/log/nova)22:45
_vinayI installed nova in my home dir  .. and then used ./ run22:45
staetualexwhat os do you run?22:45
_vinaythat thing creates 7 screen tabs22:45
*** katkee has quit IRC22:45
_vinayand all components are running there22:45
staetualexsee if in /etc/nova/nova.conf the line --verbose is present and see if there is a line for log directory22:46
_vinayubuntu 10.1022:46
*** mgius has quit IRC22:46
_vinay--verbose is present but nothing for log directory22:47
_vinayI can paste my nova.conf if you want22:47
staetualexok, you could try to add this to your nova.conf file and restart the nova services22:52
staetualexalso, mkdir /var/log/nova ; chown nova:root /var/log/nova22:53
_vinayok sure22:54
_vinayone more thing...  I started nova as ./ run22:55
_vinayand it creates a bunch of  screens from the screen_it module in nova.sh22:55
_vinaywhat is the best way to stop nova now22:56
*** PW_ has quit IRC22:56
_vinayI have tried terminate it doesnt work22:56
*** aliguori has quit IRC22:57
_vinayright now I manually kill all services and then restart22:57
staetualexsee if scrub works22:58
staetualexthat will kill virsh instances22:59
*** joearnold has quit IRC23:02
*** jkoelker has quit IRC23:02
*** amccabe has quit IRC23:06
*** dendrobates is now known as dendro-afk23:07
*** dragondm has quit IRC23:09
*** joearnold has joined #openstack23:10
*** ryker has quit IRC23:13
*** ryker has joined #openstack23:16
*** sandywalsh_ has joined #openstack23:21
*** 77CAAPWAA has quit IRC23:30
_vinayHi staetualex23:31
_vinayI restarted nova with the logs now...23:31
staetualexi'm here23:31
_vinayalso I associated public IP with the VM23:31
_vinayand can access the VM from the src host itself(where nova is running)23:31
staetualexdid the euca-authorize trick?23:32
_vinayhowever from other hosts in the network I cannot see23:32
_vinayI did euca-authorize as in the manual23:32
_vinayeuca-authorize -P icmp -t -1:-1 default23:32
_vinay euca-authorize -P tcp -p 22 default23:32
*** bkkrw has quit IRC23:32
staetualexadded iptables prrerouting line23:33
_vinayI didnt add them manually .. but they are there23:33
_vinayeuca-associate-address must have done it23:33
_vinay192.168.122.0   U     0      0        0 virbr023:34
_vinayThis is the route -N entry on my other host23:34
_vinaydont know if it looks correct23:34
_vinayIP assigned to VM is
_vinaydo you think thats right? ( gateway= ??)23:35
*** sebastianstadil has joined #openstack23:38
*** staetualex has quit IRC23:39
*** adjohn has joined #openstack23:41
*** adjohn has quit IRC23:44
*** rchavik has joined #openstack23:45
*** rchavik has quit IRC23:47
*** rchavik has joined #openstack23:47
*** zns has quit IRC23:49
*** victorozy has joined #openstack23:50
*** bluetux has joined #openstack23:50
*** victorozy has quit IRC23:51
*** AWR has joined #openstack23:51

Generated by 2.14.0 by Marius Gedminas - find it at!