Wednesday, 2011-05-25

*** dragondm has quit IRC00:02
*** jeffjapan has joined #openstack00:06
*** iRTermite has quit IRC00:08
*** iRTermite has joined #openstack00:09
*** zns has joined #openstack00:13
*** osier has joined #openstack00:15
*** stewart has quit IRC00:26
*** samba has quit IRC00:42
*** vernhart has quit IRC00:46
*** sebastianstadil has joined #openstack00:48
*** Binbingone is now known as Binbin00:50
*** kd926 has joined #openstack00:51
kd926anyone using nova on gentoo or rhel6.1?00:51
*** jdurgin has quit IRC00:55
*** mgius has quit IRC00:57
*** Ryan_Lane has quit IRC00:57
*** woleium has quit IRC01:01
*** cloudgroups has joined #openstack01:06
*** jheiss has quit IRC01:06
*** obino has quit IRC01:15
*** nelson has quit IRC01:16
*** nelson has joined #openstack01:16
*** kaz_ has joined #openstack01:18
*** cloudgroups has left #openstack01:25
*** alex_openstack has joined #openstack01:30
alex_openstackhowto custom windows image for nova?01:30
*** rchavik has quit IRC01:32
winston-dkd926 : i'm using Nova on RHEL 6.0, plan to upgrade to 6.1 in couple of days. what can i help?01:34
*** adjohn has joined #openstack01:38
kd926winston-d: did you install from rpms or source?01:38
winston-dfrom RPMS01:39
kd926which doc did you use?01:39
winston-dkd926 : doc? this
kd926ok. Did you encounter any issues with nova-manage project zipfile?01:46
kd926winston-d: the whole command01:46
winston-dkd926 : nope, i don't01:46
*** maplebed has joined #openstack01:47
kd926winston-d: do you have access to the box right now?01:47
winston-dkd926 : what box? my nova?01:47
*** hallyn is now known as hallyn_afk01:49
*** osier has quit IRC01:54
*** vernhart has joined #openstack01:57
*** obino has joined #openstack01:57
*** Binbin has quit IRC02:06
*** Arminder-Office has joined #openstack02:10
*** Arminder has quit IRC02:13
winston-dkd926 : sorry for late response. i was out for a meeting.  yes, i do have access to my nova box02:17
*** maplebed has quit IRC02:21
*** alice has joined #openstack02:22
*** Transformer has joined #openstack02:26
*** Transformer has quit IRC02:29
*** mdomsch has joined #openstack02:34
HugoKuo_well , I found a issue that while compute resource is been ran out , Nova still try to start instance on that compute node02:35
*** mfork00 has quit IRC02:37
*** sebastianstadil has quit IRC02:40
kd926winston-d: can you open a python interpretter and just try import utils for me?02:42
winston-dkd926 : something like: python; import utils ?02:43
winston-dHugoKuo_ : are you using ZoneScheduler?02:43
winston-dkd926 : is this what you want?
HugoKuo_winstond-d : is that the default one ?02:46
kd926winston-d: yep, that's what I've been seeing. thanks02:47
HugoKuo_winston-d : actually , I did not set scheduler driver flag in nova.conf............02:47
*** zns has quit IRC02:47
HugoKuo_So , It might using the default one02:47
winston-dHugoKuo_ : i don't know, since i'm using RHEL6 package, there's a default value in my nova.conf02:47
winston-dHugoKuo_ : you can use nova-mange --help to check, i guess02:48
HugoKuo_winston-d : nice recommnad , love ya02:48
HugoKuo_winston-d : so how's your issue yesterday , about limit instance numbers on comput node ...02:50
*** rpedde has joined #openstack02:51
*** gaveen has joined #openstack02:59
*** gaveen has joined #openstack02:59
winston-dHugoKuo_ : i switched to simple scheduler. can solve my problem for now.03:02
*** littleidea has joined #openstack03:04
HugoKuo_winston-d : thanks03:05
*** osier has joined #openstack03:05
*** rpedde has quit IRC03:09
*** stewart has joined #openstack03:25
HugoKuo_any clue about to check current service driver ?03:27
winston-dHugoKuo_ : what do you mean by 'service driver' ?03:28
HugoKuo_for example03:29
HugoKuo_I change to simple scheduler03:29
HugoKuo_can I use any cmd to check current scheduler driver?03:29
HugoKuo_winston-d : dude , thanks .03:35
HugoKuo_Openstack support virtual box image?    I got a serious problem with moving VM from virtualbox into OPENSTACK . Any idea ?03:37
*** AimanA is now known as HouseAway03:40
*** rchavik has joined #openstack03:42
*** rchavik has joined #openstack03:43
*** ccustine has joined #openstack03:57
*** kashyap has joined #openstack04:03
*** kashyap has joined #openstack04:04
*** stewart has quit IRC04:07
*** somik has joined #openstack04:08
*** j05h has quit IRC04:11
somikanybody around with multi-node deployment or glance-dev experience... I am running trunk across two hypervisors and a separate cloud controller node and am unable to spin-up VMs, somehow, the hypevisors not getting the image from the remote glance repo. I have the stack trace, if someone can help with it..04:11
kaz_I want to swift client supporting swauth04:12
kaz_I already test 2client. (cyberduck,cloudfuse)04:14
kaz_but this client dose not connect my swift-servers...04:15
*** mdomsch has quit IRC04:17
*** maplebed has joined #openstack04:21
zxvdrkd926, winston-d: jpwalters, mfork00 and I have all been having issues w/ RHEL 6.104:22
*** maplebed has quit IRC04:23
kpepplekaz_: you should paste the stacktrace on ... most of the US developers are gone for the day (especially jaypipes who leads glance) and will see it tomorrow morning ... i might be able to help, but i have limited glance-fu04:34
kpeppleerrr ... i think i meant somik04:36
*** maplebed has joined #openstack04:44
*** MarkAtwood has quit IRC04:46
somikkpepple: thanks - I just posted a question -
*** f4m8_ is now known as f4m804:47
somikkpepple: if you have a clue that would be cool, or you could tell what other info I should put up with the stack trace, that would be great too04:48
* kpepple looks at the question04:48
koolhead11hi all04:49
*** MarkAtwood has joined #openstack04:51
*** maplebed has quit IRC04:53
*** littleidea has quit IRC04:56
kpepplesomik: not totally sure what's going on there (you're setup is different from mine), but you are fetching a non-VHD image and it's streaming the disk image through Dom0 and it's timing out ... either due to the image not streaming or the device not being able to be created04:57
somikwhen does such a thing usually happen?04:58
somikkpepple: any idea when the device cannot be created? usually.. the nova-compute is talking to the correct glance server, so i cant imagine why the image wont stream.05:01
kpepplesomik: dunno. it seems like you issue might be in vm_utils:543 where it decides to stream through Dom0 instead of directly ... but not a Xen expert so maybe this isn't as suspicious as I think05:01
somikkpepple: shouldnt it stream direct05:01
kpepplesomik: you might want to try a VHD formatted image and see if you see the same problems ...05:01
somikkpepple: ok let me try that then. thanks for your help!05:02
kpepplesomik: np. good luck.05:02
kaz_kpepple: thanks  i upload
*** somik has quit IRC05:32
*** ccustine has quit IRC05:33
kpepplekaz_: which client is this ?05:34
kpepplekaz_: it is failing on authentication ? does it understand the openstack cloud files API ?05:35
kaz_kpepple: cloudfuse05:35
*** andy-hk has quit IRC05:36
*** andy-hk has joined #openstack05:36
*** sirp__ has quit IRC05:37
kaz_i useing cloudfuse05:38
kpepplekaz_: what are you using the very long auth_url in your .cloudfuse file ? shouldn't you use the one below it (you have commented it out) ?05:39
kpepplekaz_: also, you should look in your swift logs to see what  error you are seeing05:39
kaz_at the time swift(proxy-server) log in dose not logging05:41
kpepplekaz_: hmmm ... that makes it hard to see what specific error you are getting around your authentication. My guess is that your username, api_key and/or auth_url are incorrect.05:42
* kpepple leaves for a meeting05:43
kaz_ok . I re-try short auth url auth_url=
kaz_kpepple: thanks05:44
kaz_kpepple: I re-try and  paste short auth url
*** Binbin has joined #openstack05:58
kaz_when used with the cyberduck. swift - proxy logging HTTP status code 41205:58
kaz_  <- use cyberduck05:59
romankHi all Привет всем06:11
alex_openstackhow to custom windows image for nova and upload to nova?06:11
koolhead11alex_openstack, did you follow the documentation?06:12
alex_openstacknow i can create a winxp image with kvm06:12
*** zenmatt has quit IRC06:12
alex_openstackbut i not know how to upload to nova06:13
alex_openstackwhich documentation?06:14
alex_openstackeuca-bundle-image -i windowsserver.img06:15
alex_openstackeuca-upload-bundle -b mybucket -m /tmp/windowsserver.img.manifest.xml06:15
alex_openstackeuca-register mybucket/windowsserver.img.manifest.xml06:15
alex_openstackis this URL:   ?06:16
romankVery need help. I'm trying to set up swift on RHEL 6.1. When configuring swit-proxy on the 9 paragraph manual (
romankthe following error:06:22
romank[root@fsc2 swift]# swift-ring-builder account.builder rebalance06:22
romankTraceback (most recent call last):06:22
romank  File "/usr/bin/swift-ring-builder", line 616, in <module>06:22
romank    Commands.__dict__.get(command, Commands.unknown)()06:22
romank  File "/usr/bin/swift-ring-builder", line 477, in rebalance06:22
romank    parts, balance = builder.rebalance()06:22
romank  File "/usr/lib/python2.6/site-packages/swift/common/ring/", line 243, in rebalance06:22
romank    self._initial_balance()06:22
romank  File "/usr/lib/python2.6/site-packages/swift/common/ring/", line 394, in _initial_balance06:22
romank    while available_devs[index]['zone'] in other_zones:06:22
romankIndexError: list index out of range06:22
koolhead11obino, around06:23
koolhead11romank, kindly use pastebin :)06:23
romankok, sorry )06:23
*** thevoke has left #openstack06:24
koolhead11its just unreadable :)06:24
*** cloudgroups has joined #openstack06:29
*** dirkx has joined #openstack06:29
*** nerens has joined #openstack06:29
*** rcc has joined #openstack06:30
*** santhosh has joined #openstack06:33
*** fabiand__ has joined #openstack06:33
*** allsystemsarego has joined #openstack06:53
*** guigui1 has joined #openstack06:59
*** fabiand__ has left #openstack06:59
*** cascone has joined #openstack07:14
ttxannegentle, soren: saw ?07:19
ttxsurprisingly, it appears to be not too bad.07:19
ttxDaviey: ^07:20
*** nacx has joined #openstack07:20
*** cascone has quit IRC07:27
*** eikke has joined #openstack07:36
*** openpercept has joined #openstack07:42
*** dirkx has quit IRC07:43
*** pgregory has left #openstack07:48
*** bkkrw has joined #openstack07:49
*** MarkAtwood has quit IRC07:57
*** cloudgroups has left #openstack07:59
*** miclorb has quit IRC08:00
*** mancdaz has joined #openstack08:12
*** HugoKuo__ has joined #openstack08:19
*** AlexNeef has quit IRC08:23
*** AlexNeef has joined #openstack08:23
*** HugoKuo_ has quit IRC08:23
*** Eyk^off is now known as Eyk08:27
*** jeffjapan has quit IRC08:30
HugoKuo__Disk /dev/vdb: 10.7 GB, 10737418240 bytes08:38
HugoKuo__Disk /dev/vdb: 21.5 GB, 21474836480 bytes08:38
HugoKuo__where's /dev/vdb come from ?08:39
*** miclorb has joined #openstack08:43
sorenIt's your ephemeral storage.08:43
sorenIt's just extra space you can use.08:43
alex_openstackHugoKuo__: hi08:49
HugoKuo__alex_openstack: hi08:50
HugoKuo__Just say Hi to every one08:50
alex_openstackare you write the article for how to install windows as guest os on nova?08:50
HugoKuo__not yet , I'm working on Volume backup solution today08:51
alex_openstacki waiting for this,i can't let the instances running (winxp)08:52
HugoKuo__I'll test it tomorrow08:53
*** dobber has joined #openstack08:54
*** watcher has joined #openstack09:00
*** dobber has quit IRC09:02
*** fabiand__ has joined #openstack09:13
*** dobber has joined #openstack09:14
Davieyttx: yep!  They were at UDS.  We are helping them.09:18
*** geogdog has joined #openstack09:21
geogdogI'm new to openstack and need a little help.  I ran out of disk space when uploading by first image, and now it's stuck on status 'untarring'.  Any idea how to delete it?09:22
*** daveiw has joined #openstack09:23
*** daveiw has left #openstack09:26
*** irahgel has joined #openstack09:29
*** alice has quit IRC09:34
*** alex_openstack has quit IRC09:34
*** alice has joined #openstack09:34
*** irahgel has left #openstack09:37
nerensgeogdog: have you tried euca-terminate-instances?09:37
nerensgeogdog: sorry... image not instance... what about euca-deregister09:37
*** alice has quit IRC09:37
geogdogI'll give that a crack09:38
geogdognerens: thanks09:39
nerensgeogdog: no probs09:39
nerensdoes anyone have glance running with swift backend? care to share you config please, i'm having issues "InvalidURL: nonnumeric port" errors09:40
nerensparticularly swift_store_auth_address... would this be the correct format?
nerensand the URI from glance details all look like URI:
nerensshould that point to the api or the regisrty?09:42
*** miclorb has quit IRC09:44
*** fabiand__ has quit IRC09:48
*** irahgel has joined #openstack09:50
*** miclorb has joined #openstack09:51
*** reldan has joined #openstack09:54
*** pnavarro has joined #openstack09:54
*** justin_skariah has joined #openstack09:57
justin_skariahhello everyone i have a problem, i added a user(nova-manage user admin nova) i can see the user using the command nova-manage user list, but when i check the table user in database i am not able to see the user.10:00
uvirtbotNew bug: #788019 in swift "s3cmd swift3 authentication fails" [Undecided,New]
*** mattray has joined #openstack10:10
*** antenagora has joined #openstack10:11
*** antenagora has quit IRC10:15
*** irahgel has left #openstack10:19
*** irahgel has joined #openstack10:22
*** alice has joined #openstack10:26
*** BinaryBlob has joined #openstack10:30
*** reldan has quit IRC10:32
*** adjohn has quit IRC10:36
*** yes456 has quit IRC10:40
*** duffman has quit IRC10:48
*** miclorb has quit IRC11:00
*** mattray has quit IRC11:06
*** DodoPahim has joined #openstack11:07
*** newopen has joined #openstack11:13
newopenwho know about networking status , please11:14
*** kd926 has quit IRC11:14
newopenwhen i run a new instance, it's running :(11:14
newopeni mean it's not, it's networking11:14
*** koolhead11 has quit IRC11:14
*** AlexNeef has quit IRC11:15
*** duffman has joined #openstack11:17
*** antenagora has joined #openstack11:22
*** duffman has quit IRC11:23
*** duffman has joined #openstack11:25
*** duffman has quit IRC11:31
*** dijenerate has quit IRC11:33
*** dijenerate has joined #openstack11:34
*** keny has joined #openstack11:34
kenyI have a setup where my compute nodes launch VMs bridged to a public interface. I want my VMs to get public IPs.11:36
kenyI have built an image that will dhcp, and I can launch it fine from nova11:36
kenyproblem is, nova doesn't know what IP it gets11:37
kenyI don't want to have nova assign it a fixed IP, nor do I want nova to provide a dhcp server because I already have one... is this setup still possible?11:37
kenyCurrently, the images start and get an IP through dhcp, but nova somehow thinks its IP is (this network does not exist in nova-manage network list)11:40
*** BinaryBlob has quit IRC11:42
kenymaybe I need to setup two bridges? one for each of my interfaces, and have nova be happy with an internal IP address and have the VMs still get an external IP through the other interface?11:43
keny(I think I got a bit confused)11:43
*** dijenerate has quit IRC11:45
*** holden247 has joined #openstack11:52
*** dirkx has joined #openstack11:52
*** aliguori has joined #openstack11:59
*** holden247 has left #openstack12:02
ttxnewopen: usually means the network node fails to serve networking resourecs to the compute nodes. Look for hints of failure in nova-network.log or nova-scheduler.log12:02
ttxjustin_skariah: are you sure you look into the same database as nova is configured to use ?12:04
*** stewart has joined #openstack12:06
*** dirkx has quit IRC12:09
*** BinaryBlob has joined #openstack12:11
*** dijenerate has joined #openstack12:13
*** BinaryBlob has quit IRC12:16
*** reldan has joined #openstack12:18
*** DodoPahim has quit IRC12:20
*** adiantum has joined #openstack12:24
*** BinaryBlob has joined #openstack12:30
*** Arminder-Office has quit IRC12:30
*** DodoPahim has joined #openstack12:35
newopenhow make instances (subnet communicate with local network (
*** littleidea has joined #openstack12:37
*** AlexNeef has joined #openstack12:38
*** reldan has quit IRC12:39
*** dijenerate has quit IRC12:40
*** dijenerate has joined #openstack12:41
*** ctennis has quit IRC12:41
*** reldan has joined #openstack12:47
*** thatsdone has joined #openstack12:52
*** AlexNeef has quit IRC12:54
*** ctennis has joined #openstack12:55
*** ctennis has joined #openstack12:55
*** Shentonfreude has joined #openstack12:58
*** lvaughn has joined #openstack12:59
*** Shentonfreude has quit IRC12:59
*** primeministerp has joined #openstack12:59
*** Binbin is now known as Binbingone13:01
*** adjohn has joined #openstack13:02
*** thatsdone has quit IRC13:05
*** zenmatt has joined #openstack13:07
justin_skariahttx: i have done my complete setup(cloud controller+compute node) on my local fedora 14.. i checked the nova.conf it points to the local database..13:09
*** ameade has joined #openstack13:09
ttxjustin_skariah: strange. nova-manage doesn't magically invent its user list13:10
*** primeministerp has quit IRC13:11
*** osier has quit IRC13:14
sorenjustin_skariah: Are you sure nova-manage uses that nova.conf?13:14
*** dendro-afk is now known as dendrobates13:16
*** Eyk is now known as Eyk^off13:17
justin_skariahttx: on my system i have only one nova.conf @ /etc/nova/nova.conf
sorenjustin_skariah: ...and you're sure nova-manage uses it?13:18
*** stewart has quit IRC13:18
justin_skariahsoren : how to be sure if its using it?13:20
*** hallyn_afk is now known as hallyn13:21
*** adjohn has quit IRC13:21
sorenjustin_skariah: You could see if the user lists it outputs mathces your database :D13:21
newopen+soren:  how make instances (subnet communicate with local network (
sorennewopen: Why are you asking me?13:22
newopenok sorry13:22
sorenNo, really. Why?13:23
sorenDid we talk about this before?13:23
sorenI don't remember.13:23
justin_skariahsoren : that is my problem :( database is not getting updated..13:23
newopenabout networking yes13:23
newopen2 days above13:23
newopen2 days ago13:24
sorenjustin_skariah: Well, there you go. Try "nova-manage --flagfile=/etc/nova/nova.conf" and see if it behaves better.13:24
justin_skariahsoren : thanks i wil try that now..13:24
justin_skariahsoren : it worked.. user is getting updated in the database.. but how can this happen my system has only one nova.conf??13:26
sorenjustin_skariah: Why does the number matter?13:27
*** santhosh has quit IRC13:27
ttxsoren: which number ?13:29
sorenttx: Number of nova.conf's on the system.13:29
*** zns has joined #openstack13:29
justin_skariahfor nova-manage to work nova.conf should be present rite??13:29
sorenI just don't see why there only being one means anything.13:29
sorenjustin_skariah: nova-manage has defaults, so no.13:30
sorenjustin_skariah: They're probably wrong, though.13:30
sorenjustin_skariah: ...but strictly speaking, nova-manage works fine without a flagfile.13:30
*** mgoldmann has joined #openstack13:30
ttxsoren: the default is not to use nova.conf ?13:30
sorenThe default is to use a nova.conf found in the same directory as the binary being called (e.g. nova-manage).13:30
soren*dreadful* default, that one is.13:31
*** santhosh_ has joined #openstack13:31
*** foxtrotgulf has joined #openstack13:31
ttxsoren: haven't had the issue with ubuntu packaging... did you workaround that ?13:31
ttxthat said, I admit having always used the default backend13:32
ttxso maybe I just haven't run into it13:32
sorenttx: I do work around it in packaging, yes.13:32
justin_skariahsoren: got it.. /etc/nova/nova.conf is not the standard location of nova.conf.. can i make it default config?13:32
ttxjustin_skariah: so blame a bad default and a packaging that doesn't work around it.13:32
sorenOh, since 1066, it should find /Etc/nova/nova.conf without any help.13:33
sorenjustin_skariah: Which version of nova are you running?13:33
sorenGreat, I can remove this from the packaging. Whoo!13:34
justin_skariahsoren: i installed it from version 2011.3-dev fedora packaging..13:35
sorenzul, Daviey: Still no uploads of glance, nova, and swift to Oneiric. Shall I just do this myself?13:35
zulsoren: yeah go ahead im still stuck in blueprint purgatory13:35
sorenzul: Awesomesauce :)13:35
sorenjustin_skariah: I have no idea what 2011.3-dev corresponds to. It sounds made up.13:36
*** jfluhmann has quit IRC13:37
zulsoren: ill pick it up again next week13:37
justin_skariahsoren: i did a nova-manage --version and i got OpenStack Nova version: 2011.3-dev (2011.3-nova:tarmac-20110507153655-ilxtckixt5o03yzy)13:37
ttxsoren: opinion on bug 758133 ? Should we just tell people to use the build-depends ppa ?13:38
*** BinaryBlob has quit IRC13:39
*** jfluhmann has joined #openstack13:40
*** f4m8 is now known as f4m8_13:42
sorenttx: Yes.13:42
sorenttx: As commented on the bug just now.13:43
sorenjustin_skariah: Yup, that's a smidgeon too old.13:43
sorenjustin_skariah: Two days, to be exact.13:43
sorenjustin_skariah: Tell whoeever maintains that repository to update it.13:43
sorenReally, if people are packaging dev snapshots, they should be keeping them *very* up-to-date.13:44
justin_skariahsoren : yup i will do that.. thank you very much for the help..13:44
*** littleidea has quit IRC13:46
*** Shentonfreude has joined #openstack13:46
*** dendrobates is now known as dendro-afk13:46
*** dprince has joined #openstack13:46
*** amccabe has joined #openstack13:46
*** gaveen has quit IRC13:47
sorennewopen: I'm not sure, but I think just getting your routing sorted out should do the trick.13:47
*** zenmatt has quit IRC13:50
*** santhosh_ has quit IRC13:52
*** dprince has quit IRC13:55
*** koolhead17 has joined #openstack13:55
*** Razique has joined #openstack13:56
Raziquehi all13:56
Raziqueevery time I'm trying to sync the database from the comptute node, i've the following error13:57
RaziqueCommand failed, please check log for more info13:57
Raziqueand when I look into the nova-manage log, I see :13:58
*** justin_skariah has quit IRC14:02
*** openpercept has quit IRC14:02
*** littleidea has joined #openstack14:04
*** mgoldmann has quit IRC14:13
*** asksol has quit IRC14:14
*** Capashen has joined #openstack14:16
*** asksol has joined #openstack14:20
*** littleidea has quit IRC14:20
*** robbiew has joined #openstack14:21
*** littleidea has joined #openstack14:22
*** troytoman-away is now known as troytoman14:23
*** tjikkun has quit IRC14:23
*** tblamer has joined #openstack14:24
*** jkoelker has joined #openstack14:24
*** Arminder has joined #openstack14:26
*** alice has quit IRC14:26
*** lborda has joined #openstack14:27
CapashenDoes OpenStack has an equivalent of Amazon EBS ? I see in the blog "Nova now supports volumes residing on HP SANs and Solaris iSCSI devices." Is it directly a connection from Nova to iSCSI devices, without swift at all ?14:27
*** littleidea has quit IRC14:29
annegentlettx: I'm working with the authors of that beginner's guide to bring it into our doc set, it has great chapters that I'll mix in, CC licensed. Working on the integration now.14:29
*** dirkx has joined #openstack14:29
*** maplebed has joined #openstack14:30
*** littleidea has joined #openstack14:31
*** dijenerate has quit IRC14:34
*** mdomsch has joined #openstack14:39
ttxannegentle: cool :)14:40
ttxCapashen: there is an internal volume service that can plug iSCSI, sheepdog and various other options14:42
*** dirkx has quit IRC14:43
ttxCapashen: creiht is working on decoupling that and provide a volume as a service thing (project LunR)14:43
*** antenagora has joined #openstack14:46
*** dendro-afk is now known as dendrobates14:46
annegentleShermanBoyd: I have been trying to hunt down the origin of cc_host for a while now14:48
annegentleI think it may have existed in Bexar but I don't know if it's in Cactus, so I need to update the docs. Do you have the bandwidth to log a bug, tagged as documentation, in Nova?14:49
newopenhow to make traffic from br100 passed by eth014:50
newopenshould i add route14:50
*** dijenerate has joined #openstack14:50
newopen+soren: do you know which route should i add14:52
ShermanBoydannegentle: is this good?:
*** littleidea has quit IRC14:59
*** dragondm has joined #openstack15:00
annegentleyep, thanks. Back in Feb. I asked, "If I still have --cc_host in a nova.conf file does it still get used, or do the --ec2_blah settings take over?" but the logs on appear to be 404ing right now...15:04
*** j05h has joined #openstack15:05
uvirtbotNew bug: #788169 in openstack-manuals "cc_host flag is undocumented" [Undecided,New]
uvirtbotNew bug: #788168 in nova "Socket Closed on 50 simultaneous deletes" [Undecided,New]
*** littleidea has joined #openstack15:07
*** geogdog has left #openstack15:07
*** kashyap has quit IRC15:08
Capashenttx, will LunR project be in diablo release ?15:08
ShermanBoydis there an openstack forum?15:09
ttxCapashen: it should be supported, yes. It won't be a core project by the Diablo release, so as such it won't be in the coordinated Diablo release15:09
ttxShentonfreude: ?15:09
ttxShermanBoyd: ^15:09
*** rcc has quit IRC15:09
*** aliguori has quit IRC15:10
uvirtbotNew bug: #788175 in glance "No documentation for new results filtering in API or client" [Low,In progress]
*** troytoman is now known as troytoman-away15:11
*** bkkrw has quit IRC15:13
newopeni need to access from vms ( to another physical host (
ShermanBoydnewopen: describe your network15:15
*** rcc has joined #openstack15:15
*** alandman has joined #openstack15:18
newopenok i have a single machine installation (nova) (eth0= and network of vms is
*** sirp__ has joined #openstack15:18
newopeni have anothersingle machine installation (eth0=
*** openpercept has joined #openstack15:19
*** openpercept is now known as Guest5774215:20
newopeni want to reach from vms15:20
newopenShermanBoyd: is it clear?15:21
ShermanBoydnewopen: so everything is plugged into the same switch15:22
newopenwhich one?15:23
newopeni can reach the second machine ( from the first (
ShermanBoydnewopen: what does your network topology look like?  I'm assuming all machines are plugged into the same switch15:24
newopenthey are in the same subnet15:24
newopeni have just 2 hosts which are directly connected15:25
ShermanBoyddo you have routing_source_ip defined in nova.conf15:25
newopenvms ( in the first host are plluged inti br100 (
*** justin_skariah has joined #openstack15:26
newopeni want to ping from vms15:26
newopeni don't know which configuration should i apply to allow access15:27
*** ejat has joined #openstack15:27
ShermanBoydwhat is the gateway set to on the virtual machines?15:27
*** guigui1 has quit IRC15:27
*** ejat has quit IRC15:28
*** ejat has joined #openstack15:28
*** ejat has joined #openstack15:28
*** ejat has joined #openstack15:29
*** ejat has joined #openstack15:29
newopenyou mean br100 ip address=
*** reldan has quit IRC15:29
*** dobber has quit IRC15:31
*** ejat has quit IRC15:31
newopenShermanBoyd: does it what you mean?15:31
*** reldan has joined #openstack15:32
ShermanBoydnewopen: your physical hosts and vms are on different subnets so in order to communicate they need a router in between15:34
*** aliguori has joined #openstack15:34
ShermanBoydnewopen: this router needs to be pointed two in the VMs as the gateway15:35
*** ejat has joined #openstack15:35
*** mdomsch has quit IRC15:36
*** obino has quit IRC15:36
newopenyou mean i access with ssh to vm and add this route15:36
newopenin vm?15:36
ShermanBoydnewopen:  right add it as the default gw15:37
newopeni add eth0 ip address as default gateway?15:37
*** nerens has quit IRC15:37
ShermanBoydnewopen:  I assume15:37
*** rnirmal has joined #openstack15:38
*** RamD has joined #openstack15:38
newopen is the address of br100!15:38
newopen192.168.100.247 is the address of eth015:38
newopenvm'address is
ShermanBoydnewopen:  set the gw to
ShermanBoydor maybe openstack needs to assign another ip for routing15:39
alandmanShermanBoyd: How to you keep the br0 from using the gateway address? Is there a way to set a different address on br0 and the dnsmasq gateway?15:40
newopenok and this will permit me to ping from vm15:40
ShermanBoydjust to make it clear I'm just trying to figure things out too15:41
*** justin_skariah has quit IRC15:42
*** johnpur has joined #openstack15:42
*** ChanServ sets mode: +v johnpur15:42
alandmanShermanBoyd: My understanding is that using FlatDHCP nova-network launches dnsmasq to hand out IP addresses on the subnet15:43
*** prudhvi has joined #openstack15:44
ShermanBoydfrom that doc it seems like that the 2nd ip in a network is set as the router15:44
alandmanIt also adds ipchains rules to do snat so the computers can get out, setting itself to be the default gateway15:44
ShermanBoydmaybe that's a problem if that IP is the same as your bridge15:44
*** vernhart has quit IRC15:44
*** dendrobates is now known as dendro-afk15:45
*** jonkelly has joined #openstack15:45
newopen1# route add default gw
newopenroute: ioctl 0x890b failed: File exists15:47
alandmanCan anyone tell met what the --routing_source_ip flag in nova.conf does?15:48
*** r5 has joined #openstack15:48
alandmannewopen: What does a route -n return?15:48
ShermanBoydnewopen: can you ping
newopenyes of course15:52
newopenand others vms15:52
newopen --routing_source_ip= in nova;conf15:53
newopeni need to ping from vms15:53
*** stewart has joined #openstack15:54
*** maplebed has quit IRC15:54
newopeni can ping also (eth0 address)15:55
*** nerens has joined #openstack15:55
*** mRy has joined #openstack15:58
*** dprince has joined #openstack16:00
*** rnirmal_ has joined #openstack16:02
*** stewart has quit IRC16:03
newopenso? :(16:04
*** rnirmal_ has quit IRC16:04
*** obino has joined #openstack16:06
*** jonkelly has quit IRC16:06
*** rnirmal has quit IRC16:06
*** vernhart has joined #openstack16:06
*** rnirmal has joined #openstack16:06
alandmannewopen: So it looks like you correctly have the bridge so it seems like it might be an snat issue16:06
uvirtbotNew bug: #788211 in nova "Internal Server Errror from Compute on Create" [Undecided,New]
newopenyes i correctly have the bridge16:07
newopenbut what about snat, how can i configure it16:08
*** irahgel has left #openstack16:08
*** aliguori has quit IRC16:08
alandmannewopen: it's done through iptables rules16:09
*** jonkelly has joined #openstack16:09
alandmannewopen: You should have iptables running on
alandmannewopen: try iptables --list16:09
newopenyou mean
*** nacx has quit IRC16:10
alandmannewopen: yes16:10
*** antenagora has quit IRC16:10
newopenif i want to access from vm ( linked to (br) to
newopenwhich rule should i add?16:12
*** rchavik has quit IRC16:12
newopeniptables -A nova-network-??????16:12
alandmannewopen: That's I'm not sure... I've not used iptables in a long time and have to remember how it works16:14
*** jheiss has joined #openstack16:14
*** woleium has joined #openstack16:15
*** obino has quit IRC16:16
*** obino has joined #openstack16:16
*** RamD has quit IRC16:21
*** reldan has quit IRC16:23
*** dendro-afk is now known as dendrobates16:23
*** bhall_ has joined #openstack16:23
*** aliguori has joined #openstack16:24
*** bhall has quit IRC16:24
*** ejat has quit IRC16:26
*** zul has quit IRC16:26
*** ryker has quit IRC16:26
*** eikke has quit IRC16:26
*** ryker has joined #openstack16:26
*** ejat has joined #openstack16:27
*** ejat has joined #openstack16:27
*** reldan has joined #openstack16:27
*** duffman has joined #openstack16:28
*** maplebed has joined #openstack16:30
*** maple_bed has joined #openstack16:31
*** ejat has joined #openstack16:32
*** adiantum has quit IRC16:33
*** mgoldmann has joined #openstack16:34
*** maplebed has quit IRC16:35
*** dendrobates is now known as dendro-afk16:36
sirp__if anyone is interested in dist-sched work, we'd love some feedback on and
sirp__^^ meant for openstack-dev, repasting :)16:38
uvirtbotsirp__: Error: "^" is not a valid command.16:38
*** zul has joined #openstack16:42
*** samba has joined #openstack16:42
*** zul has quit IRC16:42
*** zul has joined #openstack16:42
*** aliguori has quit IRC16:44
*** kbringard has joined #openstack16:49
*** adiantum has joined #openstack16:49
kbringardhowdy howdy howdy16:49
*** sigos has joined #openstack16:49
*** DodoPahim has quit IRC16:51
*** Eyk^off is now known as Eyk16:52
*** dendro-afk is now known as dendrobates16:52
*** pnavarro has quit IRC16:54
*** DodoPahim has joined #openstack16:54
*** keny has quit IRC16:55
*** jdurgin has joined #openstack16:57
*** zenmatt has joined #openstack16:58
*** newopen has quit IRC16:58
*** primeministerp has joined #openstack17:00
sigosHello, I have a network problem with my openstack installation. From the controller, I am able to ssh into my VM instances. From my VM instances, I can ssh to any host in my network (and to the other VM instances). From any machine in my network (except the controller / the VM instances) when I try to ssh to any of my VMs, the ssh client freezes (last debug line is: channel 0: open confirm rwindow 0 rmax 32768) and the auth log file17:01
sigosanyone have an idea to help me?  thanks!17:01
*** Ryan_Lane has joined #openstack17:01
*** ccustine has joined #openstack17:01
*** primeministerp has quit IRC17:03
*** fabiand__ has joined #openstack17:05
*** watcher has quit IRC17:07
*** ejat has quit IRC17:08
*** ejat has joined #openstack17:08
*** ejat has joined #openstack17:08
*** dendrobates is now known as dendro-afk17:08
*** ejat has quit IRC17:08
*** dendro-afk is now known as dendrobates17:09
*** ejat has joined #openstack17:10
*** ejat has joined #openstack17:10
*** ejat has joined #openstack17:10
*** ejat has joined #openstack17:10
kbringardis there a way to force an instance to come up on a specific compute node?17:10
*** hallyn is now known as hallyn_afk17:10
*** ejat has joined #openstack17:11
*** ejat has joined #openstack17:11
*** ejat has quit IRC17:11
*** ejat has joined #openstack17:12
*** ejat has quit IRC17:12
*** ejat has joined #openstack17:14
*** ejat has joined #openstack17:14
annegentlekbringard: I'm pretty sure the answer is no, and that the work on a distributed scheduler and multi-cluster zones are part of the path to enable that scenario. See
kbringardah, cool17:18
kbringardthe way I do it now is I just fire up a crapton of instances at once17:18
kbringardand generally one lands on the node I need it to17:18
annegentlehehe, nothing like scattershot methods :)17:19
*** Guest57742 is now known as openpercept17:24
*** dendrobates is now known as dendro-afk17:24
*** openpercept has quit IRC17:24
*** openpercept has joined #openstack17:24
*** robbiew has quit IRC17:24
*** rnirmal has quit IRC17:24
*** robbiew has joined #openstack17:24
*** MarkAtwood has joined #openstack17:25
*** joearnold has joined #openstack17:25
*** ejat has quit IRC17:28
*** ejat has joined #openstack17:29
*** ejat has joined #openstack17:29
*** stewart has joined #openstack17:35
*** koolhead17 has left #openstack17:38
*** m0dpr0b3 has joined #openstack17:39
*** Capashen has quit IRC17:39
*** ejat has joined #openstack17:42
*** ejat has joined #openstack17:45
*** ejat has joined #openstack17:45
*** ejat has joined #openstack17:45
*** maple_bed is now known as maplebed17:46
*** ejat has joined #openstack17:47
*** ejat has joined #openstack17:47
*** dendro-afk is now known as dendrobates17:50
*** littleidea has quit IRC17:59
*** jedi4ever has joined #openstack18:00
*** somik has joined #openstack18:00
*** kashyap has joined #openstack18:01
*** koolhead17 has joined #openstack18:04
*** koolhead17 has joined #openstack18:04
*** koolhead17 has left #openstack18:07
*** littleidea has joined #openstack18:07
alandmanSo I have a question about nova-network18:10
*** tblamer has quit IRC18:10
alandmanWhen nova-network creates the network bridge interface it gives that interface the IP of the gateway18:12
alandmanIs there a way I can have it set a different IP address on the bridge and for the gateway?18:12
*** JuanPerez has joined #openstack18:13
*** dendrobates is now known as dendro-afk18:14
*** fabiand__ has quit IRC18:14
*** dendro-afk is now known as dendrobates18:15
*** RoAkSoAx has quit IRC18:20
*** mgius has joined #openstack18:22
*** Eyk is now known as Eyk^off18:22
ShermanBoydalandman: is this with flat or vlan?18:23
alandmanShermanBoyd: flat18:25
uvirtbotNew bug: #788265 in nova "Can't boot instance because of instance_type wrong pass" [Undecided,New]
*** rnirmal has joined #openstack18:28
*** aliguori has joined #openstack18:31
*** blamar_ has quit IRC18:32
*** drewlander has joined #openstack18:33
*** blamar_ has joined #openstack18:34
drewlanderhas anyone gotten openstack compute nodes working with hyper-V? and if so, is there any documentation? I have not been able to find anything online18:34
*** blamar_ has quit IRC18:37
drewlanderbasically, on a compute node, can I tell the compute node that the Hyper-V server(s) are at ip(s) or if the compute nodes have to be run on the hyper-v server, how to get it a compute node started in windows18:37
alandmanWe have HyperV compute nodes up and running18:37
alandmanYou need to run nova-compute on HyperV server18:38
*** primeministerp has joined #openstack18:38
alandmandrewlander: This is a starting point
drewlanderoh nice, I did not see this, thank you so much18:39
alandmandrewlander: But the instructions are far from complete or so comprehensive18:39
primeministerpgreetings programs18:39
alandmandrewlander: You will need some additional python packages18:39
*** blamar_ has joined #openstack18:39
drewlandercool. at least I have a starting point, this will absolutely help18:39
drewlanderthank you18:40
primeministerpdrewlander: we can prob share our install script18:40
*** newopen has joined #openstack18:40
drewlanderthat would be very helpful if you are able to do that18:40
primeministerpdrewlander: alandman spent a bunch of time working on it18:40
drewlandercool. thanks for pointing me to this document, cant believe i missed it18:42
newopeni have problem with nova-manage
*** blamar_ has quit IRC18:43
drewlanderalandman: fantastic, this will aboslutely help. thank you so much!18:43
primeministerpdrewlander: please run linux on your hyperv cloud18:43
*** blamar_ has joined #openstack18:44
alandmandrewlander: You will need to change some of the copies. You'll need to download swigin and mingw instead of copying them18:44
primeministerpdrewlander: w/ the linux lic's please18:44
*** clauden_ has joined #openstack18:44
drewlanderwill do.18:44
primeministerpdrewlander: awesome...  we have 70 instances running on 8 hyperv nodes right now18:45
drewlanderI saw microsoft is fully supporting CentOS as well on hyper-v , I thought that was quite good18:46
primeministerpdrewlander: yes that was a press release18:47
primeministerpdrewlander: there are packages in the opensuse build service for running centos enlightened18:47
primeministerpdrewlander: it's our msft collegues in the lab that do that dev work18:48
primeministerpdrewlander: for the linux lic's that is18:48
newopennova-manage doesn't work, no errors in log files18:49
*** dprince has quit IRC18:49
drewlanderthanks for the help again guys, I will let you know how it goes! much much appreciated18:51
*** kd926 has joined #openstack18:51
alandmandrewlander: Our nova.conf
kd926I am having issues starting the nova-compute service in RHEL6.118:52
kd926get an error "libvirtError: internal error Cannot find suitable emulator for x86_6418:52
drewlanderalandman: fantastic, thanks!18:52
primeministerpkd926: are you 32bit?18:53
kd926primeministerp: 6418:53
primeministerpkd926:  you sure18:53
primeministerpkd926: maybe you're missing a a package dependency18:54
primeministerpkd926: what's your nova.conf18:54
kd926primeministerp: probably have some unecessary flags in there18:55
*** stewart has quit IRC18:55
primeministerp you have to play w/?18:56
primeministerpkd926: that's a lot of real ips18:56
kd926primeministerp: its not attached to a physical nic18:56
kd926and its /2918:56
primeministerpkd926: yeah but those are real ips18:57
*** qdk_ has joined #openstack18:57
primeministerpkd926: which may exist out in the real world18:57
kd926primeministerp:  so far I haven't run across such a website... guess I'm lucky18:58
primeministerpkd926: yeah but why use something real when you don't have to18:58
kd926primeministerp: no reason in particular18:58
ShermanBoyddoes xen have live migration support yet in nova?18:58
kd926primeministerp: any ideas about the compute service by looking at the nova.conf?18:59
primeministerpkd926: only add what you need18:59
*** qdk_ has quit IRC18:59
primeministerpkd926: you have so much going on, including not using appropriate ips18:59
primeministerpkd926: i'd start by implementing just what you have, and make sure there are valid services listening19:00
kd926primeministerp: what is not needed19:00
kd926primeministerp: euca-describe-instance starts fine19:00
primeministerpkd926: that's a tool19:00
kd926primeministerp: everything functions except nova-compute19:00
*** zns has quit IRC19:00
kd926primeministerp: I should have said it returns instances waiting to be scheduled19:01
primeministerpkd926: i'm assuming everything is running on a single host?19:01
kd926primeministerp: yep19:01
*** vernhart has quit IRC19:01
primeministerpkd926: do you have rabbit-mq running?19:01
*** fabiand__ has joined #openstack19:01
*** woleium has quit IRC19:02
primeministerpprimeministerp: things tend to sit there when there's nothing to communication through19:02
kd926primeministerp: its running19:03
primeministerpalso you should prob remove most of your other bits, until you have each working19:03
primeministerpkd926: all that logging19:03
m0dpr0b3hello there, recently, I buy a dell, and I installed xen, I would like understand how openstack works, and how I can put it in production, please, what Openstack do ?19:03
primeministerpkd926: also you should turn up your logging level19:03
alandmanCan anyone tell me what the --routing_source_ip setting in nova.conf does or is for?19:03
primeministerpm0dpr0b3: i'd start by reading the wiki19:04
*** drewlander has quit IRC19:05
m0dpr0b3yes: I do , but what is openstack, what I will manage ? it is manager of virtual machines as xen or vmware  ?19:05
alandmankd926: What's in your nova-scheduler log?19:05
*** MarkAtwood has quit IRC19:05
primeministerpm0dpr0b3: all that's in the wiki and docs19:05
kd926alandman: compute:
primeministerpm0dpr0b3: you're asking a very general question, which is easily answered via the existing documentation.   I highly suggest you have a strong understanding of the OSI Model.  It will help in understanding where all the pieces fit in.19:07
primeministerpm0dpr0b3: to be quite honest it took me about 2 weeks of research to understand at a high level how all the parts work19:08
*** stewart has joined #openstack19:08
primeministerpm0dpr0b3: if you are just starting out w/ virtualization, openstack is probably a bit of an advanced leap conceptually which requires a strong understanding of advanced virtualization topics19:09
primeministerpm0dpr0b3: to be able to "get it" quickly19:09
primeministerpm0dpr0b3: in short it is a mgmt layer, which sits on top of your virtualization solution19:10
m0dpr0b3primeministerp: allright, thanks, your feedback is easy of understand and very intuitive :)19:10
primeministerpm0dpr0b3: i'm not trying to be discouraging, you just need to have a solid understanding of the fundamentals first19:11
primeministerpm0dpr0b3: because there's no easy answer19:11
ShermanBoydso I have a public routable ip on my cloud controller, if I'm setting up a "flat" network how do I bind br100 to that interface?19:11
m0dpr0b3primeministerp: Yes u have the reason :) again thank you19:11
ShermanBoyderm not how19:12
ShermanBoydjust "do I bind ..."19:12
alandmankd926: Are you using 32 or 64 bit python?19:13
alandmankd926: You seem to have both a /usr/lib/python2.6 and a /usr/lib64/python2.619:13
kd926alandman: python-2.6.6-20.el6.x86_6419:13
kd926alandman: You are correct, I was concerned with that as well19:14
kd926alandman: everything is from Grid Dynamics rpm repo and RHN19:14
*** Eyk^off is now known as Eyk19:16
uvirtbotNew bug: #788282 in swift "Update SAIO stats documentation for container-stats" [Undecided,New]
*** dendrobates is now known as dendro-afk19:18
kd926alandman: with the exception of qemu which had to be installed from source to get qemu*-x86_6419:18
*** primeministerp is now known as primeministerp|a19:19
*** primeministerp|a is now known as primeministerp|o19:19
*** primeministerp|o is now known as primeministerp19:19
primeministerpgah, character limit19:19
kd926ShermanBoyd: which interface is configured with your public ip?19:19
kd926ShermanBoyd: eth0?19:20
ShermanBoydkd926: eth019:20
kd926ShermanBoyd:  which distro?19:20
ShermanBoydthat's where I connect via ssh19:20
ShermanBoydubuntu natty19:20
kd926does which brctl show a binary?19:20
kd926sorry, $ which brctl19:20
*** fabiand__ has quit IRC19:21
ShermanBoyd /usr/sbin/brctl19:21
kd926ShermanBoyd: good, did you create the bridge? i.e. ifconfig br100 gives output?19:21
ShermanBoydoh wait you mean user env?19:21
kd926ShermanBoyd: you should be able to do this:19:22
ShermanBoydI threw it on eth1 though19:22
kd926$ brctl addif eth0 br10019:22
ShermanBoydwhich is a non-public face switch19:22
*** HouseAway is now known as AimanA19:22
*** MarkAtwood has joined #openstack19:22
kd926ShermanBoyd: however you will need to restart ssh and whatnot so you might lose your connection19:23
kbringardhas anyone ever seen this issue where your instances randomly lock up...19:23
kbringardso like, for instance, I'll look into an instance19:23
ShermanBoydso am I bonding it to an alias? or just replacing my eth0 config19:23
kbringardand like man curl19:23
kbringardor whatever19:23
kbringardand my shell will hang19:23
kbringardor try to edit a file, or something19:23
kbringardbut, I can log right back in19:23
kbringardthe the instance is still up and running19:24
kd926alandman: any ideas?19:24
kbringardit's intermittent, in that I can launch the same image again and it'll be fine, but once an instance is in that state, it happens consistently until it's terminated19:25
*** reldan has quit IRC19:25
kd926ShermanBoyd: that would replace your eth0 config, and eth0 would be a part of the bridge19:25
kbringardlook == log19:26
kd926ShermanBoyd: and by replace I really mean eth0 configuration is superceded by the bridges19:26
*** aliguori has quit IRC19:26
alandmankd926: Not really, unfortunately19:28
*** adiantum has quit IRC19:29
*** tblamer has joined #openstack19:29
*** rcc has quit IRC19:30
ShermanBoydkd926: that's working thanks19:31
kd926ShermanBoyd: no problem19:32
ShermanBoydam I correct if I say a flat network doesn't nat and the vlan model does19:33
Ryan_LaneShermanBoyd: yes19:33
ShermanBoydI mean nat the VMs19:33
ShermanBoydhmm that's a conundrum, I don't want to nat, but I need the compartmentalization19:34
Ryan_Lanedon't use floating IPs then :)19:34
Ryan_Laneyeah, it shouldn't NAT then19:35
Ryan_LaneI believe19:35
Ryan_Lanethough I'm more familiar with FlatDHCP19:35
*** dendro-afk is now known as dendrobates19:38
*** Ryan_Lane is now known as Ryan_Lane|food19:38
*** joearnold has quit IRC19:40
*** aliguori has joined #openstack19:40
uvirtbotNew bug: #788295 in nova "UnboundLocalError in exception handling" [Undecided,New]
*** littleidea has quit IRC19:43
*** freeflyi1g has joined #openstack19:43
*** freeflying has quit IRC19:46
ShermanBoydthe documentation states you don't need to configure br100 if you are using VlanManager , I'm starting to get an inkling that isn't correct ...19:51
kbringardShermanBoyd: if you're using vlan manager the compute and network code brings up the bridges and does the tagging19:52
kbringardbased on what is in the DB for each project19:52
ShermanBoydhow does it know which interface to bind to?19:52
kbringardit defaults to eth0 in the code, otheriwse it binds to whatever interface you have defined as --vlan_interface in the nova.conf19:53
kbringardso like, in mine19:53
kbringardI have19:53
kbringardthen you'll want to bring up eth1 with no IP19:54
kbringardat least on the compute nodes19:54
kbringardthis is how I have it set on all my machines (ubuntu specific)19:57
kbringardauto eth119:57
kbringardiface eth1 inet manual19:57
kbringardup ifconfig eth1 up19:57
vishyyou really shouldn't configure br100 ever19:58
ShermanBoydI'm going to update
ShermanBoyd * vlan_interface: This is the interface that VlanManager uses to bind bridges and vlan to.19:58
ShermanBoyd    (default: 'eth0')19:58
*** joearnold has joined #openstack19:58
ShermanBoyddoes that sound like a good description19:58
ShermanBoydvishy: you mean for VlanManager?  Or not even for FlatManager?19:59
vishymy basic principle is never use FlatManager20:00
*** jonkelly1 has joined #openstack20:00
vishyyou don't need to configure for Vlan or FlatDHCP20:00
kbringardwords to live by right there20:00
ShermanBoydso why is rackspace going flat?20:00
*** brd_from_italy has joined #openstack20:01
vishyShermanBoyd: they have external ip management services20:01
*** jonkelly has quit IRC20:01
uvirtbotNew bug: #788306 in swift "Aggregate Stats Need to Be Stored in Files Named Using the Date of Their Content" [Undecided,New]
vishyShermanBoyd: and there existing (enormous) deployment has assumptions about having a public address that is configured in the guest20:01
*** Razique has quit IRC20:02
vishyand they have the resources to do a bunch of manual external setup20:02
*** reldan has joined #openstack20:02
*** littleidea has joined #openstack20:03
ShermanBoydso I have two interfaces on each node, one connected to a switch right behind my router (pulbic) and one connected to another switch with no connection to the external workld20:04
*** zns has joined #openstack20:05
vishyShermanBoyd: that is generally a good setup20:05
ShermanBoyddo I put both the private subnets created for projects and the public ips on the "public" interfaces20:05
*** hallyn_afk is now known as hallyn20:07
ShermanBoydor can I run the private project networks on the "private" interface and route public ips from the "public" interface20:07
ShermanBoydI'm assuming that's what routing_source_ip is all about20:08
kbringardShermanBoyd: I think the public IPs all come up on the network controller and it NATs back to the private IPs of the VM20:08
ShermanBoydkbringard: that's with the floating ones?20:10
ShermanBoydthere are a few use cases where nat is a problem20:10
uvirtbotNew bug: #788314 in swift "For the Stats System, Split Raw Logs Into Date Based Files" [Undecided,New]
kbringardvishy: did you see my question earlier about VM's "locking up"20:11
ShermanBoydRyan_Lane mentioned I might just not use floating ips in these cases20:12
*** rnirmal has quit IRC20:13
vishykbringard: no20:13
*** nelson has quit IRC20:13
kbringardif you have a moment...20:13
alandmanCan anyone tell me what the --routing_source_ip setting in nova.conf does or is for?20:13
*** nelson has joined #openstack20:13
kbringardlong story short, every now and again a VM will just stop responding20:14
kbringardbut, it's still there20:14
kbringardit's just my shell seems to crap itself20:14
kbringardbut I can log into it again no problem20:14
kbringardand it's during random stuff20:14
kbringardlike editing a file20:14
kbringardor running man on something20:14
kbringardor apt-get/yum updates20:14
vishyalandman: when instances without a floating ip hit the internet, traffic is snatted to that ip20:14
*** rnirmal has joined #openstack20:14
kbringardwhen it happens to an instance, it happens to it until it's terminated, but I've fired up the same image on the same compute node and it's fine20:15
ShermanBoydis routing_source_ip only for VlanManager then?20:15
kbringardand it happens with multiple images of multiple flavors (ubuntu and cent specifically)20:15
uvirtbotNew bug: #788315 in swift "Make the Log Processor Runnable Across Multiple Systems at Once" [Undecided,New]
kbringardbut I can't seem to find any specific catalyst that causes it, it just seems to happen randomly20:17
kbringardand nothing in the log files, at least that I can tell20:17
vishyShermanBoyd: no, flatDHCP nats as well20:18
*** lorin1 has joined #openstack20:18
vishykbringard: have you tried VNC in to see if the instance has crashed or just lost networking?20:18
kbringardwell, I can ssh right back into it20:18
kbringardso the VM isn't locked up, per se20:18
kbringardit's just that particular session that's hung20:19
kbringardI can even log back into it in another window while the one is still hung20:19
alandmanvishy: Is it possible to set a gateway different from br100 using flatDHCP?20:19
kbringardbut, if I try to man or whatever again, it'll do the same thing20:19
vishyalandman: depends on what you mean by gateway?20:19
kbringardit's more or less a minefield and I've found that terminating and re-instantiating is pretty much the best choice20:19
vishykbringard: hmm, how large is the deployment?20:20
alandmanvishy: When nova-network starts it creates br100 with an IP address, seemingly the one set for the gateway used by dnsmasq20:21
kbringardcurrently it's... I think 10 compute nodes, with one scheduler, controller and network all running on the same host (which doesn't have compute on it)20:21
kbringardthere aren't that many VMs running though20:21
alandmanvishy: If I had a router on that network could I have dnsmasq start giving out that gateway20:21
kbringard26 VMs running currently20:22
vishykbringard: at a certain point you need to up ip_conntrack limits (defaults to 65535) or you will start losing connections20:22
kbringardalthough, I did just add a bunch of nodes, so for this scenario there are 420:22
vishybut i don't think you've hit that yet20:22
kbringardit seems to be IO related20:22
*** hadrian has joined #openstack20:23
kbringardlike, when I access something on disk20:23
kbringardon the VM20:23
*** MarkAtwood has left #openstack20:23
vishyalandman: not really.  We are working on another mode that would support that.  You can do it sort of, but you will lose floating ips and you will have to set up metadata forwrding rules manually20:23
kbringardbut, kind of network related as well.. because a telltale sign that it's going to happen is if you try to apt-get and the mirrors start timing out20:23
vishykbringard: what are your host oses?20:25
*** robbiew has quit IRC20:26
kbringardubuntu 10.1020:26
vishykbringard: i haven't seen anything like that, seems very strange20:26
kbringardyea, it's really really odd20:26
ShermanBoydcan anyone tell me cc_host does?  Docs say it points to the api "–cc_host ### This is where the the nova-api service lives" but don't we already have that with ec2_url ?20:26
kbringardno worries, I'll keep investigating20:26
*** robbiew has joined #openstack20:26
uvirtbotNew bug: #788318 in swift "Make the Log Processor More Memory Efficient" [Undecided,New]
kbringardI just figured I'd ask in case you (or anyone) had any ideas20:27
kbringardnext time it happens I'll try to get some straces of the kvm process on the host and collect some data if I can20:27
lorin1ShermanBoyd: I think cc_host isn't used anymore. I just searched for it in the code and couldn't find it.20:30
alandmanIs there any way to start instances that are shutdown or can they only be deleted and rerun?20:30
uvirtbotNew bug: #788320 in swift "Only Aggregate Data Files Should Appear in Their Swift Container" [Undecided,New]
kbringardthanks for your time vishy :-)20:31
*** jedi4ever has quit IRC20:34
ShermanBoydthanks everyone!20:34
*** Shentonfreude has quit IRC20:34
ShermanBoydannegentle: so it seems like cc_host is not part of cactus
ShermanBoydalandman: ahh there's a good one20:36
ShermanBoydalandman: maybe you can save it as a template then shutdown20:36
ShermanBoydalandman: then restore the template when you start?20:37
*** jobicoppola has joined #openstack20:39
ShermanBoydlorin1: thanks, if you want to chime in on the doc bug I filed it's here
*** alandman has quit IRC20:45
*** dmarkey has joined #openstack20:46
dmarkeySo, it looks like openstack is mostly python?20:50
kbringardalandman... oh, looks like he left20:50
kbringarddmarkey: yeppers20:50
dmarkeykbringard: likr.. is there any non python?20:51
vishyalandman: hmm was going to suggest euca-reboot-nstances20:51
kbringarduhm, I don't think so... it relies on external stuff that isn't python, but vishy would know best20:51
vishydmarkey: there are a few shell scripts and ti uses qute a few linux utilities20:52
dmarkeyi see, is there a database?20:52
ShermanBoyddmarkey: mysql or postgres20:52
vishydmarkey: yes, sqlalchemy, can use sqlite/mysql/postgres20:52
dmarkeyah. was hoping you were using django20:52
vishydmarkey: the dashboard uses django20:53
ShermanBoyddmarkey: I think they use sqlalchemy20:53
*** JuanPerez has quit IRC20:53
annegentleShermanBoyd: thanks for that legwork, updating now20:53
dmarkeydjango has its own ORM20:53
dmarkeyi suppose its no harm learning a new one20:53
vishydmarkey: we only use the orm at the bottom of the db layer20:53
dmarkeyi see20:54
vishyaside from that we use mostly dictionaries20:54
dmarkeyso.. you use the django url dispatcher, and templating language?20:54
vishydmarkey: only in the dashboard20:54
vishymost of openstack is not a website20:55
dmarkeyi see20:55
*** m0dpr0b3 has quit IRC20:55
dmarkeyand does everything communicate via xmlrpc or somesuch20:55
vishydmarkey: for compute mostly amqp20:55
vishyexternal apis are REST20:55
vishy(or amazon's style rpc's for ec2 compatibility layer)20:56
ShermanBoydI've got this virbr0 interface, is this just default libvirt stuff?20:56
*** mdomsch has joined #openstack20:57
ShermanBoydvishy: yes to me?20:57
dmarkeyvishy: thanks. I'll try to create an instance and have a play20:57
vishyShermanBoyd: yes20:58
ShermanBoydvishy: do I leave it be or should I get rid of it?21:00
*** allsystemsarego has quit IRC21:01
*** Ryan_Lane|food is now known as Ryan_Lane21:03
*** lorin1 has left #openstack21:05
*** h0cin has quit IRC21:07
*** Pathin has joined #openstack21:09
*** dendrobates is now known as dendro-afk21:10
*** amccabe has quit IRC21:10
*** Pathin has quit IRC21:11
*** Pathin has joined #openstack21:13
*** Pathin has quit IRC21:14
*** Pathin has joined #openstack21:15
*** Pathin has quit IRC21:17
*** Pathin has joined #openstack21:18
*** littleidea has quit IRC21:18
*** Pathin has quit IRC21:18
*** Pathin has joined #openstack21:18
*** jobicoppola has quit IRC21:20
*** jobicoppola has joined #openstack21:21
*** RobertLaptop has quit IRC21:21
*** aliguori has quit IRC21:22
*** abb17 has joined #openstack21:23
abb17where are the meeting minutes for yesterday's Quantum meeting?  I can find the regular minutes, but nothing on the NetStack meetings21:24
vishyShermanBoyd: just make sure that you kill the dnsmasq that libvirt starts21:27
vishyyou can disable the default network somehow as well but it is just the dnsmasq instance that causes issue21:27
ShermanBoydvishy: ty21:27
abb17vishy: thanks!21:27
*** Jordandev has joined #openstack21:30
*** RobertLaptop has joined #openstack21:33
*** j05h has quit IRC21:34
*** j05h has joined #openstack21:34
*** ctennis has quit IRC21:35
* ShermanBoyd inches closer to a working openstack21:35
*** aliguori has joined #openstack21:35
*** ameade has quit IRC21:37
*** nijaba_afk has joined #openstack21:37
*** nijaba has quit IRC21:38
*** mgoldmann has quit IRC21:38
ShermanBoydroot@titanium:~# euca-authorize -P icmp -t -1:-1 default21:39
ShermanBoyddefault None None icmp -1 -1
ShermanBoyd[Errno 111] Connection refused21:39
*** stewart has quit IRC21:41
kbringardhere's an interesting one21:42
kbringardgflags.IllegalFlagValue: flag --vpn_image_id: invalid literal for long() with base 10: 'ami-00000023'21:42
kbringardif I remove the ami-, it at least accepts the flag... but I think nova-manage is using it in base10, but the the API is doing it in base1621:42
kbringardbecause when I try to launch it, it tells me ami-00000017 not found21:43
*** jatsrt has joined #openstack21:44
ShermanBoydhmm most of my log files are empty21:44
ShermanBoydwith the exception of nova-manage.log21:45
*** vernhart has joined #openstack21:47
jatsrtanyone out there have experience with live migrations21:47
jatsrtor more specifically with getting nfs4 working for your instances directory21:48
jatsrtperpetual permission problems21:48
*** jobicoppola has quit IRC21:49
*** eikke has joined #openstack21:50
kbringardwhat's the status of the cloudpipe stuff?21:52
kbringardis it supposed to work?21:52
jatsrtwhat is not working for you?21:52
jatsrtand yes I beleive it is supposed to work21:52
kbringardjatsrt: well... I created the image21:52
kbringardgot it all setup like in the docs21:53
kbringardthe first thing that seems to be bugged is that the vpn_image_id only takes a base10 number (in the nova.conf)21:53
kbringardso1) you can't put ami-21:53
jatsrtthat was a tircky bit21:53
kbringardand 2) you have to convert the real image's hex number to dec21:53
jatsrtand it is long nor hex21:53
jatsrtand don't pad with 0s21:53
kbringardso, for 23 I put 3521:53
jatsrtso ami-0000000221:53
jatsrtwould just be 221:53
kbringardah, that's good to know21:54
jatsrtwould be 1021:54
kbringardyea... ami-000000023 i put 00000003521:54
kbringardand that seems to have worked (although I'll remove the 0s)21:54
jatsrtalso not that as far as I could tell br0 will not come up unless that is configured right21:54
jatsrtyeah take out the 0s21:54
jatsrtfor some reason my bridge interface came up after that21:54
kbringardso then, when I do nova-manage vpn run project21:55
kbringardit starts21:55
kbringardbut, it didn't give it the .2 that's outlined in the DB21:55
kbringardunder vpn_private_address21:55
kbringardit gave it the next available one in the normal dhcp pool21:55
kbringardINSTANCEi-000000dfami-0000002310.4.1.1010.4.1.10runningsandbox-vpn (sandbox, enc2b13)0m1.tiny21:55
kbringardinstead of
*** jonkelly1 has quit IRC21:56
jatsrthmm, I was getting the .2 ok21:56
kbringardnova-manage vpn list shows this:21:56
kbringard10.4.1.10 enc2b13 i-000000df running down21:56
jatsrtand your network is setup ok?21:57
kbringardI'm assuming the down is because it's expecting it to be on .221:57
jatsrtnova-manage network list21:57
kbringardyep, network works great otherwise21:57
jatsrtwhat is your subnet?21:57
jatsrtjust make sure you didn't make them too small21:57
kbringardfor that project,
*** lvaughn has quit IRC21:57
jatsrtopenstack gets real confused if you create a subnet of say 8 and run 10 instances21:57
kbringardnope, I have /24s for every project21:58
jatsrtnot too sure why that would happen then21:58
kbringardand I've fired up 100 instances at once, and they all work21:58
jatsrtare you running trunk or cactus?21:58
kbringardhaha, ok, cool21:58
jatsrtwell, it was working for me just yesterday21:58
kbringardwell, let me try removing the 0s21:58
kbringardmaybe somehow that's screwing it up21:59
jatsrtyeah that definately helped21:59
jatsrtI think it does21:59
jatsrtnot sure why though, I haven't dug into the code on that yet21:59
kbringardbut the fact that it launched the right image makes me wonder how that would be related to this21:59
jatsrtthat is the same behavior that I was seeing then removed 0s and things got happier.  Though I think I just never got any IP21:59
jatsrtit did launch though22:00
kbringardah, cool, let me try that22:00
kbringardthanks :-)22:00
kbringardugh, these "restarting it" instructions are super clear :-p22:01
*** jonkelly has joined #openstack22:01
ShermanBoydkbringard: which ones?22:01
kbringardthe "Restarting Cloudpipe VPN" section22:02
*** matiu has joined #openstack22:02
*** jonkelly has quit IRC22:02
*** mdomsch has quit IRC22:03
kbringardnewp, it's still getting an IP from the DHCP pool22:03
kbringardall I'm dong is nova-manage vpn run project22:04
*** zns has quit IRC22:04
*** foxtrotgulf has quit IRC22:04
kbringardvpn_public_address is the same as all the others22:04
kbringardpublic_port is unique (1000... default)22:05
jatsrtdid you start up a second project to try it out clean?22:05
kbringardnot yet, good idea though22:05
kbringardsame thing22:05
jatsrtI know for a fact that I was getting the .2 yesterday22:05
*** brd_from_italy has quit IRC22:05
kbringardbrings it up in the right subnet22:05
kbringardbut gives it an IP from the DHCP pool22:05
*** lborda has quit IRC22:06
kbringardit can't be something in the image itself, since it's nova that's dolling out the IP22:06
kbringard... thats not to say the image is right, just that it shouldn't be this problem22:06
jatsrtyou are using br0 and not eth0 correct?22:07
ShermanBoydugh those instructions are hard to follow22:07
kbringardno, each project gets it's own bridge and eth1 is my vlan_interface22:07
kbringardbut, it all works correctly in every other way22:07
*** Eyk is now known as Eyk^off22:08
kbringardso the project gets bridge br_A_B where A and B are the subnet classes22:08
kbringardso would be on br_4_322:08
*** ctennis has joined #openstack22:08
*** ctennis has joined #openstack22:08
kbringardbut like I said, that all works fine with instances22:09
kbringardthey come up right22:09
kbringardthey get tagged right22:09
*** med_out is now known as med22:09
jatsrtno sorry in the instance, it is expecting a br0 instead of eth0 for dhcp, not that it should make a difference22:09
*** med is now known as medberry22:09
kbringardoh wait22:10
*** mgius has quit IRC22:10
kbringardyou mean on the image itself?22:10
kbringardyea, that's how that's setup22:10
kbringardsorry, I thought you meant the project's network22:10
jatsrtok, so I just got in22:10
jatsrtI have it working on trunk from a few days ago22:10
jatsrtit is getting the right IP22:10
jatsrtI haven't tested connectiong to it, that was tomorrow project22:11
jatsrtwhen was the last time you updated22:11
kbringardthis morning22:11
jatsrtfrom apt22:11
kbringardi'm on 110822:11
jatsrtok, let me do an update and seee what happens22:11
kbringardto be specific22:11
jatsrtok, so I lied22:12
ShermanBoyddoes nova api open 2 ports?22:12
jatsrtit isn't getting the right IP22:13
kbringardah, OK22:13
kbringardthat's good to know22:13
kbringardhaha, in that case I'm going to call it a day and I'll dig into the code tomorrow22:13
ShermanBoyd2011-05-25 15:11:38,013 AUDIT nova [-] Starting /usr/bin/nova-api on
ShermanBoyd2011-05-25 15:11:38,014 AUDIT nova [-] Starting /usr/bin/nova-api on
kbringardShermanBoyd: yes22:13
kbringard8773 is the ec2 api22:13
kbringardand 8774 is the nova/rackspace api22:13
jatsrtyeah, has to be something screwed in the code22:13
jatsrtnot allocating the right IP22:13
jatsrtthe iptables rule is setup with the right IP, but DHCP is serving it wrong22:14
jatsrthas to be some dhcp server contention22:14
kbringardyea, which would indicate that it's getting written out to the br_....conf wrong22:14
kbringardthat's the right hostname and mac22:15
kbringardwrong IP22:15
* kbringard updates it manually and euca-reboots the instance22:15
jatsrtyep same thing22:16
*** gondoi has quit IRC22:17
kbringardwell, I need to call it a day, thanks for the help22:19
kbringardglad to know I'm not just crazy22:19
kbringarddefinitely crazy, but that's not the problem in this case ;-)22:19
kbringardI'll dig into the code some tomorrow22:19
kbringardpiece out homies22:19
*** kbringard has quit IRC22:19
ShermanBoydkbringard: peace to the middle east22:20
*** dendro-afk is now known as dendrobates22:21
*** kbringard has joined #openstack22:21
kbringardjatsrt: I just thought of something22:21
kbringard.2 is marked as reserved in the fixed_ips table22:22
kbringardperhaps that is part of the problem22:22
kbringardit would make sense that it would be reserved22:22
kbringardbut if for some reason it's bjorking because of that22:22
kbringardanyway, just a thought22:22
kbringardI'm gone for real now22:22
*** aliguori has quit IRC22:22
kbringardpiece out22:22
*** kbringard has quit IRC22:22
*** newopen has quit IRC22:25
*** nerens has quit IRC22:30
jatsrtif kwargs.get('vpn', None):             address = network_ref['vpn_private_address']             self.db.fixed_ip_associate(ctxt,                                        address,                                        instance_id)22:33
*** dendrobates is now known as dendro-afk22:34
*** littleidea has joined #openstack22:35
vishyjatsrt: what is broken?22:39
jatsrtso when running a cloudpipe instance22:39
jatsrtit is not allocating the correct IP22:40
jatsrtin VLAN mode22:40
vishyeven with cloudpipe image id set?22:40
jatsrtthat is the line in in network that looks if it is VPN and sets it up22:40
*** jkoelker has quit IRC22:40
vishyis the vpn kwarg not passed properly anymore?22:40
jatsrtdoesn't look like it22:41
jatsrtwhere does that get passed in?22:41
vishy237-247 of compute.manager22:41
vishylooks like it is still there22:42
jatsrtis_vpn = instance_ref['image_id'] == str(FLAGS.vpn_image_id)22:42
jatsrtso two of us confirmed you just get an IP from the main pool22:42
jatsrtI'm just tracing back through the code now22:43
jatsrtso the vpn_image_id is supposed to be a long22:43
jatsrtso for me =922:43
vishyperhaps the format of instance_ref['image_id'] has changed?22:43
vishyput a log in there and see what it is getting?22:43
jatsrtyeah, one sec I'll see if I can22:44
vishy(or just look in your database)22:44
jatsrtlets see that's compute, so that would need to be on the host it started on?22:44
jatsrtso sowing up in the DB as 922:45
jatsrtso that looks like it would be a match22:45
ShermanBoydhmm I don't seem to be able to get nova-api to run, even when I start it from the command-line I still can't hit the porta22:46
*** JoeJulian has joined #openstack22:47
vishyjatsrt: weird22:47
ShermanBoydfrom syslog all the other services are just constantly restarting22:49
jatsrtso in network/manage I put in a log statement and got {'vpn': False}22:49
jatsrtI'll dive into this tonight and perhaps have a patch, would be my first one to the project :-)22:49
*** miclorb has joined #openstack22:49
jatsrtok all, I'm out of here tonight22:51
*** jatsrt has quit IRC22:51
vishyjatstrt: yes i'm very curious22:55
vishyimage_id may have been changed to an int22:55
*** tblamer has quit IRC22:56
vishyhmm doesn't look like it22:56
*** littleidea has quit IRC22:56
*** somik has quit IRC23:04
*** ctennis has quit IRC23:08
*** littleidea has joined #openstack23:14
*** dragondm has quit IRC23:25
*** rnirmal has quit IRC23:32
*** johnpur has quit IRC23:37
*** abb17 has left #openstack23:42
*** jeffjapan has joined #openstack23:43
*** medberry is now known as med_out23:44

Generated by 2.14.0 by Marius Gedminas - find it at!