Friday, 2011-06-17

*** Ryan_Lane has joined #openstack00:03
*** MarkAtwood has quit IRC00:04
*** wilmoore has quit IRC00:07
*** ziyadb has quit IRC00:08
*** Ryan_Lane has quit IRC00:08
*** samba has quit IRC00:10
*** ton_katsu has joined #openstack00:10
*** Ryan_Lane has joined #openstack00:10
*** ton_katsu has quit IRC00:16
*** joearnold has quit IRC00:23
*** vernhart has quit IRC00:24
*** ton_katsu has joined #openstack00:29
*** Binbin has joined #openstack00:35
*** llang629 has joined #openstack00:39
dweimerIn the swift documentation it mentions that the cluster can use 'commodity' hard drives due to the replication and checksum features. By commodity drives is it referring to typical SATA server drives (ex. constellations or ultrastars) or is it reasonable to use desktop quality drives like deskstars or barracudas?00:40
*** bluetux has joined #openstack00:40
Ryan_Lanedweimer: I'd say it's a bad idea00:40
Ryan_Lanethere is an order of magnitude between enterprise sata and consumer sata for silent read and write errors00:41
Ryan_Laneand there's another order of magnitude between enterprise sata and scsi/sas00:41
*** dolphm has joined #openstack00:41
*** dolphm has quit IRC00:43
dweimerInteresting. So the increased number of read/write errors on a desktop drive is to large for checksum and quarantine features to keep up?00:43
*** heckj has quit IRC00:44
*** ohnoimdead has quit IRC00:45
*** jdurgin has quit IRC00:47
*** MarkAtwood has joined #openstack00:47
*** dolphm has joined #openstack00:50
*** dolphm has quit IRC00:53
*** huslage has joined #openstack00:53
*** winston-d has joined #openstack01:02
Ryan_Lanedweimer: I'm just saying it's a risk01:02
Ryan_Lanedweimer: it's possible that the multiple copies, checksums, and such are enough, but consumer sata drives are garbage, and it's likely better not to use them01:03
*** Ryan_Lane has quit IRC01:03
*** cloudgroups has joined #openstack01:04
dweimerFair point. The costs are so enticing though. Do you know if anyone has done an error rate study since the Google paper a few years ago?01:05
*** cafe has quit IRC01:05
*** dolphm has joined #openstack01:06
uvirtbotNew bug: #798486 in swift "XML 1.1 Compatibility" [Medium,In progress]
*** dolphm has quit IRC01:06
*** maplebed has quit IRC01:08
*** Binbin has quit IRC01:17
*** ton_katsu has joined #openstack01:29
*** ksteward2 has joined #openstack01:30
*** dolphm has joined #openstack01:32
*** huslage has quit IRC01:36
*** pguth66 has quit IRC01:39
*** MarkAtwood has quit IRC01:39
*** georgegoh has joined #openstack01:41
*** PW__ has joined #openstack01:43
*** watcher has quit IRC01:46
*** obino has quit IRC01:47
*** Binbin has joined #openstack01:49
*** dolphm has quit IRC01:52
*** pmullaney has quit IRC01:57
*** fysa has joined #openstack02:00
*** PW__ has quit IRC02:01
*** GeoDud has quit IRC02:27
*** zz2 has joined #openstack02:28
zz22011-06-17 10:16:23    DEBUG [routes.middleware] Match dict: {'action': u'meta', 'controller': <glance.server.Controller object at 0x1260250>, 'id': u'5'} 2011-06-17 10:16:23    DEBUG [eventlet.wsgi.server] - - [17/Jun/2011 10:16:23] "HEAD /images/5 HTTP/1.1" 200 912 0.015638 2011-06-17 10:16:25    DEBUG [routes.middleware] No route matched for GET /v1/images/6 2011-06-17 10:16:25    DEBUG [eventlet.wsgi.server] 192.1602:31
zz2this is my glance log,02:31
zz2i don't know how to solve it?02:31
*** vladimir3p has quit IRC02:32
zz2i'am sure , i hava installed the same version nova inthe nova control node and the nova compute node . and the version is version_info = {'branch_nick': u'workspace',  'build_date': '2011-04-15 02:57:55 +0000',  'clean': None,  'date': '2011-04-15 02:47:01 +0000',  'revision_id': 'tarmac-20110415024701-a9bdb77vaatk99lh',  'revno': 989}02:37
zz2nova control node can lanuch a  instance succ, but the compute node can't lanuch and give me that log message,02:39
*** Binbin has quit IRC02:45
*** odyi has quit IRC02:46
*** cbeck has quit IRC02:47
*** osier has joined #openstack02:50
*** obino has joined #openstack02:56
*** hadrian has quit IRC03:04
zz2great, :), i solved my problem。 the reason is i have installed a different version glance server in the compute node, beafore this i think the compute node don't need the glance and they don't have influence each other, it's my fault,'03:09
*** rchavik has joined #openstack03:14
*** rchavik has joined #openstack03:14
*** HugoKuo has joined #openstack03:22
*** miclorb_ has quit IRC03:24
*** Ryan_Lane has joined #openstack03:31
*** winston-d has quit IRC03:33
*** winston-d has joined #openstack03:43
*** Binbin has joined #openstack03:46
*** AimanA is now known as HouseAway03:52
*** Binbin has quit IRC03:56
*** medberry is now known as med_out04:04
*** wariola has joined #openstack04:08
*** Ryan_Lane has quit IRC04:10
*** deshantm_laptop has joined #openstack04:15
*** matiu has joined #openstack04:15
*** PW__ has joined #openstack04:17
*** matiu has quit IRC04:20
*** matiu has joined #openstack04:20
*** matiu has joined #openstack04:20
*** Binbin has joined #openstack04:21
*** deshantm_laptop has quit IRC04:24
*** stewart has quit IRC04:24
*** kashyap has joined #openstack04:31
*** koolhead17 has quit IRC04:32
*** stewart has joined #openstack04:37
*** nijaba has quit IRC04:48
*** PW__ has quit IRC04:50
*** nijaba has joined #openstack04:50
*** nijaba has joined #openstack04:50
*** cbeck has joined #openstack04:52
*** Ryan_Lane has joined #openstack04:55
*** PW__ has joined #openstack04:56
*** miclorb_ has joined #openstack05:03
*** Binbin has quit IRC05:03
*** scooterS has joined #openstack05:06
*** scooterS has quit IRC05:08
*** wariola has quit IRC05:13
*** Ryan_Lane has quit IRC05:13
*** Ryan_Lane has joined #openstack05:15
*** ivan has quit IRC05:23
*** wariola has joined #openstack05:26
*** PW__ has quit IRC05:27
*** koolhead17 has joined #openstack05:30
*** Binbin has joined #openstack05:31
*** ivan has joined #openstack05:32
*** ivan has joined #openstack05:34
*** cloudgroups has quit IRC05:35
*** cloudgroups has joined #openstack05:35
*** cloudgroups has left #openstack05:36
*** miclorb_ has quit IRC05:40
*** Binbin has quit IRC05:40
*** ksteward2 has quit IRC05:47
*** miclorb_ has joined #openstack05:54
*** Ryan_Lane has quit IRC06:12
*** jakedahn has joined #openstack06:14
*** jakedahn has quit IRC06:14
*** jakedahn has joined #openstack06:14
*** pygi has quit IRC06:16
*** matiu has quit IRC06:19
*** katkee has joined #openstack06:36
*** mcclurmc_home has joined #openstack06:38
*** koolhead17 has quit IRC06:39
*** reidrac has joined #openstack06:54
*** vernhart has joined #openstack07:01
*** sante has quit IRC07:04
*** georgegoh has quit IRC07:11
*** Binbin has joined #openstack07:18
*** Razique has joined #openstack07:24
*** po has joined #openstack07:31
*** Ryan_Lane has joined #openstack07:39
*** ovidwu has quit IRC07:44
*** irclog_ has quit IRC07:44
*** Ryan_Lane has quit IRC07:45
*** nacx has joined #openstack07:52
*** ovidwu has joined #openstack07:52
*** irclog_ has joined #openstack07:53
*** cbeck has quit IRC07:57
*** cbeck has joined #openstack07:58
*** adjohn has joined #openstack07:58
*** CloudChris has joined #openstack08:11
*** Katibe has quit IRC08:11
*** CloudChris has left #openstack08:12
*** cuzoka has joined #openstack08:13
*** Capashen has joined #openstack08:16
*** RichiH_ has joined #openstack08:16
*** gaveen has joined #openstack08:18
*** gaveen has joined #openstack08:18
*** RichiH has quit IRC08:21
*** RichiH_ is now known as RichiH08:21
*** soosfarm has quit IRC08:22
*** lool has quit IRC08:22
*** soosfarm has joined #openstack08:25
*** lool has joined #openstack08:25
*** martin has quit IRC08:26
*** miclorb_ has quit IRC08:27
*** soosfarm has quit IRC08:29
*** lool has quit IRC08:29
*** dobber has joined #openstack08:32
*** martin has joined #openstack08:32
*** Binbin has quit IRC08:33
*** soosfarm has joined #openstack08:34
*** lool has joined #openstack08:34
*** allsystemsarego has joined #openstack08:37
*** ziyadb has joined #openstack08:43
doudeHi all, I've got a problem with XenServer and Glance. I put an image in glance with VHD disk format and when I try to start an instance with this image, the Xen driver blocks on '_fetch_image_glance_vhd'. Nothing happen. The instance stay in 'building' state. I can see get request in Glance log :
doudeI didn't have this problem if I use a raw disk format image08:54
*** anonymous83722 has quit IRC08:58
*** anonymous83722 has joined #openstack09:00
*** katkee has quit IRC09:01
*** katkee has joined #openstack09:02
*** mgoldmann has joined #openstack09:07
xxtjaxxHi! How long between euca-run-instance and it actually running should there be before getting suspicious09:19
*** kennethkalmer has joined #openstack09:20
sorenxxtjaxx: Depends on the size of the image.09:20
*** cbeck has quit IRC09:22
*** cbeck has joined #openstack09:22
xxtjaxxari-lucid in this case09:23
HugoKuothe size ....09:23
xxtjaxx3,8M it says here09:24
sorenxxtjaxx: You can't run an ari. It's a ramdisk.09:24
sorenI guess we should refuse to try.09:24
sorenYou run AMI's, not ARI's.09:25
HugoKuoAMI is the point :>09:25
xxtjaxxhm trouble is they have ,-´ in them. Makes euca tools apparently cry09:26
sorenNot at all.09:27
xxtjaxxI do not joke09:28
soren*all* images have that.09:28
sorenThe EC2 API always calls them ami-XXXXXXX (if it gets to name them)09:28
*** kennethkalmer has quit IRC09:29
xxtjaxxI only had it working when I cd'd into that dir09:30
*** kennethkalmer has joined #openstack09:30
*** kennethk_ has joined #openstack09:36
*** kennethkalmer has quit IRC09:36
xxtjaxxah nevm. well schedules the proper way now i think09:43
xxtjaxxI'll see what will happen09:43
xxtjaxxits a 150M big Image now09:44
xxtjaxxsoren: so now it still doesnt boot for some reason and I used ari/ami/aki as intended by the designers09:53
*** paraneetharanc has joined #openstack10:08
*** ton_katsu has quit IRC10:12
*** ton_katsu has joined #openstack10:12
*** ton_katsu has quit IRC10:15
*** stewart has quit IRC10:23
*** stewart_ has joined #openstack10:23
*** ton_katsu has joined #openstack10:24
*** kennethk_ has quit IRC10:30
*** wariola has quit IRC10:34
*** wariola has joined #openstack10:36
*** wariola has quit IRC10:39
*** bluetux has quit IRC10:46
*** ton_katsu1 has joined #openstack10:53
*** ton_katsu has quit IRC10:53
*** ziyadb has quit IRC10:58
*** ziyadb has joined #openstack10:58
*** ziyadb has joined #openstack10:58
*** paraneetharanc has left #openstack10:59
*** Capashen has quit IRC11:11
*** katkee has quit IRC11:12
*** markvoelker has joined #openstack11:15
*** stewart_ is now known as stewart11:17
*** justin_skariah has joined #openstack11:20
*** adjohn has quit IRC11:21
*** ton_katsu has joined #openstack11:22
*** ton_katsu1 has quit IRC11:22
*** shang has quit IRC11:24
xxtjaxxHmm so now it sits with ari and ami and aki properly set in the euca-describe-instance and yet nothing happend11:25
xxtjaxxwhy? and what the hell?11:26
sorenCan I see?11:26
sorenThe output of euca-describe-instances, that is.11:26
*** ton_katsu has quit IRC11:30
*** ctennis has quit IRC11:30
xxtjaxxsoren: INSTANCE        i-00000002      ami-569b07c2                    scheduling      /root/id_rsa (some, None)   0       m1.small        2011-06-17T09:41:40Z    unknown zone11:34
xxtjaxxsoren: ^11:35
soren/root/id_rsa might be the culprit.11:37
sorenYou don't pass it a path to a key, you pass it an ID for a key that you create using euca-add-keypair.11:37
sorenEr.. BEcause?11:37
sorenThat's just how it works.11:37
sorenxxtjaxx: Anyways, do you have nova-scheduler running?11:39
*** ctennis has joined #openstack11:44
*** ctennis has joined #openstack11:44
*** osier has quit IRC11:47
*** PeteDaGuru has joined #openstack11:48
xxtjaxxsoren: that was an image of 150MB size. This box is idling11:51
*** bkkrw has joined #openstack11:53
*** sandywalsh_ has quit IRC11:55
sorenxxtjaxx: I'm not sure why you're telling me that. I told you one thing you needed to change and another thing you needed to check.12:01
xxtjaxxthe scheduler appears to  be running yes12:10
xxtjaxxsoren: ah wait12:11
xxtjaxx << appears to have a bug12:11
xxtjaxxnova      1961  1.6  1.2  91028 37960 ?        S    Jun16  24:55 /usr/bin/python /usr/bin/nova-scheduler --flagfile=/etc/nova/nova.conf12:13
*** aliguori has joined #openstack12:15
sorenxxtjaxx: Yeah, that doesn't look too good. How did you install this?12:15
xxtjaxxDebian repos.12:17
sorenYou're running this on sid?12:18
*** katkee has joined #openstack12:18
xxtjaxxpulled in experimental fpr some of the deps12:21
sorenI don't know how well to expect that to work, to be honest. I've not tested it myself. The list of dependencies is long, and I don't know the state of all of them in sid.12:22
*** nphase-work has joined #openstack12:22
sorenxxtjaxx: Which exact version of nova is this, though?12:22
*** nphase has quit IRC12:24
xxtjaxxVersion: 2011.2-112:25
*** dirkx has joined #openstack12:26
sorenThat should work nicely.12:26
*** ziyadb_ has joined #openstack12:31
*** kashyap has quit IRC12:34
*** ziyadb has quit IRC12:34
*** hadrian has joined #openstack12:35
xxtjaxxsoren: appears not to... Any hints for debugging? N12:44
*** odyi has joined #openstack12:45
*** odyi has joined #openstack12:45
*** adiantum has joined #openstack12:46
*** kashyap has joined #openstack12:51
kd926Has anyone looked into restricting privileges necessary to run nova?12:54
*** huslage has joined #openstack12:55
xxtjaxxsoren: How do I debug this?13:02
*** bcwaldon has joined #openstack13:04
sorenkd926: Like what?13:05
sorenxxtjaxx: I'm knee deep in half a million other things right now. If you can stick around for a bit longer, someone should show up you can help you get to the bottom of this.13:05
kd926Specifically, nova-network depends on sudo for things like dnsmasq, iptables, ect13:05
sorenxxtjaxx: (Most people are in the US)13:05
*** bonzay has joined #openstack13:05
sorenkd926: Right.13:06
sorenkd926: Those are privileged operations. How would you perform them without being privileged?13:07
xxtjaxxsoren: alright. I'm sorry go on Its just an eval anyway ;)13:07
katkeei created a new user and then a new project with that user as a projectmanager. this user can create instances but cannot allocate ips. i tried to give that user a netadmin role but it does not work either.13:08
sorenkatkee: What's the error?13:08
katkeehi soren, the error is 401 not authorized13:09
katkeeEC2ResponseError: 401 Unauthorized13:09
bpaluchare there any guides out there that go through launching instances that thier images are hosted on glance? I can't find anything, and glance images don't seem to have emi's.13:09
*** lborda has joined #openstack13:10
sorenkatkee: I'd probably look at the nova-api logs (and possibly the nova-network logs).13:12
sorenbpaluch: If you didn't register through the EC2 API, I believe you can just use the name as known by glance.13:13
sorenbpaluch: If you do register through the EC2 API, it gets and AMI (not EMI. Only Eucalyptus uses EMI's)13:14
bpaluchok thx, i'll look into that then13:14
bpaluchstill learning about the different api's13:14
kd926soren: dnsmasq isn't privileged, only binding to low ports is, as for iptables I believe access can be added to allow making adjustments without being root13:16
kd926soren: the intent is not have the entire service to be run as root. thats overkill13:16
katkeesoren: i didn't see any errors in the logs...13:18
*** bonzay has left #openstack13:19
katkeesoren: is it possble to have an admin user per project?13:20
katkeesoren: i don't want an admin user to see others projects13:20
*** huslage has quit IRC13:25
*** huslage has joined #openstack13:25
*** pygi has joined #openstack13:28
*** dolphm has joined #openstack13:28
*** gaveen has quit IRC13:28
bpaluchOh so if I have the Glance flags in my nova.conf, and then I use the euca commands to store an image it will use glance for storing them?13:29
bpaluchhmm maybe I spoke too soon, don't know the euca command for storing images...13:30
*** shentonfreude has joined #openstack13:33
*** mRy has joined #openstack13:34
*** mRy is now known as Guest5461913:35
*** hallyn has joined #openstack13:36
*** justin_skariah has quit IRC13:39
bpaluchoh I get it now. I was using the ubuntu cloud util tools earlier but they are just wrappers for simplifying the euca tools.13:41
*** jkoelker has joined #openstack13:42
*** foxtrotgulf has joined #openstack13:48
*** amccabe has joined #openstack13:53
*** foxtrotdelta has joined #openstack13:56
*** foxtrotgulf has quit IRC13:58
*** kbringard has joined #openstack14:04
kbringardanyone have any idea what may be causing this:
*** dgags has joined #openstack14:10
uvirtbotNew bug: #798714 in nova "Misleading error message on migrate number conflict when running unit tests" [Undecided,New]
*** troytoman-away is now known as troytoman14:13
*** lorin1 has joined #openstack14:13
*** primeministerp1 has joined #openstack14:13
*** huslage has quit IRC14:14
*** huslage has joined #openstack14:14
*** osier has joined #openstack14:15
bpaluchDo I have to run the euca authorize commands for allowing tcp on the compute node or the controller node?14:18
bpaluchand will I then have to run authorize commands for every port I want to communicate with on instances on that compute node?14:18
bpaluchproblem i'm having in a multi node set up, is that I can ping the instance i've launched by can't ssh, port refuses connection.14:19
*** lborda has quit IRC14:19
*** katkee has quit IRC14:22
*** dk_ has joined #openstack14:24
*** huslage has quit IRC14:25
*** dragondm has joined #openstack14:26
NashTrashjaypipes: Hi Jay, I have moved my Glance server and now the nova-api machine is throwing an error saying unable to connect.  Might you have a moment?14:27
jaypipesNashTrash: I can try. Could you pastebin your logs from nova-api?14:28
kbringardNashTrash: you may want to check that you have the glance_api_servers flag set in your nova.conf14:28
*** lborda has joined #openstack14:28
kbringardthat messed me up for awhile14:28
NashTrashkbringard: Have you run nova-manage role for that server?14:28
kbringarduhm... how do you mean?14:29
kbringardoh, you mean the bug I reported14:29
kbringardyea, other people confirmed it as well14:29
kbringardif you're not admin, you cannot launch instances14:29
NashTrashjaypipes: Here is the error coming from nova-api:
NashTrashkbringard: you have to give users the sysadmin and netadmin roles at both the global level and at the project level14:30
kbringardI think the flags used to be glance_host and glance_port but now it's one flag: glance_api_servers = host:port14:30
kbringardNashTrash: I know14:30
kbringardit was working fine, then I upgraded the code, and then it stopped working14:30
NashTrashkbringard: Ah, ok.14:31
kbringardplus, the error explicitly says AdminRequired: User does not have admin privileges14:31
kbringardand the code is wrapping AdminRequired()14:31
kbringardmy best guess so far is that something in the new scheduler isn't setting the context right14:31
NashTrashjaypipes: Here is my nova.conf —
jaypipesNashTrash: what version of Nova?14:32
NashTrashjaypipes: My glance had been on but I moved it to the same server as the nova-api (
NashTrashjaypipes: Diablo milestone 114:32
kbringardI'd try adding ----glance_api_servers=
kbringardonly 2 --s14:33
jaypipeskbringard: I don't think that was in D1, no?14:33
kbringardoh, milestone 114:33
kbringardmy bad14:33
kbringardlo siento14:33
kbringardit's early and my kid screamed all night14:33
kbringardmuy consado14:34
NashTrashkbringard: I hear you on that one.14:34
kbringardI think he's teething14:34
kbringardbut anyway, yea14:34
jaypipesNashTrash: hmm, ok, so can you ping the Glance API server? just use the glance CLI tool:14:34
NashTrashjaypipes: Netstat -lp shows glance-api listening on 929214:35
jaypipesglance --port=XXX --host= index14:35
jaypipesglance --port=9292 --host= index14:35
kbringardjaypipes: so long as you're here... I know it's not your area of expertise, but does a quick look at the stack trace in show any glaring issues I'm missing?14:35
kbringardnot being able to launch instances if you're not an admin seems like a pretty serious bug to me :-D14:36
NashTrashjaypipes: glance index returns good results14:36
jaypipeskbringard: yeah, that makes sense. It looks like the API controller needs to do a context.elevated() when it calls into the zone db...14:36
jaypipesNashTrash: ok, good.14:36
*** katkee has joined #openstack14:37
*** robix has quit IRC14:38
jaypipesNashTrash: ok, so I noticed that my_ip and glance_host are different...14:38
jaypipesNashTrash: one is 243 the other is 253.14:38
jaypipesNashTrash: is that the issue?14:38
RaziqueHi folks, is it possible to add a tag to as instance ? so euca-describe instance returns the tag ?14:39
NashTrashjaypipes: Uh, yeah.  MISTYPE!14:39
jaypipesNashTrash: :)14:39
Raziquecuz yesterday I terminate the wrong instance, kinda annoying14:39
NashTrashd'oh d'oh d'oh d'oh d'oh d'oh d'oh d'oh14:39
jaypipesNashTrash: I channel Homer at times as well ;)14:39
*** jonkelly has joined #openstack14:40
kbringardjay, nice14:44
kbringardI added context = context.elevated() to line 530 in the get_all_across_zones and it's working now14:45
kbringardthanks homie14:45
NashTrashjaypipes: Thanks!  That worked like a charm.  I am back the land of runnable instances14:46
*** imsplitbit has joined #openstack14:46
*** llang629 has left #openstack14:46
kbringardI monkey patched my API server and I am back in business as well14:48
kbringardnot even drinkin' time and jay has already solved 2 provlems14:48
NashTrashkbringard: Submit that patch buddy!14:48
kbringardhehe, I want to look at the context a bit more and make sure I did it in a sane way, but that is my intention14:48
*** po has quit IRC14:49
jaypipesNashTrash: excellent! :)14:50
NashTrashkbringard: I am kind of new to python, so this may be a really dumb question.  But, from the python shell if I try to import nova.api.ec2 it throws an error about _() not being defined.  Do you happen to know how to get around this?14:50
jaypipeskbringard: that worked ok, then? good :)14:50
kbringardNashTrash: I am also new to python, so I have no kind of idea14:50
kbringardjaypipes: yessir, you are the man14:51
jaypipesNashTrash: that's an easy one... do this:14:51
NashTrashkbringard: Ha.  Again, I am not alone14:51
jaypipesNashTrash: import gettext14:51
NashTrashjaypipes: What would we do without you14:51
NashTrashjaypipes: ?????14:51
reidracNashTrash: you would use google :)14:51
jaypipesNashTrash: gettext.install('nova', unicode=1)14:51
jaypipesNashTrash: do that ABOVE your import.14:52
NashTrashreidrac: So much sarcasm.  Way too early for that.14:52
reidracNashTrash: 10 minutes left to finish my week, excuse me :)14:53
NashTrashreidrac: Europe?14:53
reidracNashTrash: UK, yes14:53
NashTrashreidrac: Ah, then have a great weekend.14:54
reidracNashTrash: Thanks, you too14:54
NashTrashjaypipes: Is glance-upload documented anywhere?14:56
jaypipesNashTrash: not really. only in the --help.14:57
jaypipesNashTrash: are you using EC2 API?14:57
jaypipesNashTrash: technically, you can do everything that glance-upload does with the regular glance client... it's more of a shortcut for EC2-style images.14:58
NashTrashjaypipes: glance-upload —help goes to a trace.  I have taken a snapshot of a running machine and I want to try creating instances using that snapshot.14:58
kbringardNashTrash: if you like ruby, I wrote a rubygem to interface with the glance API and a client for uploading using it directly14:59
jaypipesNashTrash: heh, that's not good!14:59
jaypipesNashTrash: could you submit a bug on that?14:59
*** mRy_ has joined #openstack14:59
NashTrashkbringard: I am using the bash scripts you posted yesterday as a guide14:59
*** dirkx has quit IRC14:59
NashTrashjaypipes: Will do.14:59
jaypipesNashTrash: cheers14:59
sorenNashTrash: I can't imagine "—help" is a valid option to glance-upload.15:00
sorenNashTrash: "-help" *could* be, but I don't even think it is, either.15:00
kbringardfor very basic, you should be able to do glance-upload --disk-format=ami --container-format=ami --type=machine <image> <image-name>15:00
kbringardfor the image15:00
sorenNashTrash: Hint: "—" is not the same as "-"15:00
kbringardif it's an all in one (meaning the kernel and ramdisk are included in the image) then that should be all you have to do15:01
kbringardif it's not, you'll need to do some glance update magic15:01
NashTrashsoren: One dash not two.  You are correct sir.15:01
sorenNashTrash: — is not two dashes. It's a completely different character.15:01
sorenNashTrash: How did you even type it?15:02
*** Guest54619 has quit IRC15:02
kbringardsoren: is it possibly your (or his) IRC client changing it?15:02
NashTrashsoren: Very true.  The standard two dash causes a trace.  One dash give the following "glance-upload: error: argument -h/--help: ignored explicit argument 'lp'"15:02
kbringardI know friggin Word does that15:02
NashTrashIt was two dashes.  Adium "helped" me out.15:03
kbringardnot to compare your IRC client to word...15:03
NashTrashkbringard: My image has an external kernel but no ramdisk15:04
*** reidrac has left #openstack15:07
kbringardNashTrash: then you'll need to do like15:07
*** troytoman is now known as troytoman-away15:07
kbringardglance update <image id> type=machine kernel_id=<existing kernel id>15:08
kbringardalso, make sure that if you set any custom properties, you include them again in your glance update15:08
kbringardotherwise it assumes you don't want them anymore and marks them as deleted = 1 in the DB15:08
jaypipesNashTrash: cheers15:08
NashTrashkbringard: I did this — glance-upload --disk-format=ami --container-format=ami --type=machine --kernel=1 mongo_snap mongo_snap15:09
kbringardah, there you go, I think I forgot --kernel worked, haha15:10
NashTrashkbringard: It returned the following:
kbringardok, that looks good15:10
kbringardif you do a euca-describe-images15:10
kbringarddo you see it?15:11
kbringardand in theory it should have the kernel linked to it15:11
NashTrashkbringard: Yes, it is there.  I am trying a run-instances with it now....15:11
uvirtbotNew bug: #798749 in glance "glance-upload --help fails" [Low,New]
*** ziyadb_ has quit IRC15:12
*** ziyadb has joined #openstack15:13
*** ziyadb has joined #openstack15:13
NashTrashkbringard: Boo yaa.  It worked nicely.  Thanks.15:13
kbringardno prob15:14
*** mdomsch has joined #openstack15:16
kbringardand I can definitely confirm your glance-upload --help bug15:17
*** mdomsch is now known as mdomsch_westford15:17
*** Razique has quit IRC15:24
*** ameade has joined #openstack15:28
*** troytoman-away is now known as troytoman15:31
*** troytoman is now known as troytoman-away15:31
doudeHi, I meet a problem with Glance. I use the revision 139, it didn't change the las two weeks. Actualy, when I upload an Image, I didn't get any error but no data was upload in data store. The glance registery add an entry with an empty size.15:33
*** knighthacker has joined #openstack15:33
*** knighthacker has left #openstack15:34
doudeAnd this entry is in active status15:34
*** odyi has quit IRC15:34
doudei didn't change my config and worked a the begining of the week15:34
*** odyi has joined #openstack15:35
*** odyi has joined #openstack15:35
*** odyi has quit IRC15:35
*** obino has quit IRC15:36
doudeI use filesytem backend and file is created during the uplaod but this is an empty file15:36
*** odyi has joined #openstack15:38
*** odyi has joined #openstack15:38
kbringardhow are you uploading the file?15:38
*** anonymous83722 has quit IRC15:41
doudekbringard: i tried to use 'glance add' command and 'glance-uplaod' command. Same effect15:41
kbringardhmmm, when the command finishes, do you get any output?15:42
doudekbringard: yes,
*** heckj has joined #openstack15:44
*** rnirmal has joined #openstack15:44
doudekbringard: just for the command 'glance-upload'15:44
kbringardok, so when glance calculates the size, it's 0 as well... so it seems like the image isn't actually being uploaded15:44
kbringardu'properties': {u'type': u'raw'},15:45
kbringard u'size': 0,15:45
*** nphase-work has quit IRC15:45
kbringardI assume the file actually exists and is > 0 bytes and you have read permissions on it ?15:45
*** nphase has joined #openstack15:45
doudekbringard: And I can see log in glance API without errors :
kbringardyou can also md5 the image on your disk and see if it matches the md5 glance generated15:46
kbringardI'm guessing it won't though, since I'd wager glance is doing an md5 on a 0 byte file15:47
kbringardlastly I'd ngrep (or tcpdump, or something) the upload and see if it's actually sending data15:47
doudekbringard: not good, md5sum file: ec6f6893276a391442805cae1283d49615:50
doudekbringard: I'll try tcpdump15:51
kbringardyea... my guess is the file isn't actually being uploaded 15:51
kbringardthe post is happening, but the body is empty15:51
kbringardat least that's my guess15:51
*** ziyadb has quit IRC15:53
*** ziyadb_ has joined #openstack15:53
*** maplebed has joined #openstack15:54
*** dobber has quit IRC15:56
doudekbringard: i can see data in traffic. The data are correctly uploaded15:57
kbringardhmmm, that is odd15:57
kbringardI dunno... I guess I'd check permissions on your filesystem15:58
NashTrashdoude: If you go into the glance filesystem backend (/var/lib/glance/images) do you see anything there?15:59
kbringardif the data is being transmitted, but it's not being written to your glance location on disk15:59
*** med_out is now known as med15:59
*** med is now known as medberry15:59
doudekbringard: permission for the filesystem backend ?15:59
kbringardthen it seems there is some system level thing going on preventing it from writing15:59
*** mRy__ has joined #openstack15:59
doudekbringard: I set the user glance read and write to the folder15:59
doudekbringard: that's enough ?16:00
kbringardand the api/registry daemons are being run by the same user?16:00
NashTrashI have glance:root as the folder ownership16:00
doudeNashTrash: yes I can see empty files:
kbringardthat looks like the same perms I have on mine16:01
HugoKuohello guys , how's today XD16:01
kbringardstrange, I dunno what to say doude16:01
doudekbringard: yes:
NashTrashdoude: Hmm…kbringard has covered all of the bases I would have hit.16:02
*** dolphm has quit IRC16:02
*** mRy_ has quit IRC16:03
NashTrashdoude: Have you posted your glance-api.conf and glance-registry.conf?16:03
NashTrashdoude: Also, I am assuming df -f on the disk shows space available?16:04
NashTrashdoude: Sorry, just df16:04
doudeNashTrash: 3Go space left16:05
NashTrashdoude: Well, I guess that would do.  Jaypipes might be the person to help.  (Sorry Jay)16:05
doudeGlance API conf :
*** mRy__ has quit IRC16:07
doudeGlance registry conf :
*** cuzoka has quit IRC16:07
doudeStrange, that works before with the same release and config16:08
NashTrashdoude: Yeah.  Things look OK.  Sorry I can't help more.16:08
kbringardI see from your ps that it looks like glance may have been restarted recently?16:09
kbringardin your pasdte16:09
kbringardperhaps a change was made to the config awhile ago but only just now took effect due to a restart?16:09
doudekbringard: yes I tried recently to restart it16:10
doudekbringard: erase my database and created a new one16:10
kbringardheh, I'm kinda stabbing in the dark now... this is a pretty strange issue16:10
*** FallenPegasus has joined #openstack16:10
doudeI made some test with glance python client but I don't think is a problem16:11
doudeThe was just a feching image script16:11
jaypipesdoude: looking now..16:12
doudejaypipes: thanks16:13
*** joearnold has joined #openstack16:13
kbringarddoude: you can try to see if it helps any16:13
jaypipesdoude: what version of glance are you using?16:13
kbringardor, if you have a ruby env, you can use
*** FallenPegasus has quit IRC16:14
doudeI'm pretty sure, it's a stupid error but which one ?16:14
*** bkkrw has quit IRC16:15
*** FallenPegasus has joined #openstack16:15
*** FallenPegasus is now known as MarkAtwood16:15
jaypipesdoude: can you paste the script you are using, please?16:15
*** et_ has quit IRC16:15
*** MarkusWPHRorg has quit IRC16:15
*** dk_ has quit IRC16:17
doudejaypipes: first time I copy/paste the example script from doc with corect id and url :
doudejaypipes: and i improve it to work with the last librairie :
jaypipesdoude: what does ll /tmp/glance_file output?16:19
jaypipesdoude: in that script, you are writing the file (/tmp/glance_file) back onto itself...16:20
jaypipesdoude: sorry, no...16:21
jaypipesdoude: what does glance show 39 output?16:21
doudejaypipes: this image doesn't exist anymore. I reset all Glance : processes + bata base16:22
doudejaypipes: data base*16:22
jaypipesdoude: ok, so do you want to start over? Could you tell me what the problem is? I was a little late into the conversation here...16:23
*** MarkusWPHRorg has joined #openstack16:24
MarkusWPHRorgGood Morning fellow Openers16:24
doudejaypipes: When I upload a file to glance, It's ok. No error in Glance logs. The file exists in the registery and on the file system but it's empty, size zero.16:25
doudejaypipes: I snif the traffic and the file is correctly uploaded to glance16:25
jaypipesdoude: ok. and what are you using to upload to glance?16:26
jaypipesMarkusWPHRorg: Stackers :)16:26
*** ewindisch has quit IRC16:28
doudejaypipes: I use 'glance add' or 'glance-upload' commands16:28
MarkusWPHRorgjaypipes: I thought of that yesterday, but it seems to be lacking something.....16:29
*** obino has joined #openstack16:29
MarkusWPHRorgHow about "Cloud Wranglers?"16:29
jaypipesdoude: ok. can you show me the link to the paste of those exact commands, please? thx.16:29
doudejaypipes: glance add name="test" is_public=True < file16:30
jaypipesdoude: ok.16:30
jaypipesdoude: could you please tell me what version of both glance and webob you are using?16:31
*** dgags has quit IRC16:33
*** obino has quit IRC16:33
*** ohnoimdead has joined #openstack16:34
*** kashyap has quit IRC16:34
jaypipesdoude: and webob?16:35
*** osier has quit IRC16:36
doudejaypipes: Oops, sorry : 1.0.7-0~ppalucid116:37
jaypipesdoude: k, thx.16:38
*** koolhead17 has joined #openstack16:38
jaypipesdoude: ok, you up for doing a little patching on the fly to see if my hunch about the webob version is correct?16:39
MarkusWPHRorgI have a question about "euca-allocate-address"16:39
doudejaypipes: ok16:39
MarkusWPHRorgI'm trying to connect a publi ip on one interface to the ip of an instance16:40
MarkusWPHRorgplain vanilla ubuntu instance16:40
jaypipesdoude: alrighty. open up /usr/share/pyshared/glance/glance/client.py16:40
MarkusWPHRorgso I run "nova-manage floating create my-hostname <ip address>"16:40
MarkusWPHRorgand then euca-allocate-address <ip address>16:41
MarkusWPHRorgand finally euca-associate-address -i <instance name> <ip address>16:41
MarkusWPHRorgat each point, <ip address> is the public ip16:41
MarkusWPHRorgBut I get nothing but errors16:42
*** obino has joined #openstack16:42
jaypipesdoude: you there?16:42
MarkusWPHRorgMy question to all the smart learned people here is: Does that ip have to already be up on an interface somewhere?16:42
kbringardlol, gotta love it when people ask for help and then disappear ;-)16:42
MarkusWPHRorgOr does nova-network set up the interface for you?16:43
jaypipeskbringard: he's still here I think :)16:43
kbringardI know, just messin'16:43
doudejaypipes: yes16:43
jaypipesdoude: ok, search for the phrase "Chunk it, baby..." :)16:43
doudejaypipes: to much 'glance' in your path ?16:43
doudejaypipes: I get it line 13416:44
jaypipesdoude: ah, yes, sorry...16:44
jaypipesdoude: right after that comment, put this code:
*** flk_ has joined #openstack16:44
* jaypipes loves on-the-fly monkey patching...16:45
bpaluchis the vncproxy only in 2011.3 build packages? I installed using the script on the docs page, do I need to install something newer to get the vncproxy?16:45
jaypipesdoude: when done, save the file and try to add the image again...16:45
doudejaypipes: done16:45
jaypipesdoude: if my hunch is correct, that will fix the issue...16:45
doudejaypipes: no it doesn't work:
*** kashyap has joined #openstack16:54
MarkusWPHRorgSo how does openstack handle public ip addresses - does it create an aliased interface for them or soemthing?16:54
jaypipesdoude: is there an error message in the api log?16:55
MarkusWPHRorgI noticed that before I specified a public interface in my config, it would assign the ip to eth0 which didn't already have an ip16:55
doudejaypipes: no log in api and registry16:55
kbringardprobably need to turn debug = True in the conf files16:55
jaypipesdoude: did you restart the servers?16:55
*** nacx has quit IRC16:56
doudejaypipes: yes16:56
*** jdurgin has joined #openstack16:57
doudejaypipes: but the patch code concern only the client ?16:57
jaypipesdoude: the servers use the client.16:57
doudejaypipes: ok16:57
jaypipesdoude: as does the glance CLI tool.16:57
jaypipesdoude: so, no error was output at all? that seems very strange...16:57
*** koolhead17 has quit IRC16:58
*** magglass1 has quit IRC16:58
doudejaypipes: I cannot see any errors16:58
*** koolhead17 has joined #openstack16:58
doudejaypipes: sorry but I must leave. I'm in France and it's 7pm here.16:59
doudejaypipes: perhaps we can continue next monday ?17:00
jaypipesdoude: could you try: glance --debug add name="test" is_public=True < natty-server-uec-amd64.tar.gz17:00
jaypipesdoude: yes, we can continue on monday, no problem.17:00
*** adiantum_ has joined #openstack17:00
jaypipesdoude: ou, pas de probleme :)17:00
doudejaypipes: no option --debug with glance command17:01
jaypipesdoude: sorry, --verbose17:01
*** iammartian has joined #openstack17:01
doudejaypipes: just completed time added : Completed in 0.0002 sec.17:02
doudejaypipes: thanks for your help and have a good weekend17:02
jaypipesdoude: ok, you too.17:03
*** mszilagyi has joined #openstack17:06
*** obino has quit IRC17:06
*** kbringard has quit IRC17:07
*** foxtrotdelta has quit IRC17:07
*** adiantum has quit IRC17:07
*** hadrian has quit IRC17:07
*** soosfarm has quit IRC17:07
*** lool has quit IRC17:07
*** obino has joined #openstack17:07
*** joearnold has quit IRC17:08
*** kbringard has joined #openstack17:08
*** foxtrotdelta has joined #openstack17:08
*** hadrian has joined #openstack17:08
*** soosfarm has joined #openstack17:08
*** lool has joined #openstack17:08
*** joearnold has joined #openstack17:08
*** netmarkjp has joined #openstack17:09
kbringardwb everyone17:10
kbringardit was a long an arduous journey, but I'm glad we could all make it here again safely17:11
*** kbringard has quit IRC17:13
*** foxtrotdelta has quit IRC17:13
*** hadrian has quit IRC17:13
*** soosfarm has quit IRC17:13
*** lool has quit IRC17:13
*** ke4qqq_ is now known as ke4qqq17:14
*** crescendo has quit IRC17:15
*** kbringard has joined #openstack17:16
*** foxtrotdelta has joined #openstack17:16
*** hadrian has joined #openstack17:16
*** soosfarm has joined #openstack17:16
*** lool has joined #openstack17:16
*** obino1 has joined #openstack17:16
*** crescendo has joined #openstack17:16
*** ke4qqq has quit IRC17:16
*** ke4qqq has joined #openstack17:16
*** netmarkjp has left #openstack17:17
*** obino has quit IRC17:19
*** katkee has quit IRC17:20
*** joearnold has quit IRC17:24
*** adiantum_ has quit IRC17:31
*** kbringard has quit IRC17:31
*** kbringard has joined #openstack17:31
*** kbringard_ has joined #openstack17:32
*** kbringard has quit IRC17:32
*** kbringard_ is now known as kbringard17:32
*** zz2 has quit IRC17:38
MarkusWPHRorgNetsplit? Or is everyone just bailing?17:39
*** GeoDud has joined #openstack17:40
nhmlooks like there was a netsplit17:41
*** dolphm has joined #openstack17:43
*** vladimir3p has joined #openstack17:43
*** jakedahn has quit IRC17:45
*** Ryan_Lane has joined #openstack17:47
kbringardyep, netsplit is what I saw17:47
*** iRTermite has quit IRC17:49
*** arun_ has quit IRC17:50
MarkusWPHRorgCan anyone tell me how nova-net/euca sets up floating ip addresses?17:50
MarkusWPHRorgDoes it set up an aliased interface or something when you do the euca-associate?17:51
*** rchavik has quit IRC17:57
*** magglass1 has joined #openstack17:57
MarkusWPHRorgI notice that when I run "ip add list" one of my public ip's is listed there, associated/attached to my public interface17:58
MarkusWPHRorgas opposed to another publi ip which has it's own aliased interface(eth1:0)17:59
MarkusWPHRorgthe one on eth1:0 is pingable/ssh-able for outside, the one that shows ip in the ip ad list command is not17:59
MarkusWPHRorgMy problem is, I don't even know what the EXPECTED behavior is. Makes it a little hard to diagnose and troubleshoot18:00
kbringardthe floaters get added with ip addr add18:01
kbringardto the external interface on your network controller18:01
kbringardthen NAT rules get setup on the network controller as well18:01
kbringardiptables -L -vn -t nat18:01
kbringardshould, in theory, have D and SNAT rules sending packets to the private IP from the public18:02
kbringardif all that is setup, then you'll also want to make sure ip_forwarding is enabled on the network-controller18:02
kbringardif you're on ubuntu18:02
kbringardthen look in /etc/sysctl.conf for the following line18:03
MarkusWPHRorgok cool - thank you very much - that's very helpful18:03
kbringardit's commented out by default18:03
kbringardyou'll also want to find net.ipv6.conf.all.forwarding=118:03
kbringardif you're using ip618:03
kbringarduncomment them, save the file18:03
kbringardand run sysctl -p18:03
kbringardyou can set them manually, but changing it here assures they'll persist across reboots18:03
MarkusWPHRorgno ip6 just yet18:04
MarkusWPHRorgbut soon18:04
*** MarkAtwood has left #openstack18:05
kbringardanyone know how often merges propagate out to the ppa apt repo?18:05
MarkusWPHRorgI f I knew what a "ppa apt repo" was.....18:06
MarkusWPHRorgactually I know what an apt repo is18:06
MarkusWPHRorgjust not sure about ppa18:06
kbringardthe ppa apt repo is the launchpad repo where the openstack packages are18:07
MarkusWPHRorggot it18:07
kbringardat least in the context I'm using18:07
kbringardso I was asking how long after a fix has been commited, it takes for new dpkgs to be rolled and available at the repo18:09
*** tjikkun has quit IRC18:10
*** sebastianstadil has joined #openstack18:12
NashTrashCan I directly edit the "image_location" string in a Glance DB?  Is this used for anything other than user output?18:13
kbringardnot sure... I would have assumed it would tell glance where to get the image on the disk... maybe give it a try and see what happens, hah18:15
NashTrashkbringard: I am testing.  Glance stores each file in its filesystem as 1, 2, 3, 4, 5 (id of the entry).  The image_location seems more to have been set as part of the uec-publish-tarball process.  For example, using glance-upload doesn't even set that property.18:16
kbringardah, then in that case you're probably right18:17
NashTrashkbringard: Seems to still be working.  For whatever reason image_location property is what shows us as the primary descriptor for euca-describe-images.  I want to make that prettier for my users18:18
kbringardoh, interesting... I thought it was hexing the id18:19
NashTrashkbringard: Not from what I can tell.  There is a name field in the images table and the full name displayed in euca- is (image_properties.image_location)18:20
kbringardoooo, sorry, I thought you meant the AMI18:21
NashTrashkbringard: Oops.  Got that backwards.  euca- shows image_properties.image_location(
kbringardyea, I set the name when I upload18:22
*** iammartian has left #openstack18:22
NashTrashAFK for a bit18:22
uvirtbotNew bug: #798876 in nova "Consider Switching from Carrot to Kombu for AMQP" [Undecided,New]
zulsoren: bug 798878 is kind of interesting, we dont want to leak the mysql password ;)18:36
*** ahs3 has joined #openstack18:37
*** fabiand__ has joined #openstack18:38
uvirtbotNew bug: #798878 in nova "nova-XXX respawns constantly -- [Errno 111] Connection refused" [Undecided,New]
*** mrjazzcat has joined #openstack18:43
*** bibabot has joined #openstack18:47
*** zorzar has quit IRC18:50
*** zorzar has joined #openstack18:54
*** lorin1 has quit IRC18:55
*** llang629 has joined #openstack18:57
*** kennethkalmer has joined #openstack18:58
*** Razique has joined #openstack18:59
*** amccabe has quit IRC19:00
*** robjo1 has joined #openstack19:00
*** ccooke has quit IRC19:00
*** ziyadb has joined #openstack19:00
*** ziyadb has joined #openstack19:00
*** zorzar has quit IRC19:00
*** ziyadb_ has quit IRC19:00
robjo1Where would I find "nightly" tarballs of the development tree?19:00
*** amccabe has joined #openstack19:01
*** ccooke has joined #openstack19:02
*** zorzar has joined #openstack19:03
*** dgags has joined #openstack19:05
*** Razique has joined #openstack19:05
*** Razique has joined #openstack19:06
*** Razique has quit IRC19:07
*** Razique has joined #openstack19:07
*** robjo1 has left #openstack19:07
Raziquehi all19:07
heckjrobjo1 at launchpad in the various project areas. I don't believe that we're building PPA's and associated packages nightly from the development trunks.19:12
*** winston-d has quit IRC19:15
*** kennethkalmer has quit IRC19:20
kbringardI do think there is "continuous" integration with the ppa apt repo though19:21
kbringardthe trunk apt repo that is19:21
kbringardit's usually not much more than an hour or so behind19:21
*** Razique has quit IRC19:22
*** Razique has joined #openstack19:22
*** ziyadb has quit IRC19:26
*** ziyadb has joined #openstack19:26
*** renuka has joined #openstack19:27
renukahello i had a question about the order of calls in the __init__ of VolumeManager19:28
*** iammartian has joined #openstack19:34
*** iammartian has left #openstack19:34
*** renuka has quit IRC19:35
*** clauden has joined #openstack19:35
*** jdarcy_ has joined #openstack19:37
*** ziyadb_ has joined #openstack19:37
*** ziyadb has quit IRC19:39
*** jdarcy has quit IRC19:39
*** jdarcy_ is now known as jdarcy19:39
*** jdarcy_ has joined #openstack19:41
*** mdomsch_westford has quit IRC19:41
*** jdarcy has quit IRC19:41
*** jdarcy_ is now known as jdarcy19:41
*** ziyadb_ has quit IRC19:42
*** llang629 has left #openstack19:45
*** Ephur has joined #openstack19:48
*** huslage has joined #openstack19:49
adam_gim trying to get started with glance, using default_store = file.  using either uec-publish-tarball or 'glance add', images are registered but the corresponding files in /var/lib/glance/images are zero-length.  is this configuration error or is my understanding of local file backend wrong?19:54
*** amccabe has quit IRC19:54
adam_ghere is a log of a run of uec-publish-tarball:
*** Ephur has quit IRC19:56
Raziquehi adam_g19:57
Raziquewhat euca-describe-images returns ?19:58
*** jdarcy has quit IRC19:59
*** kd926 has quit IRC20:00
adam_gRazique: a good lookin list of iamges:
adam_g , i wonder.20:01
*** bibabot has quit IRC20:01
*** jdarcy has joined #openstack20:03
*** amccabe has joined #openstack20:08
*** Ryan_Lane is now known as Ryan_Lane|away20:10
Raziquetry to restart nova-objectstore, nova-api and glance services20:11
Raziqueand see if the problem occurs again20:11
Raziquefor glance : glance-control all stop && glance-control all start20:11
kbringardhey soren, you around sir?20:11
sorenkbringard: I am.20:12
kbringardyou do most of the iptables code, yea?20:12
*** cowmix has joined #openstack20:12
sorenthat's the theory.20:12
kbringardhehe, would you mind taking a look at: real quick20:13
* soren looks20:13
kbringardto see if something jumps out at you20:13
kbringardthe quick and dirty is, if you update a rule in a group with instances running on it, the nova-compute-local group on the compute nodes with instances from that group ends up updating everything to point to whatever the first instance in the chain is20:14
kbringardyou can service nova-compute restart and it goes back to what it should be20:14
*** dgags has quit IRC20:14
sorenkbringard: oh, dear.20:14
sorenDear, oh dear, oh dear.20:15
kbringardit looks like it's been that way for awhile20:15
sorenProbably. That stuff hasn't changed much in a while.20:15
kbringardthe biggest issue is that if something has a floater assigned to it20:15
kbringardit's instantly bare assed to the world20:15
kbringardat least in my environment20:16
sorenI think I know sort of where it happens.20:16
kbringardsorry, I spent the morning looking at it, and realized I wasn't familiar enough with this code to figure it out quickly, hehe20:18
sorenGimme a couple of minutes.20:18
kbringardno worries20:18
kbringardwhile you're at it... if you think it's related... it looks like --source-group stuff doesn't work as expected either20:18
kbringardbut that's less important at this very moment, imo20:18
*** huslage has quit IRC20:19
sorenkbringard: Yeah, ti doesn't.20:19
kbringardlol, cool, glad you know :-)20:19
kbringardanyway, I'll leave you be. Thanks for taking a look20:20
*** dolphm has quit IRC20:23
*** pothos_ has joined #openstack20:27
*** pothos has quit IRC20:29
*** mgoldmann has quit IRC20:29
*** pothos_ is now known as pothos20:29
*** llang629_ has joined #openstack20:34
*** llang629_ has left #openstack20:35
sorenkbringard: Man, I don't blame you for not figuring this out. It's complicated :)20:38
kbringardindeed, which is why I came to the mastah20:39
*** obino1 has quit IRC20:46
*** obino has joined #openstack20:46
MarkusWPHRorgI'm having kind of a similar problem20:47
MarkusWPHRorgFreshly booted machine running nova compute, swift, network, pretty much everything20:48
MarkusWPHRorgwhen I run euca-allocate-address list20:48
MarkusWPHRorgI get UnknownError: An unknown error has occurred. Please try your request again.20:48
MarkusWPHRorgand lots of TRACE messages in the log20:48
*** fabiand__ has quit IRC20:49
daboMarkusWPHRorg: Do any of them mention permissions? There was a bug fix today for that20:49
MarkusWPHRorgI've seen similar behavior when I used an ip address instead of hostname in my "nova-manage floating" command20:49
daboCheck your 'users' table in the nova database - if 'is_admin' is 0, try changing that column to 1 and re-run. If the errors go away, then that's the bug that was found and fixed20:50
MarkusWPHRorgNot seeing that yet - still looking20:50
MarkusWPHRorgThere is this:  ERROR nova [-] Exception during message handling20:50
*** flk_ has quit IRC20:51
*** ctennis has quit IRC20:51
kbringarddabo: are you talking about the bug when launching instances?20:51
kbringardMarkusWPHRorg: is your rabbitmq running?20:51
dabokbringard: yeah - it affected lots of things, not just launching20:51
kbringardah so, i did not know that20:52
kbringardglad I fixed it20:52
MarkusWPHRorgkbringard: how can I check?20:52
daboi am too20:52
*** markvoelker has quit IRC20:52
kbringardMarkusWPHRorg: service rabbitmq status20:52
MarkusWPHRorgya it's up20:54
kbringardwhat rev are you running?20:54
MarkusWPHRorgit was "service rabbitmq-server status"20:54
kbringard1194 I think is the newest20:54
kbringardoh, sorry20:54
MarkusWPHRorgs ok :)20:54
kbringardI always just tab complete, so I forget the exact service names, hehe20:54
kbringard<--- lazy20:54
MarkusWPHRorghow can I find out my version?20:55
MarkusWPHRorgI'm running whatever version installs with "apt-get" so I may be a ways behind20:57
kbringarddpkg -l | grep nova20:57
kbringardit'll be like20:58
kbringardhmmm, that's pretty old... is that the cactus released version?20:58
MarkusWPHRorgnot sure, but maybe20:58
MarkusWPHRorgis there a way to upgrade, or should I attempt the install from scratch?20:59
sorenUpgrades should work.20:59
sorenif they don't, please let us know.20:59
MarkusWPHRorgHow do I do the upgrade - apt-get?20:59
kbringardit seems likely that you have the "stable" repo in your sources as opposed to the trunk repo20:59
kbringardunless you've not upgraded in a couple of months21:00
MarkusWPHRorgI just installed it fresh on ubuntu 11.4 a couple weeks ago21:00
MarkusWPHRorgwhat do I need to change in my sources to get the "trunk repo?"21:00
*** ctennis has joined #openstack21:04
*** ctennis has joined #openstack21:04
*** ameade has quit IRC21:09
heckjMarkusWPHRorg: do you really want trunk, or the latest milestone release (i.e. diablo-1)?21:10
*** ewindisch has joined #openstack21:10
heckjMarkusWPHRorg: I've been using the to install the latest and greatest milestone bits21:12
heckjcurl -O
heckjchmod 755 nova.sh21:12
heckj./ branch lp:~hudson-openstack/nova/milestone-proposed21:12
heckj./ install21:12
*** foxtrotdelta has quit IRC21:19
*** kashyap has quit IRC21:19
*** Ryan_Lane|away is now known as Ryan_Lane21:25
uvirtbotNew bug: #798965 in nova "List Servers API call does not return "flavorRef" attribute. To get it we have to do Get Server API call for each server in a list." [Undecided,New]
MarkusWPHRorgheckj: Thank you - that's very helpful21:30
uvirtbotNew bug: #798966 in nova "We'd like to have a visibility attribute ('public'/'private') for Images." [Undecided,New]
uvirtbotNew bug: #798967 in nova "public Server addresses are kept in a list of private addresses while the publist list is empt" [Undecided,New]
uvirtbotNew bug: #798969 in nova "error messages should have details" [Undecided,New]
*** lborda has quit IRC21:36
*** medberry is now known as med_out21:44
*** jkoelker has quit IRC21:48
*** lvaughn has quit IRC21:56
*** lvaughn has joined #openstack21:57
*** FallenPegasus has joined #openstack21:58
*** FallenPegasus is now known as MarkAtwood21:58
*** PeteDaGuru has quit IRC21:59
*** lborda has joined #openstack22:01
*** lvaughn has quit IRC22:03
*** imsplitbit has quit IRC22:07
*** sebastianstadil has quit IRC22:12
sorenkbringard: AHA!22:15
kbringardfound it, eh?22:16
sorenI did.22:16
kbringardthat is because you rule22:16
*** amccabe has quit IRC22:16
sorenI'm stubborn. there's a fine line there.22:16
kbringardhaha, fair enough22:18
sorenYup, totally nailed this one.22:21
sorenI'm absolutely beat, though.22:21
kbringardlol, sorry22:21
kbringardthanks for fixing it :-)22:21
sorenkbringard: Not your fault :)22:21
sorenkbringard: I'll epxlain on the bug, so that I can fix it later.22:21
kbringardhehe, true, I was empathizing22:21
*** allsystemsarego has quit IRC22:23
*** jonkelly has quit IRC22:24
*** shentonfreude has quit IRC22:31
*** kbringard has quit IRC22:35
*** lborda has quit IRC22:38
*** rnirmal has quit IRC22:51
*** watcher has joined #openstack22:57
*** miclorb has joined #openstack22:59
*** miclorb has quit IRC23:05
*** lborda has joined #openstack23:11
*** freeflyi1g has joined #openstack23:16
*** freeflying has quit IRC23:19
*** pguth66 has joined #openstack23:22
*** rnirmal has joined #openstack23:24
*** rnirmal has quit IRC23:24
*** ewindisch has quit IRC23:25
*** vladimir3p has quit IRC23:26
*** ewindisch has joined #openstack23:26
*** fysa has quit IRC23:30
*** ewindisch has quit IRC23:30
*** Razique has quit IRC23:39
uvirtbotNew bug: #798998 in nova "format_images in ec2_api does not correctly return owner id" [Undecided,New]
*** fysa has joined #openstack23:51

Generated by 2.14.0 by Marius Gedminas - find it at!