Friday, 2011-02-11

*** dendrobates is now known as dendro-afk00:00
vishystill there, what is wrong with it?00:01
vishycurl -O should get it for you00:01
*** littleidea has joined #openstack00:07
*** jsgotangco has quit IRC00:09
*** dragondm has quit IRC00:10
vishykpepple still here?00:11
kpepplevishy: yes. jk0 and i are still hacking away.00:12
vishyso you guys get an OperationalError no such column?00:12
kpeppleno such table ... but only because it's looking for the sqlite db instead of the configured mysql db.00:12
vishyi get OperationalError00:13
vishylet me switch to mysql instead00:14
vishyi get no such column services.availability_zone00:14
jk0vishy: we tracked the problem down to rev 64100:14
*** hggdh has quit IRC00:14
kpepplejk0: did you paste that diff for vishy00:15
vishyi still get no_such_column00:15
jk0kpepple: which paste?00:15
kpepplejk0: the diff between 640 and 64100:15
*** westmaas1 has joined #openstack00:16
jk0here it is:
vishyi can't replicate00:16
vishywhen i run using your code compute fails with no such column: services.availability_zone00:16
kpepplevishy: to replicate, trash your nova.sqlite file. then "bin/nova-compute -f /etc/nova/nova.conf" assuming the nova.conf points to mysql00:17
vishydoes -f work?00:17
vishynever used that before00:17
*** hggdh has joined #openstack00:17
vishymine is connecting to mysql db just fine00:17
jk0did you merge with trunk or anything?00:18
vishyjk0 nope00:18
jk0well wtf00:18
vishyi have to get on a plane, but I've never used -f syntax before00:18
* jk0 is using --flagfile00:18
jk0vishy: if it helps00:18
vishyi'm using default_flagfile so that could make a diff00:19
jk0what's weird is this:00:19
vishyare you using the right flagfile when you nova-manage db sync?00:19
jk02011-02-10 18:18:25,052 DEBUG nova.root [-] sql_connection : mysql://root:nova@ from MainProcess (pid=19446) serve /home/jkearney/openstack/dynamicinstancetypes/nova/
jk0so it's seeing the sql flag, but not using it00:20
vishynova-manage --flagfile=/etc/nova/nova.conf db sync00:20
kpepplei just ran this with flagfile ... see this paste
vishyok I'll try to replicate more from the plane00:20
vishysee you later00:20
jk0thanks dude00:20
kpepplein line 53, the sql_conn is set correctly ... in line 145 (a log i added), it's not set correctly00:20
kpepplevishy: thanks00:21
*** hggdh has quit IRC00:22
*** hggdh has joined #openstack00:22
*** littleidea has quit IRC00:24
*** rpedde has joined #openstack00:25
*** littleidea has joined #openstack00:26
*** clauden_ has quit IRC00:29
*** adiantum has joined #openstack00:31
*** kpepple_ has joined #openstack00:38
*** kpepple has quit IRC00:38
*** kpepple_ is now known as kpepple00:38
*** kpepple has quit IRC00:40
*** dubsquared has left #openstack00:41
*** littleidea has quit IRC00:45
*** adiantum has quit IRC00:47
*** dendro-afk is now known as dendrobates00:56
*** MarcMorata has quit IRC00:57
*** adiantum has joined #openstack00:59
*** maplebed has quit IRC01:01
*** adiantum has quit IRC01:07
*** alekibango has joined #openstack01:07
*** MarcMorata has joined #openstack01:10
*** adjohn has joined #openstack01:11
*** adiantum has joined #openstack01:11
*** joearnold has quit IRC01:14
*** alekibango has quit IRC01:23
*** alekibango has joined #openstack01:43
*** MarcMorata has quit IRC01:44
*** dfg has quit IRC01:44
*** miclorb has quit IRC01:45
*** adjohn has quit IRC01:48
*** dendrobates is now known as dendro-afk01:49
*** azneita has joined #openstack01:52
*** adjohn has joined #openstack01:53
*** kpepple has joined #openstack02:03
*** arcane has quit IRC02:04
*** littleidea has joined #openstack02:07
*** dirakx1 is now known as dirakx_afk02:10
HugoKuohow can I know the current version of euca on my computer ?02:12
HugoKuoinstalled from package02:13
*** hub_cap has joined #openstack02:14
*** arcane has joined #openstack02:19
*** alekibango has quit IRC02:31
*** alekibango_ has joined #openstack02:31
*** cascone has joined #openstack02:32
*** adiantum has quit IRC02:34
*** adiantum has joined #openstack02:35
*** miclorb_ has joined #openstack02:35
kpeppleHugoKuo: euca-version02:36
HugoKuo1.2-31337 2009-04-04   ..... is this the correct one ?02:37
*** adjohn has quit IRC02:38
*** adjohn has joined #openstack02:40
nelsoncreiht: hey. I figured out how that code works (in
kpeppleHugoKuo: that is what i am using and it works with nova bexar +02:43
nelsonSwift3Middleware gets called. It looks at the path to figure out what controller needs to be called.02:43
nelsonIt calls the controller, which returns a webob.Response.02:43
nelsonIt has a callable which expects an environment and start_response.02:44
*** rpedde has quit IRC02:44
nelsonSo it basically looks like this: return getattr(controller, req.method)(env, start_response)(env, start_response)02:45
nelsonBut it could also look like this: return getattr(controller, req.method)(env, None)(env, start_response)02:47
nelsonbecause the first mention of start_response never gets called.02:47
*** cascone has quit IRC02:48
nelsonNeeds more comments!!!02:49
*** kpepple has quit IRC02:51
HugoKuothanks kpepple02:56
*** littleidea has quit IRC02:56
KnightHackerjoin #project102:57
KnightHackerops :)02:57
HugoKuoyou mean there's a mix cloud environment in you infrastructure ?  includ euca and nova ?02:57
*** adiantum has quit IRC02:58
*** adjohn has quit IRC03:00
*** Ryan_Lane is now known as Ryan_Lane|away03:00
*** lvaughn has joined #openstack03:02
*** adiantum has joined #openstack03:03
*** hadrian has joined #openstack03:05
*** adiantum has quit IRC03:13
*** adiantum has joined #openstack03:14
*** dirakx_afk is now known as dirakx03:18
*** adiantum has quit IRC03:28
*** whaley has joined #openstack03:29
*** burris has quit IRC03:30
*** Ryan_Lane|away has quit IRC03:30
*** btorch has quit IRC03:30
*** colinnich has quit IRC03:30
*** _0x44 has quit IRC03:30
*** RichiH has quit IRC03:30
*** rackerhacker has quit IRC03:30
*** Brainspackle has quit IRC03:30
*** devcamcar has quit IRC03:30
*** adiantum has joined #openstack03:35
*** localhost has joined #openstack03:38
*** burris has joined #openstack04:01
*** Ryan_Lane|away has joined #openstack04:01
*** btorch has joined #openstack04:01
*** RichiH has joined #openstack04:01
*** colinnich has joined #openstack04:01
*** devcamcar has joined #openstack04:01
*** _0x44 has joined #openstack04:01
*** rackerhacker has joined #openstack04:01
*** Brainspackle has joined #openstack04:01
*** sets mode: +v _0x4404:01
*** RJD22 is now known as RJD22|away04:02
*** RJD22|away is now known as RJD2204:07
*** Xenith has quit IRC04:17
*** localhost has quit IRC04:23
*** sateesh has joined #openstack04:24
*** localhost has joined #openstack04:24
*** Xenith has joined #openstack04:31
*** lvaughn has quit IRC04:35
*** dendro-afk is now known as dendrobates04:45
*** hazmat has quit IRC04:56
*** kashyapc has joined #openstack05:02
creihtnelson: ahh... cool05:16
uvirtbotNew bug: #716847 in nova "euca-delete-volume cannot delete volumes in error state" [Undecided,New]
*** kashyapc has quit IRC05:27
*** adiantum has quit IRC05:30
*** ccustine has quit IRC05:36
*** adiantum has joined #openstack05:36
*** dendrobates is now known as dendro-afk05:41
*** BK_man has quit IRC05:45
*** adiantum has quit IRC05:46
*** f4m8_ is now known as f4m805:48
*** BK_man has joined #openstack05:54
*** adiantum has joined #openstack05:58
*** littleidea has joined #openstack06:01
*** sateesh has quit IRC06:09
*** sateesh has joined #openstack06:09
*** dkocher has joined #openstack06:11
*** hadrian has quit IRC06:19
*** Ryan_Lane|away is now known as Ryan_Lane06:42
*** dkocher has quit IRC06:44
*** adiantum has quit IRC06:45
*** omidhdl has joined #openstack06:47
*** miclorb_ has quit IRC06:56
*** adiantum has joined #openstack06:57
*** littleidea has quit IRC06:58
*** kpepple has joined #openstack06:58
*** elimisteve has joined #openstack07:17
*** adiantum has quit IRC07:19
*** George has joined #openstack07:21
Georgeanyone can help?07:24
kpepplewith what ?07:24
Georgeinstalling openstack on centos07:24
kpeppleahhh ... centos 5 ... this thing, it is not easy07:24
kpeppledid you look on the wiki for vvukkans instructions ?07:25
*** CloudChris has joined #openstack07:25
Georgei just see this link07:26
kpepplethe tricky part is the openssl/m2crypto and qemu-img patch ... the rest is pretty easy07:27
*** Ryan_Lane is now known as Ryan_Lane|away07:28
*** fabiand_ has joined #openstack07:31
*** adiantum has joined #openstack07:31
Georgeso centos is not ready ,any yum update can break openstack?07:33
*** BK_man has quit IRC07:36
kpeppleyes, certain yum updates will break openstack (nova) on centos 5. when centos 6 comes out, things will be much better -- there are already RHEL 6 packages out.07:37
Georgeon instructions swig is 64bit version07:42
Georgeswig-1.3.25-1.el5.rf.i386.rpm installing 32bit is ok?07:43
Georgeor 64bit is requirement for cloud?07:43
*** befreax has joined #openstack07:44
kpepplewithout 64 bit, you probably will not have KVM or Xen support ... you will have to use UML (which is slow)07:44
kpepplei've never run it on a 32bit machine07:44
Georgexen supports 32bit07:44
kpeppleit does ? i didn't know that (i run my devel box with kvm)07:45
Georgei have a test vm on xen07:45
Georgeit works like a charm07:45
kpeppleok, i *think* that the 64 bit is mostly just so that you get kvm or xen ... so i think your 32bit swig would be fine07:46
*** adiantum has quit IRC07:47
Georgeon module load kvm fails/not found ,and i go up on screen i see that yum hasnt installed kvm07:53
Georgeis it possible openstack to work with xen?07:53
*** adiantum has joined #openstack07:53
kpepplei don't think kvm can't be loaded on a 32 bit machine .. yes, openstack runs with Xen ... but those instruction are for kvm07:54
kpepplelet me find the xen instructions. hold on.07:54
kpeppleokay ... here are the xenserver instructions -
kpepplei haven't worked with that ... so if you have problems, you might check back in the morning when jk0 or some of the other people are awake07:56
Georgexen is already installed on this machine from before07:56
kpeppleare you installing in dom 007:56
George[root@server aoetools-32]# xm list07:57
GeorgeName                                      ID Mem(MiB) VCPUs State   Time(s)07:57
GeorgeDomain-0                                   0     1754     4 r-----   2075.207:57
Georgeiam on the physical machine07:58
kpeppleah, yes. some other people have done this. you can probably skip over the modprobe and go right to the installing the python libraries07:59
Georgeeasy install 2.6 missing08:00
Georgeyum install hasnt installed python 2.608:01
Georgeit installed python 2.508:01
kpeppledid you do step #2 on ? that installs python2.608:01
Georgeon this step it didnt installed python2608:03
Georgei found now from other site a repository that contains python26 for centos08:03
kpeppleand you can't do python26 from the command line ?08:03
Georgepackage hasnt been installed08:03
Georgei found
kpeppleyou need to add the epel -- python26 is there08:04
kpepple# rpm -Uvh ''08:05
kpepplethe extra repositories ... they keep python26 and such there08:05
Georgethis is the mistake08:07
kpeppleoh ... yeah, sorry. forgot you were 32bit.08:08
Georgethats why it failed08:08
Georgebut seems like chl repository has installed python08:08
*** thimble has joined #openstack08:09
Georgeeasy_install2.6 not exists maybe i have to install python26-setuptools08:09
kpeppleyes, must have setuptools08:10
Georgei will install epel anyway to be compatible with the docs08:11
Georgei installed setup tools and still cant find easy install-2.6 so i will add the rpm of epel08:12
kpeppleisn't it easy_install26 ?08:12
George[root@server aoetools-32]# easy_install08:13
Georgeeasy_install      easy_install-2.408:13
George[root@server aoetools-32]# easy_install08:13
Georgeexcept if is easy_install08:13
kpeppleno easy_install will be a symlink to easy_install-2.408:13
Georgedo i need setuptools-devel also?08:14
*** guigui1 has joined #openstack08:14
kpepplenot sure08:15
*** adiantum has quit IRC08:15
GeorgeTransaction Check Error:08:15
George  file /usr/bin/easy_install from install of python26-setuptools-devel-0.6c9-5.noarch conflicts with file from package python-setuptools-0.6c5-2.el5.noarch08:15
Georgelets forget chl i will try epel08:16
*** adiantum has joined #openstack08:17
*** rcc has joined #openstack08:17
Georgehope now to have easy_08:23
Georgethanks in advance whatever will be the result08:24
*** perra has joined #openstack08:24
*** masumotok has joined #openstack08:24
kpeppleg'night everyone08:30
*** kpepple has quit IRC08:31
Georgegoodnight kpepple thanks08:32
*** drico_ has joined #openstack08:33
*** adiantum has quit IRC08:35
*** adiantum has joined #openstack08:40
*** miclorb has joined #openstack08:54
*** thimble has quit IRC08:56
*** masumotok has quit IRC09:00
Georgeanyone able to advise me for nova on centos?09:04
*** photron has joined #openstack09:05
openstackhudsonProject dashboard-tarmac build #2,685: FAILURE in 35 sec:
openstackhudsonProject nova-tarmac build #60,306: FAILURE in 36 sec:
*** nati_ has joined #openstack09:07
Georgei get this on last09:07
George[root@server nova]# nova-manage --flagfile /etc/nova/nova.conf db sync09:07
George[Errno 2] No such file or directory: '/etc/nova/nova.conf'09:07
GeorgeERROR:: Unable to open flagfile: /etc/nova/nova.conf09:07
ttxGeorge: do you have a flag file in /etc/nova/nova.conf ?09:08
Georgewhat is flag file?09:08
ttxa sort of config file.09:09
George[root@server nova]# ls /etc/nova09:09
George[root@server nova]#09:09
George[root@server nova]# nova-manage --flagfile /etc/nova/nova-api.conf db sync09:10
George[DEFAULT] does not match any options:09:10
George        user09:10
ttxGeorge: see my nova.conf file at
openstackhudsonProject dashboard-tarmac build #2,686: STILL FAILING in 35 sec:
openstackhudsonProject nova-tarmac build #60,307: STILL FAILING in 36 sec:
ttxnova-manage looks up nova.conf to see where the db is, by default a sqlite DB at state_path/nova.sqlite09:12
nati_Hi Do anyone knows how to use ajax_console ?09:13
*** miclorb has quit IRC09:14
ttxnati_: best doc I've seen so far is the "Description of the Change" at
nati_:ttx Thank you!09:15
ttxnati_: we need better doc for that09:15
*** BK_man has joined #openstack09:16
openstackhudsonProject dashboard-tarmac build #2,687: STILL FAILING in 37 sec:
nati_:ttx I think so too, I'm reading code now in order to figure out how to use it.09:16
openstackhudsonProject nova-tarmac build #60,308: STILL FAILING in 35 sec:
* ttx gets some coffee09:19
openstackhudsonProject dashboard-tarmac build #2,688: STILL FAILING in 36 sec:
openstackhudsonProject nova-tarmac build #60,309: STILL FAILING in 35 sec:
*** miclorb_ has joined #openstack09:22
openstackhudsonProject dashboard-tarmac build #2,689: STILL FAILING in 35 sec:
openstackhudsonProject nova-tarmac build #60,310: STILL FAILING in 35 sec:
*** calavera has joined #openstack09:30
openstackhudsonProject dashboard-tarmac build #2,690: STILL FAILING in 36 sec:
openstackhudsonProject nova-tarmac build #60,311: STILL FAILING in 35 sec:
sorenLooks like someone forgot to set a commit message.09:36
openstackhudsonProject dashboard-tarmac build #2,691: STILL FAILING in 36 sec:
sorenErr. no.09:36
openstackhudsonProject nova-tarmac build #60,312: STILL FAILING in 36 sec:
sorenLaunchpad scheduled downtime.09:37
*** Nacx has joined #openstack09:40
openstackhudsonProject dashboard-tarmac build #2,692: STILL FAILING in 35 sec:
*** Hayati has joined #openstack09:46
openstackhudsonProject dashboard-tarmac build #2,693: STILL FAILING in 37 sec:
wedjathi everyone09:47
wedjatdocumentation on swift needs to be updated here for swauth part :
openstackhudsonProject dashboard-tarmac build #2,694: STILL FAILING in 35 sec:
*** thimble has joined #openstack09:54
thimblehi all: the following is an error while runnning vlanmode. (nova.root): TRACE: Command: -E dnsmasq --strict-order --bind-interfaces --conf-file= --domain=novalocal --pid-file=/var/lib/nova/networks/ --listen-address= --except-interface=lo --dhcp-range=,static,120s --dhcp-hostsfile=/var/lib/nova/networks/nova-br100.conf --dhcp-script=/usr/bin/nova-dhcpbridge --leasefile-ro09:55
thimble(nova.root): TRACE: Exit code: 209:55
openstackhudsonYippie, build fixed!09:55
openstackhudsonProject dashboard-tarmac build #2,695: FIXED in 2.8 sec:
*** omidhdl has quit IRC09:56
*** miclorb_ has quit IRC09:59
*** Hayati has quit IRC10:06
*** nati_ has quit IRC10:06
*** miclorb has joined #openstack10:10
*** miclorb has quit IRC10:14
openstackhudsonYippie, build fixed!10:21
openstackhudsonProject nova-tarmac build #60,313: FIXED in 2.6 sec:
*** miclorb_ has joined #openstack10:24
wedjati'm searching for hours how to create an account with swauth : it simply doesn't work as suggested in the documentation10:25
*** Hayati has joined #openstack10:26
wedjatthe -A option is missing as it implies https by default10:26
wedjatthen, seems to be a syntax problem with swauth architecture and the swauth-add-user script, as it launches the exception : "Could not create account within main auth account"10:27
*** miclorb_ has quit IRC10:29
wedjatby the way, is it v1 or v2 for swauth ?10:31
*** adiantum has quit IRC10:39
ttxsoren: bug 716919 is confusing me. I don't see how ec2_host would be set to in nova-api10:45
uvirtbotLaunchpad bug 716919 in nova "FLAGS.ec2_host is set to in nova-api" [Undecided,New]
ttxI see it now10:46
uvirtbotNew bug: #716919 in nova "FLAGS.ec2_host is set to in nova-api" [Undecided,New]
*** sandywalsh_ has joined #openstack10:51
*** Hayati has quit IRC10:52
*** adiantum has joined #openstack10:52
*** rds__ has joined #openstack10:52
*** sandywalsh_ has quit IRC10:53
*** Hayati has joined #openstack10:56
*** adiantum has quit IRC10:59
*** metoikos has quit IRC11:04
*** adiantum has joined #openstack11:05
*** metoikos has joined #openstack11:08
*** metoikos has quit IRC11:13
sorenttx: This sucks. There seems to be some sort of concurrency issue somewhere. I can run hundreds of instances in a row if I do it sequentially. If I do multiple at once, a lot of them end up failing like this:,distro=maverick,userdata=no/console (search for CONTROL-D).11:16
*** metoikos has joined #openstack11:20
ttxcan you reproduce it with .. hmmm.. simpler images ? Try to put the UEC image out of the equation ?11:20
ttxsoren: ^11:20
ttxI wouldn't be surprised if there was still a race in UEC images / upstart that you end up losing under load11:21
ttxI've seen that one happening in the UEC/Eucalyptus load testing as well.11:22
* ttx breaks for lunch11:23
*** Hayati_ has joined #openstack11:26
*** guigui1 has quit IRC11:26
*** Hayati has quit IRC11:27
sorenttx: Aha! That's very useful info.11:28
sorenttx: Thanks!11:28
*** Hayati_ has quit IRC11:31
*** omidhdl has joined #openstack11:35
*** thimble has quit IRC11:39
sorenttx: Fascinating. If I run the tests on a Maverick host, it's fine.11:39
Georgei have instaleld nova on centos11:40
Georgei need a script for automatically starting/running it11:41
*** metoikos has quit IRC11:43
*** MarcMorata has joined #openstack11:43
alekibango_George: be patient... we mostly use deb systems. sometimes it takes 20 minutes before someone answers... and most people are sleeping atm, they will come in 2-3 hours11:45
*** adiantum has quit IRC11:46
alekibango_there for sure are some init scripts11:46
George:) yes timezone difference11:46
alekibango_arent there RPM packages ?  going with source is harder11:46
Georgeno rpm11:50
Georgei started services11:51
Georgeif i put it on init.d will be ok?11:52
*** allsystemsarego has joined #openstack11:56
*** allsystemsarego has joined #openstack11:56
sorenGeorge: That's where init scripts usually go :)11:57
Georgeyes but there is nowhere a start stop command11:58
Georgelike most scripts11:58
Georgestop normally has to kill each proccess11:58
*** rds__ has quit IRC12:00
sorenWe don't provide init scripts for RedHat.12:10
sorenThere are RHEL packages, though.12:10
sorenThey probably have init scripts.12:10
sorenI don't know where they are though. I don't care much about RHEL (or Fedora or CenTOS or whatever other flavour of RedHat people are into).12:11
Georgeif i start services whats next12:11
*** hazmat has joined #openstack12:15
*** ksteward has joined #openstack12:20
*** ctennis has quit IRC12:32
sorenGeorge: There are installation instructions on the wiki.12:34
*** ksteward has quit IRC12:34
sorenGeorge: Just skip the parts about installing the services, since you've already done that.12:35
*** omidhdl has left #openstack12:41
*** westmaas1 has quit IRC12:41
sorensandywalsh: Bonjour.12:47
*** ctennis has joined #openstack12:49
*** ctennis has joined #openstack12:49
*** ksteward has joined #openstack12:58
*** calavera has quit IRC13:00
sorenttx: Prime suspect right now: mountall.13:02
*** allsystemsarego has quit IRC13:02
sorenttx: Well, plymouth to some extent, but mountall should handle plymouth's failure gracefully.13:03
*** allsystemsarego has joined #openstack13:04
*** Hayati_ has joined #openstack13:05
ttxsoren: you should talk to smoser, ISTR he knows the root cause13:10
*** guigui1 has joined #openstack13:10
*** MarcMorata has quit IRC13:15
*** Hayati_ has quit IRC13:15
*** hggdh has quit IRC13:19
*** hggdh has joined #openstack13:20
*** adiantum has joined #openstack13:25
*** h0cin has joined #openstack13:29
*** metoikos has joined #openstack13:40
*** lvaughn has joined #openstack13:44
*** Hayati_ has joined #openstack13:53
*** pvo has joined #openstack13:54
alekibango_hmm i am looking for way to build my own swift bexar packages... cant find release packages in  ...13:56
alekibango_soren: this should work for packaging swift release?
*** metoikos has quit IRC13:57
*** fabiand_ has quit IRC13:58
alekibango_its not mentioned on Packaging wiki page...13:58
sorenalekibango_: That's the theory, yeah.14:02
smosersoren, in that log, is it intermixed with anything ?14:03
smoseror is that serial output of one instance boot14:03
sorensmoser: The part that ends with the CONTROL-D thing is the output of euca-get-console-output.14:04
sorensmoser: All the way up to the stuff from the kernel. EVerything before that is just noise (for you).14:04
smoserright. ok. i was wondering if it was all a single output, or if multiple instances were interweaved.14:04
sorensmoser: No, just one instance.14:05
smoserso, focusing on that.14:05
sorenI think I've fixed it.14:05
smoserhow ?14:05
alekibango_soren: will try, and update the wiki14:05
sorensmoser: In mountall.14:05
sorensmoser: Added a SIGPIPE handler that closes the plymouth connetion.14:06
sorensmoser: It doesn't seem to handle SIGPIPE at all.14:06
alekibango_btw why there are no swift bexar packages in archive? :)14:06
smoserhm.. well that sounds wonderful.14:06
sorensmoser: If Plymouth explodes and mountall attempts to write to the fd, it'll SIGPIPE.14:06
smoseri have seen this failure before14:06
*** eikke has joined #openstack14:06
smoseri dont recall where, but very rarely, you increased load is tripping it i think14:07
alekibango_wow... i just started using plymouth... sad to hear this14:07
smoserwhat was strange to me is that / was already mounted RW14:07
smoserand then init starts complaining about a failure to mount /14:07
alekibango_smoser: it might be another filesystem (eg ram)14:07
smoserbut you're probably right, soren.14:07
wedjatalekibango_: i built myself a deb package for swift 1.2.0, it's not difficult14:08
sorensmoser: certainly suggests that I'm right. Not a single failure since I applied that patch.14:09
* soren is loving using Jenkins to validate patches14:09
smosersoren, open a bug on mountall ?14:10
smoserhave you?14:10
*** hadrian has joined #openstack14:10
sorensmoser: I'm about to.14:10
sorensmoser: Just want to polish the patch a bit.14:10
sorensmoser: For one thing, it right now actually outputs a helpful message, saying that it caught a SIGPIPE.14:11
sorensmoser: That's not a very good idea.14:11
*** ppetraki has joined #openstack14:11
*** arcane has quit IRC14:11
sorensmoser: If, say, the fd that was actually unexpectedly closed is stdout. It'll just go bonkers in a tight loop trying to log stuff.14:11
ttxoh, I see you found each other.14:18
*** rds__ has joined #openstack14:19
*** Hayati_ has quit IRC14:19
pvosoren: re: you jenkins build, what is the 'no' column?14:22
*** dprince has joined #openstack14:23
sorenpvo: Whether it tries to do clever stuff with userdata.14:25
sorenpvo: If it's "no", it just logs in and echo's something. If it's "yes", it puts a script in user-data and checks that it was run correctly.14:25
*** flshn is now known as flashn14:26
*** gondoi has joined #openstack14:26
*** dendro-afk is now known as dendrobates14:29
*** gaveen has joined #openstack14:30
wedjatdoes someone know how to properly erase an account with devauth ?14:31
*** metoikos has joined #openstack14:37
*** f4m8 is now known as f4m8_14:44
*** kashyapc has joined #openstack14:49
*** arcane has joined #openstack15:00
*** jaypipes has joined #openstack15:02
*** littleidea has joined #openstack15:03
*** arcane has quit IRC15:05
*** dirakx has quit IRC15:07
*** pvo has quit IRC15:07
*** littleidea has quit IRC15:09
*** metoikos has quit IRC15:10
*** alekibango_ has quit IRC15:11
*** openstack` has joined #openstack15:17
*** heden has quit IRC15:17
*** syah has quit IRC15:17
*** dsockwell has quit IRC15:17
*** mark has quit IRC15:17
*** rdw has quit IRC15:17
*** decramy has joined #openstack15:17
*** dsockwell has joined #openstack15:18
*** openstack` is now known as openstack15:18
*** syah has joined #openstack15:18
*** soren has joined #openstack15:18
*** ChanServ sets mode: +v soren15:18
*** j05h has quit IRC15:19
*** termie has quit IRC15:19
*** johntash_ has joined #openstack15:19
*** deshantm has joined #openstack15:19
*** lvaughn has quit IRC15:20
*** Xenith has quit IRC15:20
*** blueadept has quit IRC15:20
*** wedjat has quit IRC15:20
*** tr3buchet has quit IRC15:20
*** imsplitbit has quit IRC15:20
*** londo has quit IRC15:20
*** sirp_ has quit IRC15:20
*** dabo has quit IRC15:20
*** blamar has quit IRC15:20
*** Adri2000 has quit IRC15:20
*** kashyapc has quit IRC15:20
*** gondoi has quit IRC15:20
*** lionel has quit IRC15:20
*** jbarratt has quit IRC15:20
*** annegentle has quit IRC15:20
*** uvirtbot has quit IRC15:20
*** brainproxy has quit IRC15:20
*** _cerberus_ has quit IRC15:20
*** klumpie has quit IRC15:20
*** smoser has quit IRC15:20
*** jaysun has quit IRC15:20
*** letterj has quit IRC15:20
*** sunech has quit IRC15:20
*** ctennis has quit IRC15:20
*** zul has quit IRC15:20
*** troytoman-away has quit IRC15:20
*** mtaylor has quit IRC15:20
*** HugoKuo has quit IRC15:20
*** EdwinGrubbs has quit IRC15:20
*** notmyname has quit IRC15:20
*** Jbain has quit IRC15:20
*** chmouel has quit IRC15:20
*** jeevan_ullas has quit IRC15:20
*** dovetaildan has quit IRC15:20
*** jarrod has quit IRC15:20
*** lool has quit IRC15:20
*** antonym has quit IRC15:20
*** adiantum has quit IRC15:20
*** h0cin has quit IRC15:20
*** befreax has quit IRC15:20
*** localhost has quit IRC15:20
*** mattinator has quit IRC15:20
*** pandemicsyn has quit IRC15:20
*** husimon has quit IRC15:20
*** rossij has quit IRC15:20
*** pquerna has quit IRC15:20
*** jaypipes has quit IRC15:20
*** eikke has quit IRC15:20
*** CloudChris has quit IRC15:20
*** jfluhmann has quit IRC15:20
*** iRTermite has quit IRC15:20
*** clayg has quit IRC15:20
*** gholt has quit IRC15:20
*** RJD22 has quit IRC15:20
*** yosh has quit IRC15:20
*** dprince has quit IRC15:20
*** drico_ has quit IRC15:20
*** sandywalsh has quit IRC15:20
*** fysa has quit IRC15:20
*** arun_ has quit IRC15:20
*** jtimberman has quit IRC15:20
*** filler has quit IRC15:20
*** shawn has quit IRC15:20
*** tomo_bot has quit IRC15:20
*** perra has quit IRC15:20
*** openstackhudson has quit IRC15:20
*** glenc has quit IRC15:20
*** cclien has quit IRC15:20
*** timrc has quit IRC15:20
*** photron has quit IRC15:20
*** George has quit IRC15:20
*** sateesh has quit IRC15:20
*** pothos has quit IRC15:20
*** h1nch has quit IRC15:20
*** taihen has quit IRC15:20
*** piezo has quit IRC15:20
*** soeren has quit IRC15:20
*** flashn has quit IRC15:20
*** karmabot has quit IRC15:20
*** ioso has quit IRC15:20
*** ksteward has quit IRC15:20
*** ianweller has quit IRC15:20
*** iammartian has quit IRC15:20
*** Daviey has quit IRC15:20
*** mfoemmel has quit IRC15:20
*** exlt has quit IRC15:20
*** odyi has quit IRC15:20
*** patri0t has quit IRC15:20
*** xtoddx has quit IRC15:20
*** rdw has joined #openstack15:20
*** fsaad has joined #openstack15:22
*** heden has joined #openstack15:22
*** tmarble has joined #openstack15:22
*** keekz has joined #openstack15:22
*** mark_ has joined #openstack15:22
*** ke4qqq has joined #openstack15:22
*** zedas has joined #openstack15:22
*** seats has joined #openstack15:22
*** konetzed has joined #openstack15:22
*** j05h has joined #openstack15:22
*** fitzdsl has joined #openstack15:22
*** ttx has joined #openstack15:22
*** sleepsonthefloor has joined #openstack15:22
*** jaypipes has joined #openstack15:22
*** kashyapc has joined #openstack15:22
*** gondoi has joined #openstack15:22
*** dprince has joined #openstack15:22
*** eikke has joined #openstack15:22
*** lvaughn has joined #openstack15:22
*** h0cin has joined #openstack15:22
*** adiantum has joined #openstack15:22
*** ksteward has joined #openstack15:22
*** ctennis has joined #openstack15:22
*** photron has joined #openstack15:22
*** drico_ has joined #openstack15:22
*** perra has joined #openstack15:22
*** befreax has joined #openstack15:22
*** CloudChris has joined #openstack15:22
*** sateesh has joined #openstack15:22
*** Xenith has joined #openstack15:22
*** localhost has joined #openstack15:22
*** sandywalsh has joined #openstack15:22
*** shawn has joined #openstack15:22
*** ianweller has joined #openstack15:22
*** blueadept has joined #openstack15:22
*** mattinator has joined #openstack15:22
*** jfluhmann has joined #openstack15:22
*** iammartian has joined #openstack15:22
*** fysa has joined #openstack15:22
*** pquerna has joined #openstack15:22
*** zul has joined #openstack15:22
*** troytoman-away has joined #openstack15:22
*** arun_ has joined #openstack15:22
*** pothos has joined #openstack15:22
*** openstackhudson has joined #openstack15:22
*** Daviey has joined #openstack15:22
*** h1nch has joined #openstack15:22
*** glenc has joined #openstack15:22
*** jtimberman has joined #openstack15:22
*** cclien has joined #openstack15:22
*** wedjat has joined #openstack15:22
*** Adri2000 has joined #openstack15:22
*** blamar has joined #openstack15:22
*** dabo has joined #openstack15:22
*** sirp_ has joined #openstack15:22
*** londo has joined #openstack15:22
*** imsplitbit has joined #openstack15:22
*** tr3buchet has joined #openstack15:22
*** mtaylor has joined #openstack15:22
*** patri0t has joined #openstack15:22
*** xtoddx has joined #openstack15:22
*** odyi has joined #openstack15:22
*** exlt has joined #openstack15:22
*** mfoemmel has joined #openstack15:22
*** HugoKuo has joined #openstack15:22
*** EdwinGrubbs has joined #openstack15:22
*** pandemicsyn has joined #openstack15:22
*** iRTermite has joined #openstack15:22
*** taihen has joined #openstack15:22
*** lionel has joined #openstack15:22
*** filler has joined #openstack15:22
*** notmyname has joined #openstack15:22
*** sets mode: +vvvv mtaylor exlt pandemicsyn notmyname15:22
*** tomo_bot has joined #openstack15:22
*** jbarratt has joined #openstack15:22
*** RJD22 has joined #openstack15:22
*** husimon has joined #openstack15:22
*** piezo has joined #openstack15:22
*** soeren has joined #openstack15:22
*** jaysun has joined #openstack15:22
*** Jbain has joined #openstack15:22
*** clayg has joined #openstack15:22
*** chmouel has joined #openstack15:22
*** annegentle has joined #openstack15:22
*** jeevan_ullas has joined #openstack15:22
*** rossij has joined #openstack15:22
*** flashn has joined #openstack15:22
*** sunech has joined #openstack15:22
*** letterj has joined #openstack15:22
*** smoser has joined #openstack15:22
*** klumpie has joined #openstack15:22
*** _cerberus_ has joined #openstack15:22
*** brainproxy has joined #openstack15:22
*** sets mode: +vv letterj _cerberus_15:22
*** uvirtbot has joined #openstack15:22
*** antonym has joined #openstack15:22
*** lool has joined #openstack15:22
*** jarrod has joined #openstack15:22
*** dovetaildan has joined #openstack15:22
*** gholt has joined #openstack15:22
*** yosh has joined #openstack15:22
*** karmabot has joined #openstack15:22
*** timrc has joined #openstack15:22
*** ioso has joined #openstack15:22
*** sets mode: +v antonym15:22
*** GasbaKid has joined #openstack15:22
*** bcwaldon has joined #openstack15:22
*** ttx has quit IRC15:23
*** ttx has joined #openstack15:23
*** lvaughn has quit IRC15:23
jaypipesso quiet this morning...15:24
*** rnirmal has joined #openstack15:24
*** anticw_ has joined #openstack15:25
*** dsockwell has quit IRC15:25
*** zykes- has quit IRC15:25
*** perestrelka has quit IRC15:25
*** olivier_ has quit IRC15:25
*** mjmac has quit IRC15:25
*** morfeas has quit IRC15:25
*** cw has quit IRC15:25
*** dendrobates has quit IRC15:25
*** PiotrSikora has quit IRC15:25
*** czajkowski has quit IRC15:25
*** [ack] has quit IRC15:25
*** lstoll has quit IRC15:25
*** greg_g has joined #openstack15:25
*** greg_g has joined #openstack15:25
*** eday_ has joined #openstack15:25
*** jeremyb_ has joined #openstack15:25
*** doude_ has joined #openstack15:25
*** freeflyi1g has joined #openstack15:26
*** decramy_ has joined #openstack15:26
*** 13WAA4PTR has joined #openstack15:27
*** dsockwell has joined #openstack15:27
*** zykes- has joined #openstack15:27
*** dendrobates has joined #openstack15:27
*** lstoll has joined #openstack15:27
*** perestrelka has joined #openstack15:27
*** olivier_ has joined #openstack15:27
*** mjmac has joined #openstack15:27
*** morfeas has joined #openstack15:27
*** cw has joined #openstack15:27
*** PiotrSikora has joined #openstack15:27
*** czajkowski has joined #openstack15:27
*** [ack] has joined #openstack15:27
*** sets mode: +v dendrobates15:27
*** jk0 has quit IRC15:29
*** greg-g has quit IRC15:29
*** kim0 has quit IRC15:29
*** doude has quit IRC15:29
*** jeremyb has quit IRC15:29
*** freeflying has quit IRC15:29
*** eday has quit IRC15:30
*** decramy has quit IRC15:30
*** arreyder has quit IRC15:30
*** anticw has quit IRC15:30
*** tsharju has quit IRC15:30
*** jk0 has joined #openstack15:30
*** jk0 has joined #openstack15:30
*** ChanServ sets mode: +v jk015:30
*** tsharju has joined #openstack15:30
*** msassak has joined #openstack15:30
*** mray has joined #openstack15:31
*** arreyder has joined #openstack15:34
*** fluffybunnies has joined #openstack15:36
creihtjaypipes: shhh :)15:36
*** guigui1 has quit IRC15:36
sorencreiht, jaypipes: Can you guys calm down, please? I'm trying to concentrate :)15:37
creihtwedjat: re: docs, can you file a bug? It is a lot easier to track that way, and we are less likely to forget about it :)15:38
creihtwedjat: oh actually now that I read down further, that should work fine15:39
creihtmtaylor: I'm going to bug you about swift packages again :)15:40
*** j05h has quit IRC15:40
*** j05h has joined #openstack15:40
creihtwedjat: when you are available, let me know and I can help you with the auth problems you are running into15:41
*** lvaughn has joined #openstack15:41
*** Hayati_ has joined #openstack15:41
*** rds___ has joined #openstack15:43
*** jeremyb_ has quit IRC15:44
*** jeremyb_ has joined #openstack15:44
*** jeremyb_ is now known as jeremyb15:45
*** rds__ has quit IRC15:48
*** rds___ is now known as rds__15:48
*** littleidea has joined #openstack15:53
*** arreyder has quit IRC15:58
*** arreyder has joined #openstack15:58
*** GasbaKid has quit IRC16:02
*** aimon_ has joined #openstack16:03
*** aimon has quit IRC16:03
*** aimon_ is now known as aimon16:03
*** hub_cap has joined #openstack16:03
*** dragondm has joined #openstack16:05
*** dendrobates is now known as dendro-afk16:05
*** dendro-afk is now known as dendrobates16:06
*** Hayati_ has quit IRC16:06
*** hub_cap has joined #openstack16:07
doude_Hi all, when we use Glance with a Swift backend, how we set the Swift URL proxy ? (I use the Bexar release)16:07
*** sateesh has quit IRC16:07
*** 13WAA4PTR is now known as kim0|Celebrating16:08
*** sebastianstadil has joined #openstack16:09
*** heden has quit IRC16:10
*** aryko has joined #openstack16:11
*** perra has quit IRC16:11
*** befreax has quit IRC16:15
*** fluffybunnies has quit IRC16:16
*** CloudChris has left #openstack16:21
*** vvuksan has joined #openstack16:23
*** aryko has quit IRC16:24
*** aryko has joined #openstack16:25
*** dirakx has joined #openstack16:31
*** rds___ has joined #openstack16:31
jaypipesdoude_: sorry, not quite understanding... is this a Swift question or a Glance question?  There are no configuration options for Swift in Glance.16:32
jaypipescreiht: ^^ ideas on doude_ 's question?16:32
creihtjaypipes: I was figuring it was a glance questions :)16:33
jaypipescreiht: hehe16:33
jaypipescreiht: is there a proxy server in Swift I don't know about?16:33
creihtPerhaps a better question is, how do you hook glance up to swift?16:33
*** kashyapc has quit IRC16:34
jaypipescreiht: it's just a URI... location=swift://user:pass@blah.../path/to/image16:34
creihtahh cool16:34
*** rds__ has quit IRC16:34
*** rds___ is now known as rds__16:34
jaypipescreiht: the Swift store is currently readonly. Meaning the image must already be in Swift.16:34
*** dfg has joined #openstack16:34
sirp_in bexar, swift read-only, so you just craft the URI by hand; in cactus we're gonna add POST, so we'll probably end up having a config param which contains the Swift BASE URI16:34
jaypipescreiht: working on getting Swift store read/write:
uvirtbotLaunchpad bug 713126 in glance "Swift backend does not support POST" [Critical,Confirmed]16:34
jaypipescreiht: if only sirp_ would get that done! :P16:35
* jaypipes hides in a corner16:35
jaypipessirp_: good morning, btw :)16:35
sirp_hehe, morning :)16:36
sirp_working on reply to MySQL Gone Away review...16:36
jaypipessirp_: k16:36
*** metoikos has joined #openstack16:36
wedjatcreiht: hi :)16:37
wedjatcreiht: well, apparently for swauth some steps are missing in the online documentation16:37
wedjatcreiht: no problem, i will file a new bug16:38
creihtdid you get it working?16:38
wedjatyeah, it works now16:38
wedjatbut i don't know which auth system is the best16:38
creihtwell we are going to be getting rid of devauth in cactus16:38
wedjatok, so, will it be possible to link some external auth system to swauth ?16:39
creihtswauth is just a reference implementation of auth for swift that happens to backend the auth info in swift16:39
creihtauth for swift is plugable through middleware16:39
creihtI guess I should specify wsgi middleware16:41
*** fabiand_ has joined #openstack16:41
wedjatso, we should use devauth as a template for creating it, right ?16:41
creihtwedjat: yeah you could use the devauth middleware as a template16:41
wedjatok you answered my question :)16:41
wedjatdo you know where i can find some information about http headers for swauth ?16:42
wedjati found myself all the information for devauth, but actually i still don't understand well how swauth works16:43
*** pvo has joined #openstack16:43
doude_jaypipes: I saw in the Bexar release, we can use Swift backend for Glance to fetch image, that's true ?16:44
doude_jaypipes: but only fetch for the moment, we can't post image actually16:44
jaypipesdoude_: correct.16:44
doude_jaypipes: so my question is, how configure Glance to use backend Swift16:45
doude_jaypipes: how set the Swift proxy URL ?16:45
jaypipesdoude_: I don't know what the Swift proxy URL is?16:45
*** burris has quit IRC16:45
creihtexplains a little how swauth works, but doesn't have any info on the specific requests16:46
*** rnirmal_ has joined #openstack16:46
jaypipesdoude_: to access an image that is in Swift via Glance (and store metadata about that image), x-image-meta-location header should be set to the swift URI.16:46
wedjatthanks creiht :)16:46
doude_jaypipes: when we fetch or post a file in a Swift storage, we must send request to a Swift proxy16:46
creihtIs also a good doc to look at on implementing your own auth16:46
jaypipesdoude_: how are you adding images to Glance? If you use the glance.client.Client class, it's easy enough to work with Swift objects.16:46
creihtIf that doesn't have what you need, let me know and I can see about getting more docs on there16:47
wedjatcreiht: yeah, this last one works quite well in fact :)16:47
wedjatcreiht: a last question if i can : do you know how to delete an account properly with devauth ?16:47
creihtSupport was never added for devauth for deletes16:47
jaypipescreiht: could you assist with me and doude_ ? is this something we need to add support for in Glance? I'm not sure.16:47
*** hub_cap_ has joined #openstack16:48
wedjatcreiht: that's what i thought :(16:48
*** Ryan_Lane|away is now known as Ryan_Lane16:48
doude_jaypipes: yes, but my glance.client.Client must know the URL of the Swift proxy to send the requests, no ?16:48
creihtwedjat: if you want to do it manually, then a DELETE would be need to be issued to the proxy server for /v1/ACCT_HASH16:48
*** metoikos has quit IRC16:48
wedjatcreiht: yeah i can do it manually via curl16:48
creihtthen delete the account row from the /etc/swift/account.db16:48
*** rnirmal has quit IRC16:48
*** rnirmal_ is now known as rnirmal16:48
*** burris has joined #openstack16:49
*** hub_cap has quit IRC16:49
*** hub_cap_ is now known as hub_cap16:49
wedjatcreiht: thanks ! i will try that right now :)16:49
creihtwedjat: If you are running the account reaper, then it will take care of crawling the system and deleting the customer data16:49
doude_jaypipes: no, it's Glance server which must know the Swift URL16:49
*** GasbaKid has joined #openstack16:50
*** kashyapc has joined #openstack16:50
jaypipesdoude_: I know. You pass the Swift URI in the x-image-meta-location header (or add a location key to the image_meta mapping sent to client.Client.add_image()... that's what I'm saying.16:50
*** ccustine has joined #openstack16:51
*** EdwinGrubbs has quit IRC16:52
creihtdoude_: Once you have an image in swift, the URI for it will be something like https://YOURSTORAGEDOMAIN/v1/AUTH_YOUR_AUTH_HASH/images_container/image_goes_here.some_ext16:52
creihtsounds like you send that to glance to link glance to your image16:52
*** hadrian_ has joined #openstack16:53
*** fabiand_ has quit IRC16:53
jaypipesdoude_: IOW, something like this:
jaypipescreiht: yep, exactly. as the 'location' metadata key value...16:54
jaypipesdoude_: does that make more sense now?16:55
doude_jaypipes: ok thanks16:56
jaypipesdoude_: the ability to seamlessly add an image to Glance and store the image directly into Swift is currently being worked on:
uvirtbotLaunchpad bug 713126 in glance "Swift backend does not support POST" [Critical,Confirmed]16:56
*** hadrian has quit IRC16:56
*** hadrian_ is now known as hadrian16:56
jaypipesdoude_: so, in Cactus, there'll be no need to first add the image to Swift and *then* add the metadata to Glance. It'll be just a single call.16:56
doude_jaypipes: Yes, I saw that, but I want to post manualy images to Glance/Swift, and use Nova/Glance to fetch its16:57
*** that_guy has joined #openstack16:57
*** that_guy is now known as Guest1884716:58
creihtdoude_: I think that is what jaypipes is showing you16:58
jaypipesannegentle:  Do you notice the blocks of commands under 1. and 2. are off-kilter?16:58
*** rnirmal has quit IRC16:58
doude_jaypipes: just another question, in this 'location' metadata, the 'user' vlaue contains the 'account' name and the 'user' name or juste the 'user' name ?16:58
*** littleidea has quit IRC16:58
jaypipesdoude_: that's a good question :) I'm not sure. creiht, is there a difference between a "user" and an "account" in Swift?17:00
*** dfg has quit IRC17:00
*** vvuksan has quit IRC17:01
creihtwith the default auth, you send the user:account as the user17:02
creihtwhich may not work correctly with that url structure17:02
creihtnot sure how it works17:03
doude_jaypipes: because, abitualy, when we request Swift we set 'account:user'17:03
creihtyeah that is what I meant :)17:03
Guest18847Hi again, everyone!  so, sqlalchemy is killing me with these DetachedInstanceErrors(preventing euca-describe-instances, etc. from working much of the time)17:04
Guest18847are there any hacks/fixes/etc.?17:04
*** rnirmal has joined #openstack17:04
jaypipesdoude_: one sec, looking at the source code...17:04
doude_jaypipes: ok, thanks17:06
*** maplebed has joined #openstack17:07
jaypipesdoude_: ok, so we construct (in glance) a swift.common.client.Connection() object, passing in user, key, and authurl, with the swift URI interpreted as so: swift:://user:key@authurl/container/object17:08
jaypipesdoude_: and that was taken straight from the swift.common.client source code documentation. So, I'm really not sure. :(17:08
jaypipesdoude_: so, if swift convention is swift://account:user@blah/, then we would be constructing the Connection as user=account and key=user I suppose!17:09
creihtI guess we need to make some of the naming more consistent in docs17:09
jaypipescreiht: sorry, I'm a bit lost as to whether we're doing it wrong or not. :(17:09
creihtThe vaious RS idioms thrown in there don't help17:10
*** cjreyn has quit IRC17:10
creihtto try to clear things up17:10
creihtin swift accounts are opaque strings17:10
*** eday_ is now known as eday17:11
wedjatthere can be several users for one account in swift17:11
creihtWith the reference auth, we add uses by making the account string, account:user17:11
wedjatbut a user is always linked to one account17:11
jaypipescreiht: OK, so how does that fit into standard URI naming? does that mean you get URIs like swift://account:user:pass@url/location?17:13
creihtjaypipes: so when creating the connection, user needs to be the "account:user" string and key still needs to be that users key17:13
jaypipescreiht: and that key is? assumed to be some private key file?17:13
creihtjaypipes: yeah not sure we had ever thought of it in a context like that before17:13
*** kpepple has joined #openstack17:13
jaypipescreiht: I can't pass the key in the URI, huh?17:14
creihtjaypipes: the key is basically the password17:14
jaypipeshmm, well there goes swift in glance :)17:14
creihtjaypipes: can you point me at code in glance that does this?17:14
*** RobertLaptop has joined #openstack17:14
jaypipescreiht: so, looks like we'll need to have a config variable pointing to the key file for the swift account/user?17:15
jaypipescreiht: sure, one sec.17:15
creihtjaypipes: well the key isn't a file, it is just another string that is basically the password17:15
jaypipescreiht: so you really *do* have URIs like swift://account:user:pass@url/location?!17:15
jaypipeshmm, ok.17:16
wedjatno, but this is not as difficult.17:16
jaypipesah, you just pass the password in the client Connection constructor as key kwarg. I understnad.17:16
wedjatyou'll just give the information via http headers17:16
jaypipesah, I see.17:16
jaypipesone sec, grabbing loggerhead url.17:16
openstackhudsonProject swift build #195: SUCCESS in 29 sec:
openstackhudsonTarmac: logging: use routes to separate logging configurations17:16
*** sateesh has joined #openstack17:17
creihtjaypipes: ok looking17:17
jaypipescreiht: the code in there isn't particularly clean ... sorry.17:17
creihtjaypipes: btw, the obligatory, did you guys even try this? should apply :)17:17
jaypipescreiht: this is what happens with Faked stub Swift servers. :(17:18
*** photron has quit IRC17:18
jaypipessirp_: has anyone ever tested the Swift backend properly in relation to Nova's integratoin with Glance, or is Glance always using its filesystem backend for the snapshots in XS?17:19
jaypipescreiht: the Swift code really hasn't been touched much since _0x44 wrote it early on.17:19
creihtjaypipes: no worries... we will get this figured out17:19
jaypipescreiht: yup, I'm not worried at all.17:19
sirp_jaypipes: nope, we just used filesystem backend for snapshots (since the swift adapter didn't yet have PUTs)17:21
_0x44Originally I used the stubbed swift that was included in python-cloudfiles17:21
jaypipessirp_: ah, makes sense.17:21
_0x44It was changed though, to use a different stub17:21
creihtahh.. so in cloud files, there is no reall concept of users with accounts17:22
jaypipes_0x44: no worries, I wasn't blaming you for anything :) and the new stub is the same as the old stub.17:22
*** Hayati_ has joined #openstack17:22
creihtso in that scenario, everything would work as expected17:22
jaypipes_0x44: AFAIK. test/unit/ or something like thta...17:22
_0x44jaypipes: Ah, ok... when I was adding in S3 I must have missed it in the stubs17:22
jaypipes_0x44: no worries mate17:23
doude_jaypipes: I must leave the chat, Can I ask you again on monday for this point ?17:23
jaypipes_0x44: now go back to your earthquakes.17:23
jaypipesdoude_: surely.17:23
doude_jaypipes: thanks17:23
* _0x44 goes back to trying to shake the earth.17:23
_0x44jaypipes: I'm trying to do it with my brain.17:25
jaypipescreiht: so... looks like we need to add a configuration variable for setting the password for the swift user?17:27
*** MarcMorata has joined #openstack17:27
creihtjaypipes: thinking....17:27
jaypipescreiht: not too hard. earthquakes, remember.17:27
creihtjaypipes: parsed_uri is an object returned from urlparse?17:28
*** joearnold has joined #openstack17:28
*** dprince has quit IRC17:29
*** dprince has joined #openstack17:30
jaypipescreiht: yes17:30
JordanRinkeso are there some bexar release packages? how/where do I point to grab them?17:31
openstackhudsonProject swift build #196: SUCCESS in 28 sec:
openstackhudsonTarmac: Changes to support deployment on IPv617:31
creihtjaypipes: so the quick fix is to change line 10217:32
creiht user, key = creds.rsplit(':',1)17:32
creihtthat should still give you the right creds17:32
creihtand the uri could be represented as:17:33
jaypipescreiht: yup, that's what I was thinking too...17:34
justinsbOops... hi!  Could I get a review of please?17:34
creihtlong term I'm not sure if that is how we want it17:34
jaypipesjustinsb: never! :)17:35
jaypipescreiht: me neither, but it's a fix for now..17:35
justinsbjaypipes: It was starting to feel that way :-)17:35
jaypipesjustinsb: :) unfortunately, I can't review much more than style concerns on that one. just have no experience in it. :(17:36
jaypipesjustinsb: but I'll do my best :)17:36
justinsbjaypipes: Well, if there are any style concerns, let's start the ball rolling!17:36
jaypipesjustinsb: will do17:36
creihtjustinsb: you send me an HP SAN, I'll review it for you :)17:36
justinsbcreiht: I know!  I was borrowing a Rackspace test SAN over a VPN.  Not the fastest configuration :-)17:37
jaypipesok, I've officially listened to the entire Collective Soul album now three times in a row... time for some new music...17:37
sirp_"Where the river flows.... <bam...bannnanana>"17:39
jaypipessirp_: :)17:40
*** vishvananda has joined #openstack17:40
*** reldan has joined #openstack17:40
*** nelson has quit IRC17:43
justinsbHas any work been done on metadata?  I rely on this blueprint for passing capability requests as metadata, but I don't see any code for it yet:
*** Nacx has quit IRC17:48
*** hub_cap_ has joined #openstack17:48
*** syah has quit IRC17:48
*** rdw has quit IRC17:48
*** vishvananda has quit IRC17:50
*** rdw has joined #openstack17:50
*** johnpur has joined #openstack17:50
*** ChanServ sets mode: +v johnpur17:50
*** freeflyi1g has quit IRC17:50
*** syah_ has joined #openstack17:50
*** Hayati_ has quit IRC17:51
*** rnirmal has quit IRC17:51
*** hub_cap has quit IRC17:51
*** hub_cap_ is now known as hub_cap17:51
*** freeflying has joined #openstack17:51
*** littleidea has joined #openstack17:51
*** dprince has quit IRC17:51
*** dprince has joined #openstack17:52
*** iRTermite has quit IRC17:54
*** iRTermite has joined #openstack17:55
justinsbdprince: Hi!  I think the blueprint I'm talking about is yours.  Do you know if there's any implementation work on it?
openstackhudsonProject nova build #503: SUCCESS in 1 min 31 sec:
openstackhudsonTarmac: Create a new AMQP connection by default.17:57
*** littleidea has quit IRC17:58
*** BK_man has quit IRC17:59
dprincejustinsb. Hi. We were planning on working on the API peices of that. No implementations yet. I think we need the metadata-svc implementation first however.17:59
justinsbWhat's the metadata service?  I don't really understand the metadata-svc blueprint.18:01
wedjatcan swift work with python3 ? :p18:02
wedjatshould be interesting to test it18:02
dprincejustinsb: Yeah. Not a lot of information on the metadata-svc blueprint.18:03
dprinceI just created an etherpad for openstack-api-metadata here:18:03
*** vvuksan has joined #openstack18:03
dprinceIf you want to gather information there that would be great.18:03
westmaasjustinsb: I believe the metadatasvc is intended to be the code that actually reads from and writes to a persistent store.  the API piece is just exposing that via REST18:04
*** Nick_ has joined #openstack18:06
*** Nick_ is now known as Guest8075218:06
westmaashub_cap: does that seem like a reasonable explanation?  We should sync up at some point.18:06
*** sateesh has quit IRC18:08
*** MarkAtwood has joined #openstack18:09
justinsbwestmaas, dprince: Thanks.  Why is the metadata service different from the normal API?  In other words, why isn't the instance metadata stored in the same database as the primary instance data?18:14
*** ke4qqq has quit IRC18:14
*** ke4qqq has joined #openstack18:14
*** GasbaKid has quit IRC18:14
*** Guest80752 is now known as Sateesh18:15
*** GasbaKid has joined #openstack18:19
jk0vishy: kpepple got it sorted out last night, thanks for looking at it with us18:19
*** rds__ has quit IRC18:22
dprincejustinsb: I think the idea was that the API would make use of an underlying metadata service.18:23
justinsbdprince: So the metadata service will replace the current MySQL/SQLite store?  (It might make use of a RDBMS under-the-covers, of course)18:25
dprincejustinsb: where that data is ultimately stored wouldn't matter to the end user. It could use the nova compute DB or it could also use its own metadata storage DB.18:26
dprincejustinsb: sure18:26
justinsbdprince: I guess I want to know how to implement instance metadata - it feels like that could just be put in another table in the DB in the current architecture.  When the metadata service comes along, it can refactor that into a separate service if needed.  But it doesn't feel like a hard dependency...18:27
*** arcane has joined #openstack18:28
*** rds__ has joined #openstack18:28
*** pvo has quit IRC18:30
dprincejustinsb: That seems like a reasonable approach I guess. To be honest I'm not entirely sure what the grand plan for the metadata-svc is in terms of storage.18:32
*** CloudChris has joined #openstack18:33
*** kashyapc has quit IRC18:34
*** dubsquared has joined #openstack18:34
justinsbdprince: OK, thanks!  I'll code up a branch and submit and see how it goes, I guess!18:34
dprincejustinsb: If you have a concrete idea of how to implement you might comment on the etherpad and then we can push the Blueprint for approval.18:36
vishyjk0: cool, what was it?18:39
jk0I think it was the way he was loading the tests18:39
jk0it was a very odd/one off problem18:39
vishycool, well glad it is fixed18:39
vishyis Salvatore Orlando on here?18:40
*** dprince has quit IRC18:40
comstudit sounds like we want to keep the guest agent a separate project?18:43
kpepplejustinsb: i saw your email this morning "[Openstack] Allowing clients to pass capability requests through tags?" ... can you look at and make sure you take this into account ?18:43
justinsbdprince: Good idea - done!18:43
*** aryko has quit IRC18:44
devcamcarhey gang, I have a suggestion about migrations in nova18:44
*** aryko has joined #openstack18:44
*** aryko has joined #openstack18:44
devcamcarright now it seems like we are trying to maintain giant per-release migrations18:45
devcamcarreally we should have an individual migration for each individual piece18:45
kpeppledevcamcar: +1 ... this is more the rails style and lets us rollback individual features ... if people write intelligent rollbacks18:46
devcamcarthis allows for migrations to work against trunk in non official releases18:46
kpeppledevcamcar: but this does mean that we will need to coordinate on numbering / naming18:46
justinsbkpepple: Sure.  Just looked at it.  Again, complementary proposals I think.  Some of the capability requests can indeed be put into the instance type (e.g. GPU, no GPU)  Some won't make sense (requesting location near another item)   I would suggest that instance types could also be extensible using the same metadata approach, so that you don't need to pre-define all the dimensions of instances that you support (e.g. GPU, or SS18:46
justinsbD, or Quantum Offload Engine)18:46
devcamcaryea, right now it's like 3_cactus, where it should be like 3_cactus_feature name18:46
devcamcarkpepple: yes but thats really just a merge issue18:47
kpeppledevcamcar: will that preserve migration order ? RoR went to a timestamp_name nomenclature to support distributed teams18:47
*** rcc has quit IRC18:47
devcamcarnot sure with sqlalchemy18:47
justinsbdevcamcar: It is bugging me that we're going to have those merge issues, but I wonder if a single migration is more 'solid' for production18:48
*** mdomsch has joined #openstack18:48
kpeppledevcamcar: that can become a build / debug nightmare ... and migrate is already ... tricky18:48
justinsbdevcamcar: So are we trading off developer joy vs production pain?18:48
devcamcarjustinsb: how are individual migrations a build or debug nightmare?18:49
devcamcaras long as we test migrating from official release to official release, i don't see a problem18:49
justinsbdevcamcar: Agreed, I'm not clear on that.  A single migration should run in one big transaction, so either succeeds or fails.  We risk having weird intermediate states if we go outside a transaction18:50
justinsbdevcamcar: though MySQL doesn't support DDL transactions anyway18:50
hub_capwestmaas: was afk for lunch, reading now18:51
justinsbdevcamcar: I'm definitely fine with trying it out.  If it becomes a problem in future we can bunch them into one migration script as part of the release process.18:51
devcamcarsqlalchemy doesn't supporting test runs? migrating without committing?18:51
devcamcarseems like a pretty glaring hole if it doesnt18:52
jaypipesdevcamcar: ++ on rails-style migrations. that's what we're doing in glance.18:52
devcamcari'll add a blueprint for it18:52
hub_capwestmaas: yes the metadata svc is exposed via a api, and is a persistent store for data that is intended to be used by a client, such as agent-smith. Currently ozone is using XenStore to do this via a file. XS has a way to mount as a filesystem (its a hierarchical store), and the agent is polling that "file" for changes.18:53
hub_capits a one way write tho, to the api itself. the client never writes to the api, just writes to its filesystem18:53
jaypipesjustinsb: ya, that's a good point about mysql not supporting DDL transactions...can't really do anything about that :(18:54
*** jc_smith has joined #openstack18:56
*** GasbaKid has quit IRC18:56
hub_capwestmaas: just a fyi, even though the metadata svc is assigned to me, i doubt ill get around to working on it. its not something we need ATM. So if you or anyoen else wants to take it on, feel free. The torch was passed to me when i thought it was a 2 way communication between guest/host, but its _def_ not that18:57
*** sebastianstadil has quit IRC19:01
*** nijaba has quit IRC19:04
jc_smithwith regard to opstack API in Nova, when I run curl -D - -H "X-Auth-Key: admin" -H "X-Auth-User: admin" http://localhost:8774/v1.0/   , I get back a header with X-Server-Management-Url: http://localhost:8774/v1.0/ , in the RS docs for the API there is a number after the /v1.0/ part of the URI which I think corresponds to an account ID. I don't see this in my X-Server-Management-Url, anybody seen this before?19:06
*** nijaba has joined #openstack19:07
*** nijaba has joined #openstack19:07
*** dendrobates is now known as dendro-afk19:09
*** dendro-afk is now known as dendrobates19:09
*** morfeas has quit IRC19:10
*** BK_man has joined #openstack19:11
*** morfeas has joined #openstack19:11
*** BK_man has quit IRC19:11
westmaashub_cap: got it.  We will see if we get to it as we try to bring the API up to spec.  Thanks for the info.19:11
hub_capwestmaas: np! ping me if u have any other Qs or would like some cycles19:12
*** kpepple has quit IRC19:12
*** rnirmal has joined #openstack19:13
*** reldan has quit IRC19:13
*** GasbaKid has joined #openstack19:14
*** adiantum has quit IRC19:17
jaypipesfigure this will make antonym, rackerhacker and JordanRinke cringe:,-The-Server-Room-Switch,-Woody,-and-More.aspx#Pic319:17
jc_smithah, I guess OpenStack API doesn't need an ID in the uri, nm19:18
rackerhackerjarrod: i feel nauseous19:19
rackerhackeroof, that was meant for jaypipes19:19
* rackerhacker tab-failed19:19
dubsquaredlol that second to last picture is all kinds of win19:19
*** burris has quit IRC19:21
*** westmaas has quit IRC19:21
*** msassak_ has joined #openstack19:21
*** h0cin has quit IRC19:22
*** msassak_ has quit IRC19:22
annegentlejc_smith: Yep, that number you see before the /v1.0/ is a Rackspace Cloud account ID19:22
*** msassak_ has joined #openstack19:22
*** msassak has quit IRC19:22
annegentlejc_smith: yep that's for our particular implementation, you do what you like with OpenStack for auth/accounts/billing/etc19:22
jc_smithannegentle: thanks19:23
annegentlejaypipes: omg that even makes me a little queasy19:23
*** ke4qqq has quit IRC19:26
*** ke4qqq has joined #openstack19:27
jaypipesdubsquared: right on... I love the little inset in the commenter's picture, too :)19:32
*** zul has quit IRC19:34
daboWorking on nova/compute/ today, I noticed that several methods use **_kwargs instead of the more common **kwargs. Is there a reason for the version with the underscore? Or just a personal habit?19:35
*** kpepple has joined #openstack19:46
sandywalshdabo, don't ask jaypipes ;)19:48
sandywalshoh, it's ok, he's had his coffee by now19:48
daboHeh, I ignore him anyway. :)19:48
jaypipesdabo: not sure, but I think that convention stems from pylint, where an error is thrown if an argument isn't actually used in the function. Could be wrong, though.19:48
*** clauden_ has joined #openstack19:48
jaypipesdabo: and the underscore marks it "ignored" or "unused" in pylint I think.19:49
dabojaypipes: yeccch19:49
jaypipesdabo: take it up with whoever wrote the code, not me :)19:49
*** omidhdl has joined #openstack19:49
daboThe tool is dumb, so let's make the code ugly to accommodate the tool19:50
dabojaypipes: yeah, I know it wasn't you (bzr blame absolves you)19:50
*** westmaas has joined #openstack19:51
creihtwestmaas: re: python 3, not likely as not all of its dependencies work with python 3 yet19:54
*** joearnold has quit IRC19:56
westmaascreiht: think you meant someone else :)19:56
*** jc_smith has quit IRC19:56
creihtwedjat: -^19:57
creihtwe have too many people in this channel :)19:57
*** GasbaKid has quit IRC19:58
*** brd_from_italy has joined #openstack20:02
*** reldan has joined #openstack20:05
*** dprince has joined #openstack20:06
*** gaveen has quit IRC20:07
*** lvaughn has quit IRC20:09
*** jc_smith has joined #openstack20:12
*** msassak_ has quit IRC20:14
*** joearnold has joined #openstack20:17
*** sandywalsh has quit IRC20:18
*** sandywalsh has joined #openstack20:18
*** omidhdl has quit IRC20:19
*** gaveen has joined #openstack20:20
hub_capSo my cohorts and myself have a openvz driver issue. Openvz needs a integer value for unique container id. We _could_ use the id in instance_ref, but its not exactly future proof. It could, at any time, change to a uuid. Also, the auto_increment wont recycle ids (until it reaches is max val for a int).20:29
hub_capthe ovz docs state using a integer representation of the IP address is a good way to assign the unique id20:30
hub_capbuuut, the address object (which is set up before the spawn call) in the compute manager20:30
hub_capdoesnt pass down the address information in the instance_ref20:30
*** fabiand_ has joined #openstack20:31
hub_capoops wrong window :P20:31
hub_capcompute/ line 19520:31
imsplitbithub_cap: lol20:31
hub_capis anyone opposed to adding the ip to the instance table / ref object for this case?20:31
imsplitbitI am not opposed at all and frankly I think it is a great idea20:34
hub_capimsplitbit: of course u do, u are writing the ovz driver20:34
imsplitbitdamn, that matters?20:34
imsplitbitand I might add that I find that the guy writing the ovz driver is stunningly handsome20:35
*** fabiand_ has quit IRC20:35
imsplitbitin all seriousness tho, can anyone find a reason not to push this value down to the driver via the instance object?20:35
hub_capi think everyone went to lunch, or to a super secret room20:36
*** CloudChris has quit IRC20:36
edayjaypipes, creiht: re the swift user/account/key URL discussion earlier, I would probably not put as part of the URL. that format for http:// implies the standard Authorization header. if you *do* want to still use it, you can always just escape 'account:user' so : is '%3A' (standard URL encoding rules)20:37
*** troytoman-away is now known as troytoman20:37
*** troytoma_ has joined #openstack20:37
*** sandywalsh has quit IRC20:39
hub_capimsplitbit: ping the mailing list20:40
*** sandywalsh has joined #openstack20:41
justinsbhub_cap: Can't you get a list of the in-use instance-ids on the compute node and just pick one that isn't in-use?20:41
hub_capthe ip is already allocated at this point20:41
hub_capto the instance20:42
hub_capits just not sent down to the virt driver20:42
jaypipessirp_: let's agree to discuss the whole config option processing stuff later. I approved the patch to get the fix for the bug in. we can always discuss The Right Way (if there even is one ;) ) laterz. :)20:42
*** lvaughn has joined #openstack20:43
*** Ryan_Lane has quit IRC20:44
justinsbhub_cap, imsplitbit: Does a vzid need to be globally unique if we're going to support migration?20:45
hub_capjustinsb: i wouldnt say globally, across DCs, or endpoints20:45
hub_capbut within a Zone, yes20:45
hub_capif we do DC -> DC migrations, we will probably have apis calls for that20:45
justinsbhub_cap: But can't be done with only per-host knowledge20:46
*** msassak has joined #openstack20:46
hub_capthe ipspace will be unique across the hosts in a dc, no20:46
sirp_jaypipes: cool20:46
justinsbhub_cap: You can't ask OpenVZ on one host which vzids are in use and just pick an unused one, because two hosts might pick the same vzid and then migration between those nodes would be broken20:46
sirp_jaypipes: though one way to make this problem go away is to just rip out command-line options and require config file; the code would get dramatically simpler20:47
justinsbhub_cap: So iSCSI and I think the address code uses pools of resources; which is one way - you'd need a pool of vzids20:47
justinsbhub_cap: Otherwise you're going to have a hard time squeezing an IPv6 address into 32 bits :-)20:47
hub_capjustinsb: thats why we want to use the IP address as a int value20:47
hub_caphehe good call on ipv620:47
hub_capwe wanted a solution that wasnt overly complicated. a pool of avail vz ids and a service to dole them out is overly complex for the now solution, but is def where we want to go for future20:48
hub_capbased on teh convo's imsplitbit and i had over skype20:48
*** Ryan_Lane has joined #openstack20:49
*** Pentheus has joined #openstack20:53
*** reldan has quit IRC20:55
*** reldan has joined #openstack20:57
imsplitbityeah, definitely something that needs to be done, just not now21:02
*** hadrian has quit IRC21:02
hub_capyup, FTW21:03
*** GasbaKid has joined #openstack21:05
*** DigitalFlux has joined #openstack21:06
*** GasbaKid has quit IRC21:06
annegentlelooks like a couple of nova.conf flags are bye-bye. If I still have --cc_host in a nova.conf file does it still get used, or do the --ec2_blah settings take over?21:07
DigitalFluxHi Guys21:07
DigitalFluxI was wondering, is the instances IDs totally unique ?21:08
DigitalFluxeven across different OpenStack projects ?21:08
*** GasbaKid has joined #openstack21:08
annegentlesorry I mean --ec2_url seems to be gone... if I still have --ec2_url in a nova.conf file, will nova use it?21:10
hub_capDigitalFlux: if its in the same DB, it should be, the ID is a sequence21:10
*** GasbaKid has quit IRC21:11
vvuksanDigitalFlux: they are not unique21:11
hub_capvvuksan: crap we were relying on this, how are they not unique given a single db?21:11
vvuksanthey are unique21:11
vvuksanbut in sequence21:11
vvuksannot world unique21:12
hub_capu scared the crap out of me21:12
hub_capsure unique to the DB instance itself, no?21:12
vvuksansorry about that21:12
*** GasbaKid has joined #openstack21:12
*** dendrobates is now known as dendro-afk21:14
DigitalFluxvvuksan: hub_cap: your final decision ? unique or not unique ? :-)21:14
vvuksanit's unique21:14
hub_capdo i have to put all 1 million dollars on it?21:14
DigitalFluxThey are not world unique but project unique, correct ?21:14
hub_capThey are a sequence in the DB in a single table21:15
hub_capso as long as you have a single DB21:15
hub_capthey are unique21:15
vvuksani think they are DB unique21:15
DigitalFluxhub_cap: a single table for all projects ?21:15
*** GasbaKid has quit IRC21:15
edayDigitalFlux: unique if they share a DB, not glonally unique. I tried making them globally unique via uuids, but that pach died a controversial death. perhaps some day others will see the value of unique IDs across clusters21:15
edayerr, globally21:15
DigitalFluxeday: i already see the value21:16
hub_capeday: i would share that view of the need of global uniqueness, although we are hackishly using that value for our openvz driver21:16
hub_capDigitalFlux: id = Column(Integer, primary_key=True, autoincrement=True) <-- unique to the database u use, a single table (instances) within that table21:16
*** GasbaKid has joined #openstack21:16
*** dirakx has quit IRC21:18
DigitalFluxOK, thanks guys21:18
*** blueadept has quit IRC21:19
hub_capDigitalFlux: np21:19
*** GasbaKid has quit IRC21:19
*** GasbaKid has joined #openstack21:20
*** GasbaKid has quit IRC21:25
Ryan_LaneI really don't understand nova-manage service list21:25
Ryan_LaneI swear my services are always down. never up21:25
Ryan_Laneand I have no idea why21:25
*** GasbaKid has joined #openstack21:26
kpeppleRyan_Lane: you think they are up but service list shows them down or they are down and you don't know why ?21:27
Ryan_Lanethey are up21:28
Ryan_Laneand it shows them down21:28
vvuksanRyan_Lane: for me it shows half the info21:28
kpepplewhat does you services db table look like ?21:29
*** GasbaKid has quit IRC21:29
DigitalFluxAnybody played with python-novatools (cloudservers) with nova Bexar ?21:29
DigitalFluxI simply exported the vars mentioned in novarc21:29
DigitalFluxand ran cloudservers list against it21:29
DigitalFluxGetting ""No project with id openstack (HTTP 500)"21:30
*** dendro-afk is now known as dendrobates21:30
Ryan_Lanekpepple: am I supposed to be looking at the disabled field?21:30
*** GasbaKid has joined #openstack21:30
Ryan_Lanebecause disabled show 0 on all of them21:31
Ryan_Laneso it's just mis-reporting?21:31
kpeppleyou on trunk ?21:31
*** grapex1 has joined #openstack21:31
Ryan_LaneI should really switch to the bexar release21:31
Ryan_Lanein fact, let me do that21:32
DigitalFluxSo the cloudservers just assumes that the project_id of openstack is "openstack"21:32
DigitalFluxthis is not configurable ?21:32
Ryan_Lanegah, what's the release ppa link?21:32
*** burris has joined #openstack21:33
Ryan_Lanethanks :)21:33
*** GasbaKid has quit IRC21:33
kpepplei think that enabled is set of which is a straight boolean test21:34
DigitalFluxRyan_Lane: You're welcome :)21:34
*** GasbaKid has joined #openstack21:35
*** Pentheus has quit IRC21:35
*** elimisteve has left #openstack21:36
DigitalFluxIt is true !21:36
*** maplebed has quit IRC21:36
DigitalFluxthe cloudservers command just assumes that the project name is called "openstack"21:36
*** Pentheus has joined #openstack21:36
*** hggdh has quit IRC21:37
*** maplebed has joined #openstack21:37
*** GasbaKid has quit IRC21:37
*** kbringard has joined #openstack21:37
annegentleDigitalFlux: You can change that in nova.conf with the --default_project flag21:38
mrayWhere can I find the nova-import-canonical-imagestore referenced in
*** GasbaKid has joined #openstack21:39
dragondmDigitalFLux: ya, currently it just defaults to the 1 project name.  That will be changing (hopefully for cactus).21:40
*** hggdh has joined #openstack21:41
*** jc_smith has quit IRC21:41
annegentlemray: hm. it installed when I installed Bexar from the tarball...21:41
DigitalFluxannegentle: dragondm: Thanks21:41
mrayannegentle: I can't find it in the apt repo21:41
*** lvaughn has quit IRC21:41
DigitalFluxAny idea where can i get a RackSpace test image (template) to upload to Openstack ?21:41
*** dprince has quit IRC21:41
*** GasbaKid has quit IRC21:42
*** lvaughn has joined #openstack21:42
annegentlemray: sorry I'm not much good at unravelling packaging.21:43
mrayapt-cache search reveals nothing21:43
*** GasbaKid has joined #openstack21:43
mrayhmm… maybe it's not used anymore21:44
annegentlemtaylor: any ideas? or soren? ^^21:44
*** GasbaKid has quit IRC21:45
*** ctennis has quit IRC21:46
*** jc_smith has joined #openstack21:46
Ryan_Lanebexar release shows most of my services down too21:46
vvuksanDigitalFlux: what template ?21:47
kpeppleRyan_Lane: are the services updating ?21:49
Ryan_Lanelooks like it21:50
*** DigitalFlux has quit IRC21:50
vvuksankpepple: they update for me as well21:51
vvuksanbut some just don't show21:51
kpepplehold on ... need to reboot my server to bring it to trunk21:51
Ryan_Lanemaybe I need to drop my database and re-create it21:52
Ryan_Lanewho knows if the schema is right or not21:52
jaypipesjustinsb: k, review of your hpsan branch done.21:52
jaypipesjustinsb: as much as I could, of course :)21:53
mtaylorannegentle: I blame jaypipes21:54
jaypipeswhat did I do now?21:54
mtaylorjaypipes: probably nothing21:55
creihtmtaylor: !!!!21:56
creihtmtaylor: packages? :)21:56
creihtis lp slow for anyone else right now?21:56
creihtI'm only getting 110KB/s from them right now both at home and my slice21:56
mtaylorcreiht: packages? have I dropped the ball on a packaging task I was to be working on?21:57
Ryan_Lanewell, a fresh db didn't help21:57
Ryan_Lanethough it's good to know I'm starting fresh :)21:57
jaypipescreiht: yet another LP maintenance window starting. They are trying to upgrade the database and LP software. It failed the last time.21:57
jaypipescreiht: I know.21:57
Ryan_Laneheh. poor lp people. I know their pain :)21:57
jaypipescreiht: lovely having LP maintenance windows in US work times...21:57
* creiht doesn't see any 1.2.0 packages :)21:59
creihtmtaylor: there were also 2 ppas added with nothing in them21:59
creihtnot sure what they are for, or if we intend on changing ppas?22:00
* mtaylor probably just had a brain hemorrage or something... lemme figure out what I was thinking22:01
annegentlemtaylor: no head 'splosions! :)22:04
* mtaylor makes a popping noise22:04
* mtaylor deflates22:04
*** j05h has quit IRC22:04
*** ctennis has joined #openstack22:05
uvirtbotNew bug: #717431 in glance "Swift backend not taking into account new Swift account:user:key in URI" [High,Confirmed]
sorenmtaylor: You never responded to my question about changing the e-mail in the ppa script back to mine.22:07
mtaylorsoren: uh, sorry. I don't remember the question... :)22:07
mtaylorsoren: oh right, I think I remember that now22:07
mtaylorwhy would you want the script to be doing things with your name?22:08
soren21:11 <+soren> mtaylor: Would you mind terribly if I changed it back? It's the only way I can think of that I can get notified about build/upload failures.22:08
*** j05h has joined #openstack22:08
mtaylorsoren: sorry - missed that. can we try to figure out another way to get you email failure notifications?22:08
sorenDoes it hurt anything that it's my addresss?22:08
sorenWhat was the motivation for changing it?22:09
mtaylorsoren: my pedantic side ... you are not actually the one building those packages, and you are not personally signing them22:09
mtaylorsoren: hudson is building and signing them, so having the entry be soren hansen is not reflective of reality22:09
sorenmtaylor: There's some truth to that for sure.22:09
mtaylorof course - it doesn't _actually_ hurt things, it just doesn't seem 'right'22:09
*** hub_cap has quit IRC22:13
sorenmtaylor: Apart from the notification thing, I also do like that people installing the packages can go and see in the changelog and find a real person they can e-mail if it's failing for them somehow.22:13
sorenmtaylor: Being from a PPA and all, lots of people know that the Maintainer: field can be somewhat misleading.22:14
mtaylorsoren: well, we could fill out the "Uploaders" field appropriately22:14
mtaylorsoren: but totally your call. at some point I'd love to find a 'good' solution to these problems (and also would like to again register my dislike for the motu-in-maintainer field thing)22:15
kbringardif I upload a raw image that has the kernel and ramdisk bundled inside it into glance, can I launch it with nova?22:16
tr3buchetrecent pull of trunk, getting:
tr3buchetany trouble shoot ideas?22:18
*** dprince has joined #openstack22:20
kpeppletr3buchet: it's in servers, something about getting your key pairs22:21
sirp_tr3buchet: did you use euca to add-keypair?22:22
kpeppletr3buchet: the authmanager is saying you don't have any22:23
tr3bucheti haven't changed anything since it worked earlier22:23
tr3buchetthat's the odd thing22:23
*** allsystemsarego has quit IRC22:23
sirp_tr3buchet: are you pointing at the sqlite db now instead of your mysqldb?22:23
*** westmaas has quit IRC22:24
*** dprince has quit IRC22:26
tr3buchetnot sure if this helps, but i was trying to do a cloudservers boot when it happened22:27
jaypipesgholt: hey, great series of articles on your blog about consistent hashing. well done.22:27
tr3buchetimage-list flavor-list and list all work just fine22:27
kpeppletr3buchet: probably have to go into your db and do a 'select * from key_pairs;' to see if your key pairs are really there22:31
kbringardthis feels like a dumb question, but how do I get nova to use glance to look for available images?22:31
tr3buchet"Empty set (0.00 sec)" hmmm how did that happen22:32
kpeppletr3buchet: that would be bad22:32
tr3buchetkbringard: you have to add the images to glance22:32
kbringardI did that22:32
tr3buchetkbringard: are you using hte cloudservers tool?22:32
kbringardI see the image when I do a get on images and a head shows the image attributes22:32
kbringardah, no, I was using the euca2ools22:33
kbringardI guess that would explain it, eh?22:33
tr3buchetdo you get anything out of euca-describe-images?22:33
kbringardyea, but only the images I've uploaded with the euca2ools22:33
tr3buchetkpepple: any idea how i managed to not have any keys in there?22:33
kbringardevery time that's happened to me, it's been because I created them with the wrong configuration22:34
tr3buchetkbringuard: i've only ever used images i've specificly uploaded22:34
kbringardhmm, well, I have not restarted the API and stuff since I got glance running, maybe that's it22:35
tr3buchetweird, i haven't changed any configurations, only pulled trunk22:35
kbringarddid you do a nova-manage db sync, maybe it blasted your settings?22:35
kbringardis everything else in there still?22:35
*** kpepple has quit IRC22:36
*** vvuksan has quit IRC22:37
antonymtr3buchet: last time i got that, i had to run this step, making sure that euca2ools was installed euca-add-keypair nova_key > nova_key.priv22:37
*** hggdh has quit IRC22:38
tr3bucheti did that though22:38
tr3buchetand again22:38
tr3buchettore down the whole db and recreated, recreated the project, unzipped, then added the keypair22:38
*** dragondm has quit IRC22:39
sirp_tr3buchet: you've verified that you can see the key-pair in the db?22:39
*** hggdh has joined #openstack22:39
tr3buchetkey_pair is not in the DB22:39
*** dirakx has joined #openstack22:39
sirp_tr3buchet: ah then the keypair add step was borked for some reason22:40
*** hggdh has quit IRC22:40
*** m_3 has quit IRC22:40
sirp_is it in nova.sqlite db in the rootdir?22:40
*** hggdh has joined #openstack22:40
tr3bucheti'm using mysql22:43
tr3buchetomg, tell me that isn't22:43
tr3bucheti've added something to to bin-manage for create networks and forgot to comment it out for trunk22:43
tr3buchetthanks sirp_ antonym kpepple22:44
kbringardso, the aws api doesn't use glance?22:45
sirp_kbringard: yeah, right now we're focuing on getting it to work with OpenStack API22:47
*** ppetraki has quit IRC22:47
kbringardok, cool, that's good to know22:47
kbringardso if I've uploaded an image into glance, I should be able to see if if I use the OpenStack API22:47
sirp_kbringard: yep22:48
kbringardawesome, thanks sirp_, one last question... are there any command line tools akin to the euca2ools for the openstack API?22:49
sirp_kbringard: contains some of the steps we're using to register images and spin them up22:49
sirp_kbringard: yeah we're using novatools aka 'cloudservers'22:49
kbringardperfect, I was crawling the wiki but hadn't gotten there yet22:49
*** westmaas has joined #openstack22:55
Ryan_Laneugh. for some reason with bexar release, nova-network isn't setting up the SNATTING table22:56
*** jtran has joined #openstack23:00
Ryan_Lanewhy isn't it doing init_host?23:01
*** mdomsch has quit IRC23:02
kbringardhmm, OK, so when I do "cloudservers image-list" I still only see the images I uploaded with the euca2ools23:02
kbringardbut when I get /images on my glance API host, I see the image I uploaded23:03
kbringardwhen compute, scheduler and network start, I see it setting glance_host correctly23:03
*** troytoman is now known as troytoman-away23:04
*** troytoma_ has quit IRC23:04
*** brd_from_italy has quit IRC23:04
Ryan_Lanewell, I found a bug :(23:06
Ryan_Lanewhen FlatDHCP starts, it calls super before init_host23:06
Ryan_Lanethe init_host method in super tries to assign the floating IPs23:06
Ryan_Lanewhich doesn't work so well when the firewall chains don't exist23:07
Ryan_Lanewhich causes the network manager to fail to start23:07
* Ryan_Lane goes to file a bug23:07
*** lvaughn_ has joined #openstack23:09
*** lvaughn has quit IRC23:11
*** nelson has joined #openstack23:11
*** reldan has quit IRC23:12
nelsonDoes the AUTH_hex key account ever time out?23:12
kbringardah, I figured it out23:12
kbringard<--- idiot23:12
nelsonI was just trying to do a demo of Swift for Danese, and started getting access denied (401) errors.23:13
Ryan_Laneso, lp717469 is a pretty nasty bug in bexar23:16
openstackRyan_Lane: Error: "717469" is not a valid command.23:16
*** johnpur has quit IRC23:16
Ryan_Laneanyway, there's no way to make nova-network start when this occurs, unless you want to disassociate all floating IPs23:17
creihtnelson: Is it possible the token expired?23:17
*** reldan has joined #openstack23:17
Ryan_Laneor comment out code, restart, uncomment the code, and restart23:17
nelsonIt's a public container, but I'm guessing that the account times out anyway.23:17
Ryan_Laneor know all of the rules that need to be created23:17
creihtnelson: are you by chance sending creds even though it is a public container?23:18
creihtif it is a public container, then it shouldn't 40123:18
nelsonNo, not sending creds, just the AUTH_(hex string) account name.23:18
nelsonthat was the problem. Had to fetch new certs.23:19
nelsonI'm still running 1.1; if it's fixed in 1.2, well, too bad for me.23:20
uvirtbotNew bug: #717467 in nova "migrate versioning import exceptions fails" [Undecided,New]
PentheusDoes 1.2 have working public containers?23:23
creihtPentheus: yes23:23
Pentheuscreiht: It uses a different path then RS' .make_public() though still?23:24
Pentheuscreiht: since theirs is really for cdn23:24
Pentheuscreiht: the r: .* foo23:24
Pentheuscreiht: thanks :)23:24
uvirtbotNew bug: #717469 in nova "Network manager fails to start after system restart if floating IP addresses are in use" [Undecided,New]
kbringardthanks everyone, have a fab weekend23:26
*** kbringard has quit IRC23:26
*** gondoi has quit IRC23:27
*** Guest18847 has quit IRC23:29
iammartianI just bought $19000 worth of servers I could not afford, to add "cloud based computing" to my hosting company offering.  Could openstack fill that role it its current state?23:30
iammartianself service by customers is key for me23:31
*** ccustine has quit IRC23:31
*** ryker has joined #openstack23:32
iammartiani've been using citrix xenserver up until now23:32
*** grapex1 has left #openstack23:32
*** ccustine has joined #openstack23:35
rykerHi all.  Does anyone have any suggestions for setting up a fresh openstack install for testing purposes using chef?  I found nova-solo, but i'm not sure how up to date it is.
openstackhudsonProject nova build #504: SUCCESS in 1 min 33 sec:
openstackhudsonTarmac: This fixes a lazy-load issue in describe-instances, which causes a crash.  The solution is to specifically load the network table when retrieving an instance.23:37
JordanRinkeanyone used availability zones?23:38
JordanRinkeryker: pop a tweet to @mattray23:39
JordanRinkehe is doing a ton of openstack chef stuff23:39
rykerJordanRinke: thx23:39
westmaasryker: dovetaildan (aka dprince) also does a bunch of chef stuff based on work vishy did23:40
JordanRinke(mattray = opscode dude, working with vishy also)23:40
JordanRinkeryker: I just talked to him, he said email him and he will hit you up23:42
JordanRinkehe was just shutting down or he would hop in IRC23:42
vishyryker: chef stuff with vagrant is all on ansolabs blog, putting in stuff for multinode at the moment23:42
*** sebastianstadil has joined #openstack23:43
rykerJordanRinke and westmaas:  thx, i'll email him.23:43
rykervishy: I haven't tried using vagrant yet.23:43
rykeras a matter of fact, I'm not even sure what vagrant is. :)  I keep hearing about it, but i haven't looked into it yet.23:44
*** mray has quit IRC23:45
jtrananyone here familiar w/ the openstack-compute gem?23:48
*** kpepple has joined #openstack23:53

Generated by 2.14.0 by Marius Gedminas - find it at!