Wednesday, 2011-04-20

*** kyzh has joined #openstack00:02
*** kyzh is now known as Guest1329900:02
*** Guest13299 is now known as kyzh_00:02
*** taihen has joined #openstack00:03
*** kyzh_ has quit IRC00:03
ShermanBoydKnuckleSangwich: they opened the summit to anyone who shows up00:04
ShermanBoydKnuckleSangwich: you just can't eat the food and go to the official parties00:05
*** adiantum has joined #openstack00:06
*** Zangetsue_ has joined #openstack00:08
*** Zangetsue has quit IRC00:10
*** BK_man has quit IRC00:13
*** clauden has quit IRC00:14
*** BK_man has joined #openstack00:14
*** nelson has quit IRC00:18
*** nelson has joined #openstack00:18
*** dendrobates is now known as dendro-afk00:21
*** BK_man has quit IRC00:26
*** adiantum has quit IRC00:28
*** dweimer has quit IRC00:29
*** joearnold has quit IRC00:29
*** ewindisch has quit IRC00:33
*** adiantum has joined #openstack00:33
*** adjohn has joined #openstack00:34
uvirtbotNew bug: #766697 in nova "Instance on the pure compute node cannot reach the meta data server" [Undecided,New]
*** nijaba has joined #openstack00:37
*** mlanner has quit IRC00:37
uvirtbotNew bug: #766703 in nova "Need documentation for nova-manage image ... commands " [Undecided,New]
*** dragondm has quit IRC00:50
*** miclorb_ has quit IRC00:51
*** adiantum has quit IRC00:51
*** adiantum has joined #openstack00:51
*** adjohn has quit IRC00:52
*** adjohn has joined #openstack00:52
*** j05h has quit IRC00:56
*** deepy has quit IRC00:59
*** rnirmal has joined #openstack01:00
*** BK_man has joined #openstack01:00
*** miclorb has joined #openstack01:02
*** j05h has joined #openstack01:02
*** vvuksan has joined #openstack01:05
*** deepy has joined #openstack01:06
kpeppleannegentle: regarding 766703 ... isn't that documented on the man page (doc/source/man/novamanage.rst) ? or are you asking for more detail ?01:06
*** ewindisch has joined #openstack01:12
*** BK_man has quit IRC01:15
*** adiantum has quit IRC01:18
*** ovidwu_ has joined #openstack01:21
*** ovidwu_ has quit IRC01:21
*** benbenhappy has joined #openstack01:23
*** adiantum has joined #openstack01:24
*** ovidwu_ has joined #openstack01:24
*** ovidwu_ has quit IRC01:26
*** littleidea has quit IRC01:32
*** miclorb has quit IRC01:35
winston-dkpepple: hi01:35
kpepplewinston-d: hey01:35
*** Ryan_Lane has quit IRC01:36
winston-di was wondering if you may answer my question:
* kpepple looks at launchpad01:37
*** maplebed has quit IRC01:37
*** ryker has quit IRC01:40
kpepplewinston-d: can't use just use a different libvirt template with the libvirt_xml_template flag ?01:41
*** ryker has joined #openstack01:41
kpepplewinston-d: or do you want each image (or groups of images) to use different libvirt templates ?01:42
*** littleidea has joined #openstack01:43
*** AimanA has quit IRC01:43
winston-dkpepple : yeah, something like different image(s) to use different libvirt templates, or a super-flexible template that can adapt to different images.01:44
*** miclorb has joined #openstack01:45
kpepplewinston-d: (not being a glance expert) it seems like you could store a libvirt_template_type metadata in glance registry and then have use that when it starts a vm ... however, you'd need some mechanism to distribute the templates to all the nodes (maybe store them in the glance registry and retrieve them ?)01:47
*** AimanA has joined #openstack01:48
*** adiantum has quit IRC01:48
winston-dkpepple : well, didn't notice that glance supports that kind of metadata 'cos last time I tried glance, it had some problem with euca2ools.  But thanks for the hint! I'll try glance with Cactus release.01:49
*** vvuksan has quit IRC01:49
kpepplewinston-d: it'll take some hacking in still ... by default, it's just using the libvirt_xml_template flag01:50
winston-dkpepple : by the way, I feel that my need should be common.  Or, I got something wrong or had wrong assumption?01:50
*** ksteward has quit IRC01:50
kpepplewinston-d: no, i think people will want more and more types of images supported01:50
winston-dkpepple : then this should be done in Nova or Glance, right?  I mean adding such feature seems quite straight-forward if people do need more images.01:52
*** Zangetsue has joined #openstack01:53
*** adiantum has joined #openstack01:53
winston-dkpepple: I'd check with jaypipes if this feature is desireable, I can contribute by providing a prototype patch.01:56
*** santhosh has joined #openstack02:00
winston-dkpepple : and for now, are all images being used the same in terms of image partition? (i.e. image is partitioned, not LVM;  and the '/' is always located at 1st partition)02:03
kpepplewinston-d: sorry stepped away cooking dinner :)02:04
kpepplewinston-d: all are the same for libvirt ... other drivers (vmware, etc.) probably use other schemes02:04
*** santhosh has quit IRC02:04
kpepplewinston-d: most of this will need to be done in nova02:04
*** ttx has quit IRC02:05
*** dendro-afk has quit IRC02:05
winston-dkpepple: is it too late for dinner? :)02:05
*** ttx has joined #openstack02:05
*** ttx has quit IRC02:05
*** ttx has joined #openstack02:05
*** dendro-afk has joined #openstack02:05
*** aliguori has quit IRC02:05
*** azneita has joined #openstack02:05
*** aliguori has joined #openstack02:07
kpepplewinston-d: no, but you'll have to hurry over :)02:10
*** adiantum has quit IRC02:11
*** dweimer has joined #openstack02:12
winston-dkpepple : For custom images, I think it's better to add extra attributes while bundling/uploading images to glance, then Nova can utilize that extra attribute to form a correct libvirt XML.  Meantime, libvirt xml tempate requires some minor modification to accommadate information like location of root partition.02:13
*** adiantum has joined #openstack02:16
*** jfluhmann_ has joined #openstack02:16
*** jeffjapan has quit IRC02:17
*** KnuckleSangwich has quit IRC02:18
*** jeffjapan has joined #openstack02:20
*** matiu has quit IRC02:21
*** littleidea has quit IRC02:25
*** Ryan_Lane has joined #openstack02:25
*** ChameleonSys has quit IRC02:34
*** ChameleonSys has joined #openstack02:35
*** ChameleonSys has quit IRC02:35
*** ChameleonSys has joined #openstack02:36
*** adiantum has quit IRC02:39
*** ovidwu_ has joined #openstack02:39
*** ovidwu_ has quit IRC02:41
*** ovidwu_ has joined #openstack02:44
*** cascone__ has quit IRC02:44
*** adiantum has joined #openstack02:46
*** mdomsch has joined #openstack02:54
*** cascone__ has joined #openstack02:56
*** Ryan_Lane has quit IRC02:57
thickskinsomebody tell me which is the best way to install nova(multiple server) in one physical node?????02:59
*** aimon has quit IRC03:00
*** aimon has joined #openstack03:00
*** jkoelker has joined #openstack03:02
*** jkoelker has quit IRC03:04
*** littleidea has joined #openstack03:05
winston-dthickskin : depends on what Linux distribution u r using03:10
thickskinwinston-d : ubuntu10.4 LTS03:11
winston-dthickskin : then you should follow the doc here :
thickskinI`ve done several times before.03:14
thickskinwinston-d : But I can`t accesed instance.03:15
winston-dthickskin : what's the output of 'euca-describe-instance' ?03:15
thickskinwait a second please.03:15
thickskinI`ll check the log.03:16
*** littleidea has quit IRC03:17
thickskinIt seems like runnig successfully, but instance start with 192.168.30.***.03:20
thickskin192.168.30.*** is physical node`s network segement.03:22
thickskinI don`t know why instance is running with 192.168.30.***03:22
HugoKuo_check your database03:25
HugoKuo_take a look at TABLE : fixed_ip and networks03:26
HugoKuo_you'll know why :>03:26
thickskinHugoKuo : nice to meet you again.03:26
thickskinI know that.03:26
HugoKuo_so what's your ip list in fixed_ip  ?03:27
thickskinI want to know why instance is running with 192.168.30.***03:27
thickskinyou mean $B"*!!(BDesired network + CIDR for project (normally x.x.x.x/24)03:28
thickskinI`ve setup
HugoKuo_ok let's do it manual03:28
HugoKuo_plz empty your fixed_ip table03:28
HugoKuo_and network table ~03:29
thickskinNow, my computer is already empty.03:29
thickskinubunt10.4 was installed.03:29
HugoKuo_then #nova-manage network create 1 25603:30
thickskinI`ve installed ubuntu 10.4 LTS 5minutes before.03:30
thickskincould you tell me please step by step.03:31
thickskinNow my physical node is CLEAN!!03:31
winston-dthickskin : your instance got 192.168 something because you (the script?) configure it to do so03:31
HugoKuo_Let's take time to do that03:32
*** adiantum has quit IRC03:32
openstackjenkinsProject swift build #250: SUCCESS in 27 sec:
openstackjenkinsTarmac: allowed_headers are returned from a GET but not HEAD03:32
HugoKuo_first   #apt-get update03:32
winston-dthickskin : by 'CLEAN' you mean there's only OS, Nova hasn't been installed yet?03:32
thickskinwinston-d : that`s right03:32
HugoKuo_There's a fresh ubuntu I think03:32
thickskinHugoKuo: the first step is apt-get update ?03:33
HugoKuo_yes do that03:33
winston-dthickskin : our nice guy Hugo will walk u through the installation process.  but i think you should install Nova uself.  We can help with your configuration.03:33
HugoKuo_and then open CC-install script , try to rewrite something03:34
thickskinthanks guys!!!03:34
*** adiantum has joined #openstack03:34
winston-dthickskin : this time, i prefer you install Nova manually.  'cos the installation script is mixed with some configuration too.03:35
HugoKuo_I afraid nova.conf will drive him crazy :>03:36
thickskinI`ve already chech the url. But it dosen`t work03:36
HugoKuo_thickskin : open that script and jump to line 52203:37
winston-dHugoKuo_ : how's Cactus working for you?03:37
thickskinIs the script is <<<wget --no-check-certificate>>> right?03:37
HugoKuo_winston-d : dude it's working now03:37
thickskinand next is03:38
HugoKuo_thickskin : yes03:38
HugoKuo_jump to line 52203:38
HugoKuo_replace FlatManager to FlatDHCPmanager03:38
HugoKuo_how many NIC do you have in your host03:39
-thickskin- 513 --sql_connection=mysql://root:$MYSQL_PASS@$MYSQL_HOST_IP/nova03:39
-thickskin- 514
-thickskin- 515 NOVA_CONF_EOF03:39
-thickskin- 51603:39
-thickskin- 517 echo "Initializing database"03:39
-thickskin- 518 nova-manage db sync &>> $LOGFILE03:39
-thickskin- 519 sleep 103:39
-thickskin- 520 echo "...done..."03:39
-thickskin- 521 echo03:39
-thickskin- 52203:39
thickskinyou mean 514 line??03:39
HugoKuo_yes , your 514  ...... paste on pastebin next time :)03:39
thickskinI`ve done.03:40
HugoKuo_add line  13     17   18  19  20  21  into the script03:42
HugoKuo_and change parameter to your need :>03:42
thickskinok. wait a second please.03:43
HugoKuo_this my single machine configure ..... but I combine instance network with my corporate network , so that my fixed_range is 172.16.203:44
HugoKuo_your flat_network_dhcp_start is
*** BK_man has joined #openstack03:45
thickskinand the next step is........03:46
HugoKuo_run the script03:47
openstackjenkinsProject swift build #251: SUCCESS in 27 sec:
openstackjenkinsTarmac: Allow for double quarantining of dbs.  Just appends a uuid if there's naming conflict.03:47
HugoKuo_becareful of your input parameter03:47
thickskinOK, I`ll try again.03:47
HugoKuo_I'll help you03:47
thickskinthank you.03:47
HugoKuo_step by step03:47
HugoKuo_I paste something wait03:48
*** miclorb has quit IRC03:49
*** miclorb has joined #openstack03:50
*** miclorb has joined #openstack03:52
HugoKuo_what's your host IP03:52
HugoKuo_is ?03:52
thickskinget from DHCP server03:53
HugoKuo_admin name ?03:53
HugoKuo_give me a project user name  and project name03:54
thickskinuser name is novaadmin and project name is novaproject03:54
thickskincat /etc/network/interfaces03:55
thickskinauto lo03:55
thickskiniface lo inet loopback03:56
*** BK_man has quit IRC03:56
thickskinauto eth003:56
thickskiniface eth0 inet dhcp03:56
thickskinis default.03:56
*** hadrian has quit IRC03:56
thickskinnot static03:56
vishymy guess is something else is giving out dhcp addresses03:56
HugoKuo_how many interface do you have03:56
*** localhost3 has joined #openstack03:57
thickskinjust one.03:57
thickskinbut i cat add one more.03:57
HugoKuo_check my paste, follow the parameter03:57
thickskinBut the install script is still running...03:58
HugoKuo_thickskin : vishy is expert  in network :)03:58
HugoKuo_thickskin : what do you mean still running ?03:58
thickskinI hope your help.03:58
*** cascone__ has quit IRC03:59
thickskinInstall script is done.03:59
*** adiantum has quit IRC03:59
HugoKuo_start up an instance03:59
*** ewindisch_ has joined #openstack04:02
*** dendro-a` has joined #openstack04:02
*** dendro-afk has quit IRC04:03
*** ewindisch has quit IRC04:03
*** localhost2 has quit IRC04:03
*** ewindisch_ is now known as ewindisch04:03
HugoKuo_go for lunch04:04
*** adiantum has joined #openstack04:05
thickskinI`ll be back 1hour later. Is it ok?04:05
*** AimanA is now known as HouseAway04:07
*** azneita has left #openstack04:09
*** rchavik has joined #openstack04:14
*** cascone__ has joined #openstack04:16
*** littleidea has joined #openstack04:17
*** mdomsch has quit IRC04:21
*** kashyap has joined #openstack04:23
*** Ryan_Lane has joined #openstack04:30
*** koolhead17 has left #openstack04:47
thickskinHugo kuo : I`m ready for start up an instance.04:50
*** adiantum has quit IRC04:53
*** f4m8_ is now known as f4m804:55
*** soosfarm has quit IRC04:57
*** soosfarm has joined #openstack04:58
*** photron_ has joined #openstack04:58
*** adiantum has joined #openstack05:00
*** Ryan_Lane has quit IRC05:01
*** Ryan_Lane has joined #openstack05:03
HugoKuo_does it work05:05
thickskinI`m glad to see you again soon.05:05
thickskinCould you tell me please the next step is what for start up an instance.05:08
HugoKuo_check image ami05:10
HugoKuo_and then use euca-run-instance05:10
HugoKuo_how colud you start up instance before @@?05:11 is done and I didn`t nothing after that.05:13
HugoKuo_type command05:14
HugoKuo_check all service status05:14
HugoKuo_is there any image registered to objectstore ?05:15
thickskinyou mean >>
*** ovidwu_ has quit IRC05:15
*** ramkrsna has joined #openstack05:15
*** ramkrsna has joined #openstack05:15
thickskinNow I don`t have any image.05:16
HugoKuo_excute line 3~5 now05:16
thickskin. /root/creds/novarc is not nessesary ??05:17
*** adiantum has quit IRC05:18
HugoKuo_oh sorry , excute that first05:18
thickskincommand . /root/creds/novars is in but it seems like doesn`t worked.05:19
*** photron_ has quit IRC05:19
*** zaitcev has quit IRC05:19
HugoKuo_euca-describe-images to check image05:19
*** littleidea has quit IRC05:20
thickskinresult is
HugoKuo_you already got image05:22
*** littleidea has joined #openstack05:22
HugoKuo_next step is followe line 7~1405:22
*** adiantum has joined #openstack05:24
*** zenmatt has quit IRC05:31
*** kashyap has quit IRC05:37
*** adiantum has quit IRC05:45
*** guigui1 has joined #openstack05:49
*** adiantum has joined #openstack05:50
*** aimon_ has joined #openstack05:55
*** aimon_ has quit IRC05:58
*** aimon has quit IRC05:58
*** aimon has joined #openstack05:59
*** aimon_ has joined #openstack06:05
*** jeffjapan has quit IRC06:06
koolhead11hi all06:08
*** aimon has quit IRC06:08
*** aimon_ is now known as aimon06:08
*** mgoldmann has joined #openstack06:11
*** kashyap has joined #openstack06:13
thickskinIs there anyone know that my nova instance is always get IP from master DHCP SERVER.06:13
*** jeffjapan has joined #openstack06:15
*** aimon has quit IRC06:17
*** aimon has joined #openstack06:18
*** adiantum has quit IRC06:26
*** ovidwu_ has joined #openstack06:31
*** ovidwu_ has quit IRC06:32
*** rsaidan has joined #openstack06:35
*** rchavik is now known as CakeBot06:37
*** daveiw has joined #openstack06:44
*** JulioBell has joined #openstack06:45
*** littleidea has quit IRC06:46
*** gaveen has joined #openstack06:52
*** rcc has joined #openstack07:11
*** lionel has quit IRC07:11
*** lionel has joined #openstack07:12
*** miclorb has quit IRC07:14
*** daveiw has quit IRC07:18
*** antenagora has joined #openstack07:19
*** BK_man has joined #openstack07:25
*** nacx has joined #openstack07:35
*** Ryan_Lane has quit IRC07:39
*** BK_man has quit IRC07:42
*** BK_man has joined #openstack07:42
openstackjenkinsProject burrow build #5: SUCCESS in 14 sec:
openstackjenkinsTarmac: Pylint cleanup, reorganized code, added interactive shell.07:47
*** allsystemsarego has joined #openstack07:50
*** atul_ has joined #openstack08:08
HugoKuo_anybody here , I got a question problem with glance.......while I upload image08:08
HugoKuo_socket.error: [Errno 111] Connection refused08:08
HugoKuo_Checking bucket: glancefailed to upload kernel08:09
*** koolhead11 has quit IRC08:11
*** lee__ has joined #openstack08:14
*** daveiw has joined #openstack08:16
HugoKuo_nova-objectstore service need alive or glance will failed to upload image ....08:17
*** anticw has quit IRC08:20
*** uksysadmin has joined #openstack08:22
*** fabiand__ has joined #openstack08:25
*** anticw has joined #openstack08:27
*** CloudChris has joined #openstack08:28
*** CloudChris has left #openstack08:28
*** BK_man_ has joined #openstack08:28
winston-dHugoKuo_ : you can config glance to use swift if you prefer08:28
*** aimon has quit IRC08:28
HugoKuo_that's my next step08:29
HugoKuo_winston-d : are you using SWIFT now?08:29
*** BK_man has quit IRC08:29
*** BK_man_ is now known as BK_man08:29
HugoKuo_winston-d : I tried SWIFT before 3 months , It's hard to install08:30
winston-dHugoKuo_ : I do have a swift setup but for image service, no08:30
winston-dHugoKuo_ : to me SWIFT is 1^100,00,00,00 times easier than Nova08:30
HugoKuo_how did you do that ?08:31
HugoKuo_So what's your SWIFT for ,  not for image store ?08:32
*** atul__ has joined #openstack08:40
*** ramkrsna has quit IRC08:41
*** adiantum has joined #openstack08:41
*** ramkrsna has joined #openstack08:42
*** ewindisch has quit IRC08:42
*** atul__ is now known as koolhead1108:43
rsaidanHey all, does anyone have nova-volume running? I have a question regarding the –iscsi_ip_prefix option in nova.conf08:44
HugoKuo_set to your nova-volume host08:44
*** atul_ has quit IRC08:44
benbenhappyhi HugoKuo_,I installed nova manual ,no use the script ,after running /etc/init.d/networking restart, the network is disconnect , how to connect to the net?08:45
HugoKuo_Check you iptables08:46
HugoKuo_benbenhappy : or you can flush all rules to verify the problem is caused by 's iptable08:46
*** ovidwu has quit IRC08:47
*** ovidwu has joined #openstack08:47
benbenhappyflush all rules ?08:47
rsaidanhey HugoKuo_ , iscsi_ip_prefix is supposed to be an ip prefix? like 192.168.08:47
*** BK_man has quit IRC08:47
*** BK_man has joined #openstack08:47
HugoKuo_rsaidan : yes man,08:47
*** kashyap has quit IRC08:47
HugoKuo_rsaidan : mine is
rsaidanmy problem is if i put that flag into nova.conf with a double - then all nova-services fail when I restart them... If I use one dash it seems to work but nova-manage doesn't?08:48
rsaidanHugoKuo_: so –iscsi_ip_prefix must be a full ip and not just a prefix.. will try that08:49
rsaidanHugoKuo_: I tried putting -–iscsi_ip_prefix= but that causes all nova services to fail too?08:53
rsaidanApr 20 10:51:12 hera init: nova-scheduler main process (31495) terminated with status 108:53
rsaidanApr 20 10:51:12 hera init: nova-scheduler respawning too fast, stopped08:53
rsaidanApr 20 10:51:12 hera init: nova-network main process (31502) terminated with status 108:53
rsaidanApr 20 10:51:12 hera init: nova-network respawning too fast, stopped08:53
rsaidanApr 20 10:51:12 hera init: nova-compute main process (31510) terminated with status 108:53
rsaidanApr 20 10:51:12 hera init: nova-compute respawning too fast, stopped08:53
rsaidanApr 20 10:51:12 hera init: nova-volume main process (31515) terminated with status 108:53
rsaidanApr 20 10:51:12 hera init: nova-volume respawning too fast, stopped08:53
*** gaveen has quit IRC08:53
HugoKuo_how about log file ?08:53
HugoKuo_did you have --volume_group=  flag in your nova.conf?08:54
*** adiantum has quit IRC08:54
rsaidanwhat you see above is what I get in the logs.. no, what is that flag for?08:54
rsaidanis that for the lvm vol group?08:55
rsaidanit defaults to nova-volumes doesn't it?08:55
HugoKuo_for vg08:57
HugoKuo_which you created for nova-volume08:57
HugoKuo_so do you  have a VG named "nova-volume "?08:58
rsaidanI have one called nova-volumes which I understand is what nova-volume looks for08:59
rsaidancan you paste your nova.conf file for me to see?08:59
HugoKuo_ok wait09:00
*** BK_man has quit IRC09:02
*** flopflip has joined #openstack09:02
*** BK_man has joined #openstack09:03
*** guigui4 has joined #openstack09:03
doudeHi all, I use Nova with Glance (files backend). I started an instance from a glance template. I deleted the template in glance repository and I try to snapshot the instance. When compute make the snapshot, it ask glance to get informations of the template used to start the instance. But in my case, the template doesn't yet exist. So the snapshot fails. Do you think its a bug ?09:04
*** mmaudet has joined #openstack09:04
*** metoikos has joined #openstack09:05
*** flopflip_ has quit IRC09:05
*** ovidwu_ has joined #openstack09:05
*** guigui1 has quit IRC09:06
*** gaveen has joined #openstack09:06
rsaidanHugoKuo_: thank you :)09:07
*** maat has joined #openstack09:07
rsaidanHugoKuo_: it's working, thanks for your help09:09
rsaidanHugoKuo_: again :)09:09
*** daveiw has quit IRC09:09
*** jeffjapan has quit IRC09:10
HugoKuo_rsaidan : cheers09:11
*** ovidwu_ has quit IRC09:12
HugoKuo_Well ,  Success to use Glance for image service , but instance failed to fetch metadata , due to "500 Internal Error"   is this a BUG?09:13
*** BK_man has quit IRC09:15
*** BK_man has joined #openstack09:16
uvirtbotNew bug: #766957 in glance "adding/removing image  problem with wrong url" [Undecided,New]
*** ovidwu_ has joined #openstack09:16
*** ovidwu_ has quit IRC09:16
*** kashyap has joined #openstack09:21
*** BK_man has quit IRC09:21
*** BK_man has joined #openstack09:24
*** nelson has quit IRC09:28
*** ramkrsna has quit IRC09:28
*** nelson has joined #openstack09:29
*** daveiw has joined #openstack09:43
*** lurkaboo is now known as purpaboo09:46
*** BK_man has quit IRC09:48
*** BK_man has joined #openstack09:50
*** guigui1 has joined #openstack10:00
*** kirshil has joined #openstack10:02
*** benbenhappy has left #openstack10:04
*** jfluhmann_ has quit IRC10:08
doudeI try to make a snapshot on Nova (bzr1003) with libvirt/KVM, but compute node fails with error "Error: 'architecture'". I use Glance (bzr116) with file backend. What the mistake ? I tried to add the properties 'architecture' to image but it changes nothing.10:11
*** jfluhmann_ has joined #openstack10:21
*** BK_man_ has joined #openstack10:22
*** BK_man has quit IRC10:22
*** maat has quit IRC10:30
*** mmaudet has quit IRC10:42
*** nagyz has joined #openstack10:47
nagyzcan anyone help me in setting up iSCSI? according to /proc/net/iet/volume, the volume is exported, but for some reason I can't discover it from an other machine10:48
nagyz(so nova-compute can't attach it..(10:48
uvirtbotNew bug: #767027 in glance "Update one property erase others" [Undecided,New]
*** ewindisch has joined #openstack10:59
*** ewindisch has quit IRC11:01
*** miclorb_ has joined #openstack11:06
*** kyzh has joined #openstack11:12
*** kyzh has joined #openstack11:12
*** antenagora has quit IRC11:14
*** onlany has quit IRC11:23
*** adiantum has joined #openstack11:32
*** ctennis has quit IRC11:32
*** mihgen_ has joined #openstack11:44
*** DigitalFlux_ is now known as DigitalFlux11:48
alekibangonagyz: if you solve problems like this, please describe it somewhere (wiki?)  so others can learn11:50
*** DigitalFlux has quit IRC11:50
*** DigitalFlux has joined #openstack11:50
*** ctennis has joined #openstack11:51
*** ctennis has joined #openstack11:51
*** adjohn has quit IRC11:58
mihgen_Have anyone tried to install Nova on Fedora from rpm packages provided for RHEL6? It seems to be packages needs python2.6, but I have python2.7 on Fedora.11:58
nagyzalekibango, the problem was that open-iscsi wasn't installed on the initiator.12:02
*** vvuksan has joined #openstack12:02
*** adiantum has quit IRC12:03
*** shentonfreude has quit IRC12:08
*** adiantum has joined #openstack12:11
alekibangoah :) still might be valuable for someone :)12:12
*** miclorb__ has joined #openstack12:18
nagyzif I'd like to run more than one compute instance, which directories do I need to share between them?12:19
nagyzis there a doc somewhere?12:19
*** jfluhmann_ has quit IRC12:19
dabonagyz: compute instances don't share directories; they share queues and a database if they are in the same zone.12:21
*** miclorb_ has quit IRC12:21
nagyzok, but what about images?12:21
nagyzalso, is there an API to query which instances are run on a specific compute node?12:21
uvirtbotNew bug: #767103 in nova "CreateVolume API response doesn't match AWS EC2" [Undecided,New]
nagyzall I see is compute_api.get_all()12:22
alekibangonagyz: basically yuo need shared db (mysql) and shared image registry (glance)12:22
daboimages are retrieved via the image service, which is usually glance12:22
nagyzah, okay. the shared db is already running12:22
nagyzI'll look into glance later then12:22
nagyzwhat about the API?12:22
alekibangowhat is about api?12:22
alekibangonagyz: glance is used via cmdline and via nova directly12:23
nagyzalso, is there an API to query which instances are run on a specific compute node, instead of querying all nodes by compute_api.get_all()?12:23
nagyzinside nova.12:23
nagyzI'm doing some API extension12:23
alekibangonagyz: i cant tell about nova api much... try looking into python-novaclient  -- that might help12:25
alekibango(as it uses that api)12:25
nagyznovaclient is the external API, I'm playing around with the internal one12:26
nagyzbut I guess I can just filter the resulsts12:26
nagyzso no biggie12:26
*** cascone__ has quit IRC12:34
*** joloughlin has joined #openstack12:36
joloughlinhi, since catcus upgrade my additional compute nodes wont show up as enabled12:37
joloughlinare there any know problems?12:39
*** miclorb__ has quit IRC12:40
*** mgoldmann has quit IRC12:45
*** mgoldmann has joined #openstack12:46
*** gaveen has quit IRC12:46
*** shentonfreude has joined #openstack12:57
*** hggdh has quit IRC12:58
*** hadrian has joined #openstack13:02
*** j05h has quit IRC13:02
*** hggdh has joined #openstack13:02
*** j05h has joined #openstack13:02
*** mray has joined #openstack13:05
*** fabiokung has joined #openstack13:05
*** nwww has quit IRC13:07
*** adiantum has quit IRC13:08
*** lborda has joined #openstack13:09
*** Zangetsue has quit IRC13:09
*** Daviey has quit IRC13:13
*** dendro-a` is now known as dendro-afk13:23
*** daveiw has left #openstack13:23
joloughlinhi, after cactus+glance install my compute nodes no longer show as being unavailable13:24
joloughlinany ideas appreciated13:24
*** adiantum has joined #openstack13:25
*** deshantm_laptop has joined #openstack13:29
joloughlinsorry, i mean my compute nodes are no longer enabled (nova-manage service list)13:30
*** kashyap has quit IRC13:30
*** photron has joined #openstack13:33
*** Daviey has joined #openstack13:35
*** dscastro has joined #openstack13:38
*** kakella has joined #openstack13:39
*** kyzh has quit IRC13:39
*** kakella has left #openstack13:39
*** kyzh has joined #openstack13:40
*** kyzh is now known as Guest9202113:40
*** donkdonk has joined #openstack13:40
*** j05h has quit IRC13:40
*** donkdonk has left #openstack13:40
*** f4m8 is now known as f4m8_13:41
*** Guest92021 has left #openstack13:41
*** j05h has joined #openstack13:41
*** dspano has joined #openstack13:44
*** Zangetsue has joined #openstack13:45
uvirtbotNew bug: #767185 in nova "additional compute nodes not enabled" [Undecided,New]
*** gondoi has quit IRC13:46
*** mdomsch has joined #openstack13:53
*** dendro-afk is now known as dendrobates13:55
*** dendrobates has joined #openstack13:55
*** ChanServ sets mode: +v dendrobates13:55
*** zenmatt has joined #openstack13:57
nhm_woo, I finally got multi-node flatdhcpmanager working.  yay14:00
*** antenagora has joined #openstack14:01
* nhm_ graduates from novice to guppy.14:01
nagyzare there plans somewhere to integrate HA volumes?14:02
*** rnirmal has joined #openstack14:02
*** santhosh has joined #openstack14:08
*** Zangetsue has left #openstack14:10
*** Zangetsue has joined #openstack14:10
uvirtbotNew bug: #767202 in nova "returned in create server password is not set to root" [Undecided,New]
ccookehmm. is there any more up-to-date documentation for getting (citrix) Xen to work ?14:11
uvirtbotNew bug: #767203 in glance "Update 'Name' attribute doesn't work" [Undecided,New]
ccookethe documentation on the wiki seems... old14:11
ccooke(and following it doesn't work :-)14:12
nhm_ccooke: yeah, I've got to assume it's primarily because development has been so fast. :)14:12
ccookeThat would be good :-)14:13
ccookebut the lack of any documentation that would allow it to be tested is a bit of a problem14:13
nhm_ccooke: I just struggled for a couple of weeks to get multi-node compute working and just finally did last night.14:13
ccookeheh. I'm waiting on spare hardware to test that14:14
ccookebut I've been given a xen pool to try to integrate, so...14:14
nhm_ccooke: big lesson I learned is that openstack doesn't really clean up after itself very well yet, so if you change a lot of things you'll probably have some left-over cruft in places that causes exceptions.14:14
ccookefound some isses there, yes14:15
ccookeI'm itching to submit  but need to prove that OS is viable before anyone in the legal department here would sign the requisite documents...14:16
ccookesubmit patches, that should be14:16
blamarccooke:  I take it you're using which you're having issues with?14:17
nhm_ccooke: luckily I work in academia so things are a bit easier here on that front.14:17
ccookenhm_: ah, yes :-)14:17
ccookeblamar: correct14:17
*** adiantum has quit IRC14:18
elasticdogwhich direction is the community moving toward for multi-node installations...puppet, chef, crowbar (if dell actually releases it), nova-deploy, scripts?14:21
elasticdogit seems like everything is at different stages with different ideas on how things should be architected14:22
nhm_elasticdog: I'm using puppet14:22
nhm_elasticdog: I'm using ubuntu hosts with a master node running dhcp3, tftd, and puppet.  compute nodes network boot, and install via kickstart, then puppet takes over, installs all packages and necessary pieces for nova.14:23
elasticdognhm_: that seems like a good just seems like the community is split and moving in all sorts of directions14:24
nhm_elasticdog: we are split and moving all sorts of directions even internally to our organization, and not just for openstack. ;)14:25
ccookeUse packages with sensible defautls and avahi ;-)14:25
*** vvuksan has quit IRC14:25
ccooke(security issues with that model, of course)14:26
*** adjohn has joined #openstack14:27
nhm_elasticdog: I think ultimately it will be important that openstack can integrate well into whatever solution organizations are already using, be that cfengine, chef, puppet, etc.14:28
*** CakeBot is now known as rchavik14:29
elasticdognhm_: good point14:31
*** Ryan_Lane has joined #openstack14:31
*** adiantum has joined #openstack14:32
*** cole has joined #openstack14:32
*** onlany has joined #openstack14:35
nhm_Are there any centos images around that work out of the box?14:36
*** naehring has joined #openstack14:36
*** joloughlin has quit IRC14:37
vishyjoloughlin: do you mean they show as "disabled" or as "XXX"14:38
vishydisabled can be fixed by nova-manage service enable14:38
*** kbringard has joined #openstack14:38
vishyXXX is usually time skew, use ntpdate and probably install ntp14:38
vishynhm_: I have one but it has one bug that stops it from booting14:39
vishyI haven't had time to fix it and upload a new one14:39
vishythe error is just fstab needs to be edited14:39
kbringardmaybe we should make ntpdate a dep in the ppa repos and add an ntpdate script to /etc/cron.daily/ as well14:39
kbringardseems like that happens a lot14:40
nhm_vishy: cool, thanks.14:40
vishynhm_: if you download that image and mount ami-centos/image loopback you need to change vda1 and vda2 in etc/fstab to vda and vdb respectively14:40
nhm_vishy: ok.  I'll probably give it a try.  Want to demo my multi-node install to some people here now that it's working.14:41
vishy(You can convert the image to the new format by sticking the three a*i-centos folders in /some/dir14:42
vishyand using nova-manage image convert /some/dir )14:42
nagyzvishy, is there any reason to use the default encoding when talking thru rabbitmq?14:42
vishygl! I will eventually get around to changing the fstab and reuploading14:42
nagyzinstead of using pickle, for example14:42
blamarvishy: Smoke tests need to run on a single node install maybe?14:42
vishyblamar: no they work multinode too14:43
blamarhmm, k, 100% sure problem on my end :)14:43
vishyblamar: they need to be run on network host though or you have to authorize port 22 and icmp14:43
ccookeso... has anyone been able to make the xen support work?14:43
nhm_vishy: excellent.  I'll let you know if there are any problems.  Seems pretty straightforward.14:43
kbringardif you're looking for a CentOS image you can run on OpenStack I've had decent luck with this one14:44
kbringardI uploaded it with it's associated kernel and ramdisk, then did yum upgrades on it, and snapshotted the running instance with the qemu-img stuff14:45
*** lborda has quit IRC14:46
kbringardthen reupped the new images and had a Cent 5.6 all-in-one image14:46
kbringard(not sure if that's what you guys were talking about, but thought I'd toss that out there)14:46
*** lborda has joined #openstack14:46
*** berendt has joined #openstack14:46
naehringHi! We have a little glance related problem. We've successfully uploaded an image with 5GB size (machine) to glance, the file in the filesystem is 5GB. "glance index" shows 2GB and the file which is downloaded to the node is 2 gb. We are running ext3 and glance is using the filesystem-backend. Is there any known problem, e.g. a fixed size of 2. GB?14:48
*** shentonfreude has quit IRC14:49
kbringardnaehring: it's probably qcow214:49
naehringno, the img is raw.14:49
kbringardoh, hmmm14:49
*** gondoi has joined #openstack14:49
kbringardif you launch it and do a df -h on it, does it say it's 2GB as well?14:49
kbringardI've personally never seen that problem, but I also make all my images qcow2 though14:50
berendtthe images has 5 gbyte on the glance api host14:50
berendtand 2 gbyte (2 ^ 31 - 1 bytes) on the compute node14:50
nagyzdoes nova-network usually run on only one host?14:51
kbringardnagyz: yea14:52
*** knorthcut has joined #openstack14:53
*** berendt_ has joined #openstack14:53
*** berendt has quit IRC14:53
*** ccustine has joined #openstack14:57
uvirtbotNew bug: #767252 in nova "Documentation incorrect for nova username on MySQL database in nova.conf" [Undecided,Fix released]
*** dragondm has joined #openstack15:02
*** johnpur has joined #openstack15:03
*** ChanServ sets mode: +v johnpur15:03
*** nwww has joined #openstack15:03
*** gondoi has quit IRC15:04
*** uksysadmin has quit IRC15:05
fabiand__good day. can someone tell me to what architecture parts the following files belong: nova-{console,ajax-console-proxy,debug,direct-api,logspool,spoolsentry,vncproxy} and stack, all in /usr/bin15:07
uvirtbotNew bug: #767268 in nova "EC2 API methods for dealing with security groups doesn't support GroupId" [Undecided,New]
*** MarkAtwood has joined #openstack15:16
*** shentonfreude has joined #openstack15:17
*** markvoelker has joined #openstack15:18
*** gondoi has joined #openstack15:19
*** shentonfreude has quit IRC15:22
*** markvoelker1 has joined #openstack15:24
*** markvoelker has quit IRC15:24
vishykbringard: interesting15:25
kbringardvishy: ? the cent stuff?15:26
vishynaehring: there was a bug in glance at one point about the maximum size of image15:26
vishykbringard: yes, does that euca version have some kind of init script to get the keys?15:26
naehringvishy: it looks like.15:26
kbringardvishy: yea, I think so15:26
berendt_vishy, the upload of the image is working fine.. i think only the size entry in the index is wrong15:26
vishynaehring: I thought it was fixed15:26
kbringardbut, it was awhile ago that I started my iteration15:26
kbringardso I don't remember if it had it, or if I hacked it in15:26
vishyberendt_: did you upgrade from an old version of glance?15:27
vishynaehring: ^^15:27
berendt_reset database?15:27
vishybecause the image_size was an int32 before15:27
vishygot changed to a BigInt recently15:27
*** adjohn has quit IRC15:27
berendt_| size             | int(11)      | YES  |     | NULL    |                |15:28
berendt_vishy, yes it's a bigint after resetting the database :)15:29
naehringhm, we executed the migration script manually... this may be the reason15:29
berendt_naehring, we never excecute a migration script for glance15:31
*** fabiand__ has quit IRC15:32
uvirtbotNew bug: #767279 in glance "is not possible to upload files bigger than 2 ^ 31 - 1 bytes (2 GBytes)" [Undecided,Invalid]
*** dendrobates is now known as dendro-afk15:32
naehringsure... :)15:32
berendt_i thought glance handles the database stuff automatically.. i never created the tables in the glance database15:32
berendt_and ignore the bug above.. already closed the report as invalid ;)15:32
*** enigma has joined #openstack15:33
kbringardthis looks useful15:33
*** rcc has quit IRC15:34
*** joloughlin has joined #openstack15:39
*** shentonfreude has joined #openstack15:39
joloughlinany idea what this means: Found insatnce ... in DB but no VM. State=015:39
*** naehring has quit IRC15:39
kbringardgenerally it means the VM went missing, but the DB still thought it was running15:40
vishybrendt_ there is a db sync command15:40
kbringardI think state = 0 is shutdown15:40
kbringardbut vishy can correct me on that if I'm wrong15:40
joloughlinkbringard: im using glance, you think perhaps the problem is there?15:41
*** metoikos has quit IRC15:41
kbringardjoloughlin: hmmm, it could be anywhere... I'd check the nova-compute.log on the compute instance where the VM was supposed to be running15:42
joloughlinthats the message i see in nova log on the node where it should be running15:42
kbringardpersonally, when I've had that problem, it's been when there was an issue with the backing file in _base/, or when my kernel/ramdisk weren't loaded properly15:43
doudeI try to make a snapshot on Nova (bzr1003) with libvirt/KVM, but compute node fails with error "Error: 'architecture'". I use Glance (bzr116) with file backend. What the mistake ? I tried to add the properties 'architecture' to image but it changes nothing.15:43
kbringardso the VM launched but never booted15:43
*** spectorclan_ has joined #openstack15:44
joloughlini've just moved to glance after cactus install15:44
joloughlinmust have done something wrong15:44
joloughlinall is ok on the cloud controller node15:44
kbringarddoes euca-describe-images work for you?15:44
joloughlinyeah, thats ok15:44
joloughlinuec-publish does not work15:45
joloughlinso i use nova-manage image15:45
joloughlinto upload images15:45
kbringardI've been using glance-upload15:45
kbringardmaybe try this:15:45
spectorclan_The OpenStack Cactus Webinar is back on with a new tool tomorrow at 4pm EST; check out to register. Sorry again for the confusion15:45
kbringardto upload your images15:45
kbringardit's a wrapper I wrote to make sure the kernel and ramdisk get linked properly15:45
joloughlini'll give that a go, thanks15:45
*** lool has quit IRC15:45
kbringardand the type=machine attribute gets set15:46
kbringardI have no idea if that's your problem15:46
kbringardbut it could be15:46
joloughlinthanks for help15:46
kbringardnp, hopefully it was actually helpful :-D15:46
*** lool has joined #openstack15:47
*** shentonfreude has quit IRC15:48
*** dragondm has quit IRC15:50
vishyjoloughlin: did you install cactus from packages?15:51
joloughlinfrom the ppa15:51
vishyjoloughlin: are you running nova-objectstore?15:51
joloughlinon the node glance is installed on all is ok15:51
joloughlinjust when i add extra compute node15:52
joloughlini get a problem15:52
*** deshantm_laptop has quit IRC15:52
*** z0 has quit IRC15:52
*** deshantm_laptop has joined #openstack15:53
*** shentonfreude has joined #openstack15:55
vishyjoloughlin: you have to run nova-objectstore to use uec-publish15:57
vishyjoloughlin: the images will still end up in glance15:57
kbringardvishy: do you have to set the glance backend to s3, or will it automatically sort itself out?15:58
vishyjoloughlin: all of the nodes need --glance_host=<ip> and --image_service=nova.image.glance.GlanceImageService15:58
joloughlinvishy: so I run nova-objectstore and glance?15:58
vishykbringard: no, it still backends normally15:58
kbringardah, that's cool, I did not know this15:58
joloughlinvishy: thats in my config15:58
*** nagyz has quit IRC15:59
vishykbringard: ec2_api will grab the part files from nova-objectstore (S3), combine and decrypt them and upload them to whatever image service you have specified15:59
kbringardthat is really cool15:59
kbringardgood work on that15:59
jaypipeskbringard: actually, it's kinda overly complicated :)15:59
jaypipesas I'm sure vishy would agree ;)16:00
kbringardhaha, well, it sounds cool to me :-)16:00
vishyjaypipes: aye.  I blame amazon's retarded image upload model16:00
vishyjaypipes: with snapshotting in we never have to deal with it again (I hope)16:01
*** maplebed has joined #openstack16:01
jaypipesvishy: it's OK, you can blame me instead if you want :) I'm an easy target. :)16:02
jaypipesvishy: "Blame it in the Glance, yeah, yeah!"16:02
*** shentonfreude has quit IRC16:02
kbringardhaha, jaypipes, remind me to buy you a beer next week16:02
vishyjaypipes: meh, it isn't glance's fault16:02
kbringardor a beverage of your choosing16:03
*** mdomsch has quit IRC16:03
jaypipeskbringard: heh, beer is fine :)16:03
vishyBlame it on the G-G-G-Glance-ahol?16:03
jaypipesvishy: lol16:03
jaypipesvishy, kbringard: in all seriousness, would be great to have a single tool that removes the need for doing the bundling and uploading/publishing separately, and can be used in both EC2 and OpenStack APIs...16:04
vishystack compute snapshot?16:05
kbringardcat uec-publish-tarball glance-uploader.bash > new-tool.sh16:09
*** guigui1 has quit IRC16:16
*** rchavik has quit IRC16:20
jaypipesvishy: were you being serious about stack compute snapshot? is that an actual command?16:21
*** mahadev has joined #openstack16:21
*** rapchaz has joined #openstack16:22
*** rapchaz2 has joined #openstack16:22
*** deshantm_laptop has quit IRC16:22
*** z0 has joined #openstack16:22
*** nacx has quit IRC16:24
kbringarddoes live_migration work on kvm?16:26
*** MarcMorata has joined #openstack16:29
ccookeSo... Is there anyone I can talk to about getting Xen working?16:29
*** MarcMorata has quit IRC16:29
*** z0 has quit IRC16:30
berendt_sorry.. wrong window16:30
*** z0 has joined #openstack16:34
colekbringard: direct kvm support, yes. UEC, no16:34
kbringardcole: cool, thanks16:34
*** vvuksan has joined #openstack16:35
*** dendro-afk is now known as dendrobates16:35
coleuec = nova..sorry! wrong channel! ;)16:36
*** jaypipes has quit IRC16:36
*** dprince has joined #openstack16:37
colekbringard: see for more info16:37
kbringardcole: so, if I'm using KVM as my HyperVisor, I can't "nova-manage vm live_migration ..." and have it work16:38
vishyvishy: that syntax is wrong, but yeah we have snapshotting in, and there is a compute api command for it16:39
vishyjaypipes: ^^16:39
ShermanBoydkbringard: I noticed that kvm live migration has a bug16:39
ShermanBoydkbringard: in the cactus release notes ...16:39
colekbringard: but it's bugged16:39
kbringardah, thanks16:39
kbringardyou guys rule :-)16:39
ShermanBoydis anyone taking video at the summit?16:49
ShermanBoydI'm still thinking about flying up there ...16:50
*** blamar has quit IRC16:50
alekibangomushrooms? :)16:50
alekibangoah summit!16:50
ShermanBoydno to san jose!16:51
*** jaypipes has joined #openstack16:51
*** rapchaz2 has quit IRC16:52
*** rapchaz has quit IRC16:52
jaypipesdprince: running your clear_image_props branch locally... gonna play a bit before I approve it, ok?16:54
kbringardjaypipes: is that to address the image properties getting marked as deleted if you don't specify them every time on the command line?16:55
kbringardbecause that would be ever so lovely16:55
jaypipeskbringard: hehe, no actually, it's a separate issue :)16:56
kbringardah, n/m then :-)16:56
uvirtbotNew bug: #767344 in glance "glance add silently failing with vhd/ovf" [Undecided,New]
jaypipeskbringard: but I can work on that later... thinking about just putting a switch in the glance client (--no-delete-props or something liek that...)16:57
kbringardeh, no worries, it's easy enough to work around16:57
kbringardjust a slight pain when I forget to add one of the props then have to go into the DB and fix ti :-D16:57
*** blamar has joined #openstack17:01
dprincejaypipes: gotcha.17:01
*** joearnold has joined #openstack17:04
berendt_is it possible to change the default value of use-cow-images to false?17:04
berendt_i always forget to set this parameter in nova.conf... :(17:04
jaypipesdprince: you're right about the in-memory vs. disk-based. I'm just making a slight tweak to how the test calls things. I'll push in a sec.17:04
vishyberendt_: but true is so much better!17:05
berendt_vishy, but you can't use cow with xen17:05
vishyberendt_: oh... libvirt xen?17:05
berendt_viirya, yes17:05
berendt_oh.. wrong tab completion17:06
berendt_vishy, yes17:06
vishyberendt_: there is logic in there to turn it off for lxc iirc, we should duplicate it for xen17:06
berendt_vishy, would be nice.. should i create a report for that?17:06
vishyberendt_: sure17:07
*** antenagora has quit IRC17:10
dprincejaypipes: Sure. NP. So do I need to merge lp:glance again to reconcile with the new change?17:13
*** z0 has quit IRC17:13
*** z0 has joined #openstack17:14
jaypipesdprince: nope :) I'll just ask you to review a branch on LP, and then if you like it, just bzr merge it into your local branch, commit, then push...17:15
*** Ryan_Lane has quit IRC17:15
jaypipesdprince: it just attempts to clean up a little of the hackiness by making a decorator @runs_sql that can be applied to functional test cases where a disk-based data store needs to be queried...17:15
uvirtbotNew bug: #767356 in nova "change default value of --use_cow_images to false while running nova-compute on a Xen dom0 " [Undecided,New]
dprincejaypipes: yes sir. Will do.17:16
*** kashyap has joined #openstack17:16
*** blamar has quit IRC17:18
*** mihgen_ has quit IRC17:19
*** blamar has joined #openstack17:19
comstudjay: is 'is_public' new ?17:20
*** lvaughn has quit IRC17:22
*** lvaughn has joined #openstack17:22
*** dendrobates is now known as dendro-afk17:23
*** dendro-afk is now known as dendrobates17:23
*** blamar has quit IRC17:24
j05hcomstud: I don't think so. At least, I know you could always get that info.17:25
*** timrc has quit IRC17:26
*** timrc has joined #openstack17:26
comstudi've added ami's without is_public=True before, and they are public17:26
comstudbut the behavior must be different with vhd/ovf17:27
comstudadded a comment to my bug (767344)17:27
termiesoren: what would you say is the best way to predict review days? is there a schedule from when you/something updates the wiki?17:28
termiesoren: i was checking the wiki pretty often after my last one since i was at the end of the list and wanted to see the new list but at some point it hadn't changed for a while so i stopped checking17:30
*** blamar has joined #openstack17:31
*** Shentonfreude has joined #openstack17:32
*** joearnold has quit IRC17:34
*** joearnol_ has joined #openstack17:34
jaypipesdprince: ok, lemme know what you think:
jaypipestermie: heya, soren's on vacation this week, so might not get back to you in due time :)17:35
termiejaypipes: ah17:35
termiejaypipes: do you have any insight into the process?17:35
termiei'd love to be able to mark a calendar day to go check the list or something17:35
jaypipestermie: there is an email that is sent (from soren I think) when it is your review day.17:36
jaypipestermie: other than that, I don't know if there's a "process" at all ;)(17:36
termiejaypipes: aye, but that isn't a heck of a lot of warning ;)17:36
*** joearnol_ has quit IRC17:36
jaypipestermie: ya, I think he'd agree. something to automate next week, eh?17:36
*** bcwaldon has joined #openstack17:36
*** joearnold has joined #openstack17:36
jaypipescomstud: hey, sorry, yes, that has been the behaviour of glance add the whole time. glance-upload, however, defaults to is_public=True because it had a different origin (used in Xen snapshotting..)17:37
comstudi used glance add for my ami's17:38
comstudand they're all public17:38
comstudand i didn't specify is_public=True there17:38
comstudso, one or the other sounds like a bug17:39
comstudie, 'glance add' behavior is inconsistent between ari/aki/ami and vhd17:40
comstudari/aki/ami seem to all be public by default17:40
jaypipescomstud: :) not sure what to tell you... the behaviour you are seeing is the correct behaviour. I'm not sure why it would have behaved that way before, unless it had to do with this bug:
uvirtbotLaunchpad bug 736295 in glance "glance client update metadata is_public=True fails" [High,Fix released]17:41
jaypipescomstud: I'll look into it.17:42
*** dmshelton has joined #openstack17:42
comstudit looks like the this might have changed since last week17:43
comstudargh, trying to make lunch at the same time here17:43
comstudit looks like adding a ramdisk (ari) is now private by default17:44
*** joearnold has quit IRC17:44
comstudwasn't the case last week, so... i guess this was 'fixed'17:44
comstudtime to update our wiki, i guess17:44
*** joearnold has joined #openstack17:44
jaypipescomstud: well, none of that code changed this week...17:44
comstudwell, strange17:44
jaypipescomstud: you absolutely sure you weren't using glance-upload before, and not glance add?17:44
*** joearnold has quit IRC17:45
*** joearnold has joined #openstack17:45
comstudit's possible I used glance-upload... it would mean the doc I followed has changed since last week17:45
comstudwhich is possible17:45
jaypipescomstud: XenServerDevelopment?17:45
comstudour redmine version17:45
jaypipescomstud: yep, that changed. :) jk0 added a bunch of changes...17:45
comstudi saw some of his changes, i didn't think that changed tho17:46
jaypipescomstud: I saw them pop up since I have a watch on that page.17:46
comstudbut i guess it did :)17:46
comstudyeah, same here17:46
comstudi was going to go check the emails17:46
jaypipeshehe :) either way, it needs fixing (glance-upload and glance add should be aligned....17:46
comstudanyway, that's for the infos :)17:46
jk0yeah, changed to useing 'glance add' instead of glance-upload17:46
comstudyeah, that sounds right17:46
comstudjk0: ok, yeah.  we need is_public=True added to those lines17:46
jaypipescomstud: I'm going to change your bug to be about aligning glance-upload and glance add. ok with you?17:46
*** clauden_ has joined #openstack17:47
comstudjay- definitely17:47
jaypipescomstud: feel free to comment your opinion on justinsb's bug here, too, please:
uvirtbotLaunchpad bug 755917 in glance "is_public defaults to false on image upload" [Low,Triaged]17:47
comstudthnx :)17:47
jaypipesof course17:47
* jaypipes grabs another coffee...17:47
dprincejaypipes: run_sql decorator is in. Just pushed. Should be good to go. The run_sql_cmd function is useful for backend DB validations in the future as well.17:47
dprincejaypipes: Hopefuly it is useful that is.17:48
jaypipesbcwaldon: feel free to gives a once over...17:48
bcwaldonjaypipes: on it17:48
jaypipesdprince: absolutely. i hoped the decorator would make it easy to use in the future.17:48
jaypipesdprince: just prevents all the other test cases from having to use the slower disk-based data store if they don't need to, that's all...17:49
*** woleium has joined #openstack17:50
woleiumHello peeps.17:51
woleiumI'm looking onto Openstack, and am finding it hard to work out what hypervisor to use. Has anyone published a comparison of implemented features?17:52
*** kashyap has quit IRC17:53
dprincejaypipes: One more thing on the decorator. Would we want to support mysql as well?17:53
woleiumI understand for instance that live migration was supposed to be included for KVM in cactus (I don't know if it made it though)17:53
jaypipesdprince: hmm, could do...17:54
jaypipeswoleium: we try not to make judgment calls on which hypervisor to choose :)17:55
kpepplewoleium: depends on what you want. yes, live migration for kvm made it into cactus.17:55
jaypipeswoleium: everyone has their own opinion on what's best.. ;)17:55
dprincejaypipes: Okay. So if the connection string starts with mysql I'll just skip the connection URL munging altogether.17:55
jaypipesdprince: or just change it to respect the env GLANCE_TEST_SQL_CONNECTION variable?17:55
dprincejaypipes: Sure.17:56
*** clauden_ has quit IRC18:01
*** mahadev has quit IRC18:01
*** mahadev_ has joined #openstack18:01
*** mahadev has joined #openstack18:03
*** mahadev_ has quit IRC18:03
berendt_how can i find some documentation how to use the meta data server?18:04
*** mray has quit IRC18:04
berendt_how = where18:04
*** ryker has quit IRC18:04
woleiumjaypipes: I'm sure everyone has an opinion, but at he risk of starting a flame war I'd love to compile a wikipedia style comparison table to help new users decide18:05
kpepplewoleium: +1 for flame wars :)18:05
*** dragondm has joined #openstack18:06
*** mahadev has quit IRC18:07
kpepplewoleium: doing a full blown "this hypervisor vs that hypervisor" feature comparison might be difficult but a "what hypervisor supports what OpenStack feature" would be helpful ... I'd help out with that if you started a page on http://wiki.openstack.org18:07
*** mahadev has joined #openstack18:08
*** koolhead17 has joined #openstack18:08
* woleium wanders over to the wiki and pokes it with a blunt stick.18:08
koolhead17hi all18:10
* notmyname that doesn't seem to do anything18:10
*** Ryan_Lane has joined #openstack18:10
*** czajkowski has quit IRC18:12
annegentlewoleium: poke at for starters18:15
annegentlewoleium: the tables are kinda nice18:15
*** lvaughn_ has joined #openstack18:15
annegentlewoleium: thanks to jaypipes for the nifty tables18:15
*** lvaughn has quit IRC18:18
woleiumThanks  annegentle. I'll get on it as soon as my wiki account welcome email arrives :-)18:18
*** mahadev has quit IRC18:18
*** mahadev has joined #openstack18:19
woleiumooh, no email verification - how trusting :-)18:19
*** fabiand__ has joined #openstack18:26
HugoKuo_how to start up an instance with the ttylinux_image which I added by #glance add ~~18:28
HugoKuo_14               hugo test                      raw                  ovf                        2434610618:29
HugoKuo_am I wrong :<18:30
*** smeier00 has joined #openstack18:38
*** grape has quit IRC18:39
*** smeier00 has quit IRC18:41
*** grape has joined #openstack18:43
*** dendrobates is now known as dendro-afk18:45
kbringardHugoKuo_: if you run euca-describe-images does it show up?18:45
kbringardand does it reference the kernel and ramdisk properly?18:45
*** photron has quit IRC18:45
openstackjenkinsProject burrow build #6: SUCCESS in 13 sec:
openstackjenkinsTarmac: Added client API interfaces (Account and Queue), modified HTTP response formatting, misc doc/code cleanup.18:47
kbringarddid this ever get implemented?18:49
*** knorthcut has quit IRC18:50
*** knorthcut has joined #openstack18:50
*** joearnold has quit IRC18:52
*** smeier00 has joined #openstack18:52
berendt_is there a difference how nova boot and euca-run-instances initiate the creation of a new machine?18:53
jaypipesannegentle: that Feature matrix should really be expanded at the summit... you should grab the PTLs for 15 minutes and cram a bunch of bullet points into that table...18:53
johnpurjaypipes: good call18:54
jaypipesjohnpur: dammit, I hate when my boss notices when I'm working...18:54
jaypipesjohnpur: lurker. :P18:54
johnpurjaypipes: lurker, or everpresent and omniscient?18:57
bcwaldonway to class it up18:57
jaypipesjohnpur: hmmm...lurker. :P18:57
*** Zangetsue has joined #openstack18:58
*** smeier00 has left #openstack18:59
openstackjenkinsProject nova build #834: SUCCESS in 2 min 30 sec:
openstackjenkinsTarmac: attempts to make the docstring rules clearer18:59
*** ryker has joined #openstack19:03
*** mray has joined #openstack19:04
*** arun_ has quit IRC19:06
*** mihgen has joined #openstack19:13
openstackjenkinsProject nova build #835: SUCCESS in 2 min 29 sec:
openstackjenkinsTarmac: Docstring cleanup and formatting. Minor style fixes as well.19:14
*** arun_ has joined #openstack19:18
*** HugoKuo_ has quit IRC19:23
*** cw has quit IRC19:23
*** HouseAway is now known as AimanA19:25
*** ahale has quit IRC19:38
woleiumstarted a page for hypervisor feature comparison (, not much there at the moment, but I'll add to it as my research progresses. thanks for the pointers annegentle.19:42
*** RickB17 has quit IRC19:45
* nhm_ bangs his head against iptables19:46
*** adiantum has quit IRC19:51
uvirtbotNew bug: #767545 in nova "Password reset not ran on instance create in XenServer" [Undecided,New]
*** fabiand__ has left #openstack19:55
*** troytoman-away is now known as troytoman19:56
*** clauden_ has joined #openstack19:58
koolhead17nhm_: hehe19:58
nhm_koolhead17: nova-network and my original iptables rules do not like each other at all.20:00
*** mahadev_ has joined #openstack20:03
*** bcwaldon has quit IRC20:04
sorennhm_: How so? I tried pretty hard to make them not interfere with existing rules.20:04
*** mahadev has quit IRC20:04
sorennhm_: In fact, could you please file a bug about it?20:05
nhm_soren: Sure, let me try to figure out what's wrong first though.  It's entirely possible it's my own idiocy. :)20:05
*** joearnold has joined #openstack20:06
nhm_soren: I'm really bad at iptables.  I hate it, so that is likely a big part of it. ;)20:06
nhm_soren: Right now I just disabled all of my own rules and am using the nova-network stuff and just trying to get the VM nat forwarding rules working.20:07
nhm_soren: I'm doing br100 on a single NIC with flatdhcpmanager.  I think I'm running into the same problem as this guy:
nhm_soren: so once I get that solved then I can figure out why my other NAT rules seem to break.20:09
nhm_(for the management network)20:09
*** Ryan_Lane has quit IRC20:10
vishynhm_: one interface?20:10
nhm_vishy: yeah.20:10
vishynhm_: what exactly is failing?20:10
nhm_vishy: The VMs can all talk to each other and the nova-network/nova-api/etc host, but not out to the net.20:11
vishynhm_: is this you, or someone else with the same problem ?20:11
nhm_vishy: so I figure there is something going on with the forwarding rules.20:11
vishyi has magic for you20:11
nhm_vishy: that's someone else20:11
vishyon network host: ip link set promisc on dev br10020:12
nhm_vishy: no dice...20:13
vishytook me forever to figure that one out :)20:14
nhm_vishy: I bet. :)20:14
vishyand i still don't know why it works20:14
vishyyou have ip_forward enabled on network host?20:14
nhm_vishy: I do see packets increasing slowly in the nova-network-FORWARD chain for br100 in.20:15
*** bcwaldon has joined #openstack20:15
nhm_vishy: yep, my management network used to forward stuff from the host nodes to the internet until that broke. ;)20:15
*** mahadev_ has quit IRC20:16
nhm_wow, I mangled that sentence.20:16
*** KnuckleSangwich has joined #openstack20:16
KnuckleSangwichIs there an agenda for days 3/4 of the summit next week?20:16
nhm_the same node that runs nova-network used to provide NAT forwarding for the host nodes on the management network.  Does that sound better?20:16
*** mahadev has joined #openstack20:16
*** ctennis has quit IRC20:18
vishyyes i understand that20:18
vishynetwork debugging is fun!20:19
nhm_vishy: hah!20:19
vishydo you have --private_range set?20:19
nhm_I do20:19
vishydo you have --my_ip set?20:20
nhm_does --public_interface have an effect on the nat rules?20:21
*** kakella has joined #openstack20:21
nhm_Perhaps I'm setting that wrong...20:21
uvirtbotNew bug: #767623 in swift "Stats system reports are broken" [Undecided,New]
*** kakella has left #openstack20:21
*** dendro-afk is now known as dendrobates20:22
*** sleepybox is now known as lazybox20:24
*** dprince has quit IRC20:27
vishyhmm well let me think20:29
vishy--public_interface is where floating ips are created20:29
*** ctennis has joined #openstack20:30
*** bcwaldon has quit IRC20:32
nhm_vishy: Ok, I think I know at least part of my problem.  I use a single interface for nova (ie all compute nodes are on eth0, nova-network is on eth0, etc).  The nova-network machine has a second interface talking to the internet with no nova stuff associated with it.  So I need to setup the iptables rules so that it forwards through that interface.20:32
nhm_So I think I'm "one interface" as far as the bridget shares the same interface as my management network, but not in that I have to route through the second interface on the nova-network machine.20:33
vishyhmm gotcha20:34
vishyso the default SNAT rule should be handling that20:34
nhm_vishy: Like I said, I suck at iptables. ;)20:34
vishybut perhaps routing_source_ip is not set?20:35
*** jarrod has quit IRC20:35
nhm_vishy: mind if I pastebin some stuff for you?20:35
berendt_why are my xen instances are always in state "blocked"? i can't start live migration (because nova-manage want to have state running)20:39
kbringardnhm_, vishy: I had issues where routing_source_ip wouldn't detect properly and it was routing all outbound traffic from the host node through lo20:40
kbringardso until I forced it on the network controller, I would lose all external routing on said machine20:40
kbringardnot sure if that relates to what you're debugging, but there it is :-)20:40
vishynhm_: what is your external ip?20:41
nhm_vishy:, bcast is
nhm_vishy: just didn't want it in the pastebin since I set it public.20:41
vishyso you will need routing_source_ip=
vishyoh i see20:42
vishyyou changed them all to 172.xxx20:42
vishywhich interface is connected to the other nodes?20:42
nhm_vishy: eth020:43
vishyand eth1 is the public?20:43
nhm_vishy: yep20:43
vishyso public_interface should be eth120:43
nhm_vishy: yeah, I was wondering if that was set wrong...20:43
vishyi don't see the ip on eth120:43
*** joearnold has quit IRC20:44
vishy172.0.0.0/12 whatever that is really needs to not conflict with your external ip20:44
*** knorthcut has quit IRC20:45
nhm_vishy: internal network is, vm network is, and external ip is
annegentleberendt_: live migration is only supported with KVM hypervisor, right? See:
kbringardto setup multiple zones, I just specify --node_availability_zone in the config file of each node?20:46
vishyah yeah that is a problem20:46
*** joearnold has joined #openstack20:46
vishy172.0.0.0/12 conflicts with
vishymaybe try ?20:47
uvirtbotNew bug: #767667 in nova "nova-manage vpn list is broken" [Low,In progress]
uvirtbotNew bug: #767676 in swift "Bug in x-object-manifest value handling" [Medium,Confirmed]
uvirtbotNew bug: #767677 in nova "migration not possible because instance is in state blocked (xen)" [Undecided,New]
*** bcwaldon has joined #openstack20:48
nhm_vishy: internal network is the management network for the hosts... I don't think I'm setting /12 anywhere for that?  I am setting though...20:48
*** mgoldmann has quit IRC20:49
*** ryker has quit IRC20:49
vishyis 1/12 valid?20:49
vishyclearly not, it is converting it to
vishytry 1/1620:50
jaypipesbcwaldon: we good to merge
nhm_vishy: sorry if this is a stupid question but... where?20:51
vishywhich is what is used to generate the snatting rules20:51
vishy-A nova-network-snat -s -j SNAT --to-source
nhm_ah, ok.  so change from to
*** pquerna has left #openstack20:54
nhm_bah, I see why that was invalid.20:55
nhm_thanks for the help20:56
vishyhopefully that gets you farther20:56
vishythe guy on the ml has a very similar setup and he got it working20:56
*** bcwaldon has quit IRC20:57
*** dmshelton has left #openstack20:59
*** jonkelly has joined #openstack21:02
*** dendrobates is now known as dendro-afk21:02
*** nelson has quit IRC21:06
*** nelson has joined #openstack21:06
*** ryker has joined #openstack21:07
*** dscastro has quit IRC21:09
openstackjenkinsProject nova build #836: SUCCESS in 2 min 27 sec:
openstackjenkinsTarmac: Added an option to so you can run just pep8. So now you can:21:09
openstackjenkins./ --just-pep821:09
openstackjenkins./ -p21:09
*** markvoelker1 has quit IRC21:12
*** lborda has quit IRC21:14
*** nwl has joined #openstack21:15
*** allsystemsarego has quit IRC21:20
*** aimon has joined #openstack21:21
openstackjenkinsProject nova build #837: SUCCESS in 2 min 27 sec:
openstackjenkinsTarmac: Another small round of pylint clean-up.21:24
*** rnirmal has quit IRC21:31
*** rnirmal has joined #openstack21:37
*** kakella has joined #openstack21:40
*** kakella has left #openstack21:40
*** Ryan_Lane has joined #openstack21:45
*** thielmann_ has quit IRC21:49
*** troytoman is now known as troytoman-away22:01
*** mahadev has quit IRC22:02
*** dspano has quit IRC22:02
*** matiu has joined #openstack22:03
*** kbringard has quit IRC22:04
*** spectorclan_ has quit IRC22:08
termievishy: any reason we are using a lockfile for synchronised vs an in-process lock? do we expect multiple processes of the same service to be running on the same machine?22:13
*** matiu has quit IRC22:14
*** cw has joined #openstack22:16
termievishy: oh, it looks like we only use an external one if it is asked for explicitly22:16
*** Shentonfreude has quit IRC22:18
*** MotoMilind has quit IRC22:19
Ryan_Laneso, since glance doesn't have an upstart, what's a good way to ensure it starts on boot?22:19
*** MotoMilind has joined #openstack22:20
Ryan_Lanealso, the release ppa seems to have the gamma2 packages. when will the finalized release packages be released?22:24
Ryan_LaneI also get the following error when trying to upgrade glance:22:24
Ryan_LaneERROR: unrecognized command 'db'22:24
Ryan_Lanedpkg: error processing glance (--configure):22:24
*** jonkelly has quit IRC22:25
*** matiu has joined #openstack22:25
Ryan_Lane\o/ seems the newer glance packages have an upstart :)22:26
Ryan_Laneone problem solved.22:26
woleiumCompute development has been on Ubuntu 10.04 LTS, but do you use the 32 or 64bit version?22:28
woleiumty :-)22:28
Ryan_Lanedo you mean, me, or the community? :)22:28
woleiumlol, both I guess. I'm trying to decide what to use22:29
Ryan_Lane64 bit is likely best22:29
woleiumI guess it doesn't matter that much, but i was siding with 64 anyway :-)22:29
woleiumRyan_Lane: I guess you could create your own init.d script for glance. look at /etc/init.d/skeliton for ideas (or copy the ssh one)22:31
uvirtbotNew bug: #767798 in nova "model used in method get_cpu_info() in nova/virt/ is not available in capabilities on a xen system" [Undecided,New]
Ryan_Lanewoleium: no need, there is an upstart in the newer package version22:31
woleiumoh, good stuff :-)22:31
*** adiantum has joined #openstack22:34
*** enigma has quit IRC22:35
uvirtbotNew bug: #767806 in swift "CSV files generated by stats system are broken" [Undecided,New]
*** vernhart1 has quit IRC22:44
uvirtbotNew bug: #767809 in swift "Object replicator breaks when running object replicator with no objects on the server" [High,In progress]
*** vernhart has joined #openstack22:47
*** shentonfreude has joined #openstack22:48
*** sdadh01 has quit IRC22:50
*** rsaidan has quit IRC22:50
*** shentonfreude has quit IRC22:52
*** shentonfreude has joined #openstack22:52
*** miclorb_ has joined #openstack22:53
*** vvuksan has quit IRC22:59
*** j05h has quit IRC22:59
*** vvuksan has joined #openstack22:59
*** shentonfreude1 has joined #openstack22:59
*** sdadh01 has joined #openstack23:00
*** shentonfreude has quit IRC23:02
*** vernhart has quit IRC23:03
*** vvuksan has quit IRC23:04
*** shentonfreude has joined #openstack23:05
*** j05h has joined #openstack23:06
*** shentonfreude2 has joined #openstack23:07
*** shentonfreude1 has quit IRC23:08
*** matiu_ has joined #openstack23:10
*** matiu_ has joined #openstack23:10
*** shentonfreude has quit IRC23:10
*** shentonfreude2 has quit IRC23:11
*** matiu has quit IRC23:13
*** matiu_ is now known as matiu23:13
*** vvuksan has joined #openstack23:20
*** rnirmal has quit IRC23:20
*** vernhart has joined #openstack23:22
uvirtbotNew bug: #767849 in nova "kernel-id and ramdisk-id are not showing up in metadata" [Medium,In progress]
*** aimon has quit IRC23:37
*** jeffjapan has joined #openstack23:40
*** santhosh has quit IRC23:50
*** matiu_ has joined #openstack23:52
*** Zangetsue has quit IRC23:54
*** matiu has quit IRC23:54
*** matiu_ is now known as matiu23:54

Generated by 2.14.0 by Marius Gedminas - find it at!