Wednesday, 2011-08-31

*** adiantum has joined #openstack00:00
*** lorin1 has quit IRC00:01
*** GeoDud has joined #openstack00:03
*** ColdDog has quit IRC00:04
*** nati has quit IRC00:04
*** pguth66 has quit IRC00:06
*** wilmoore has quit IRC00:07
*** obino has quit IRC00:07
*** jeffjapan has joined #openstack00:07
*** shentonfreude has joined #openstack00:08
*** obino has joined #openstack00:11
*** adiantum has quit IRC00:12
*** ncode has quit IRC00:12
*** joearnold has quit IRC00:14
*** tcampbell has joined #openstack00:14
*** dh has quit IRC00:15
*** dh__ has joined #openstack00:15
*** morfeas has quit IRC00:16
*** bonzay has quit IRC00:16
*** slriv has quit IRC00:16
*** sam_itx1 has quit IRC00:16
*** dabo has quit IRC00:16
*** ccooke has quit IRC00:16
*** keekz has quit IRC00:16
*** lynxman has quit IRC00:16
*** Eyk^off has quit IRC00:16
*** jdurgin has quit IRC00:16
*** bengrue has quit IRC00:16
dysinger1Is there an easy way to disable (not delete) an account in swift temporarily ?00:17
*** rnorwood has joined #openstack00:17
*** slriv has joined #openstack00:17
*** bengrue has joined #openstack00:17
*** adiantum has joined #openstack00:18
*** lorin1 has joined #openstack00:19
*** ccooke has joined #openstack00:20
*** obino has quit IRC00:20
*** keekz has joined #openstack00:20
*** lynxman has joined #openstack00:20
*** nati has joined #openstack00:22
*** lorin1 has quit IRC00:23
*** adiantum has quit IRC00:24
*** nati has quit IRC00:25
*** nati has joined #openstack00:25
*** morfeas has joined #openstack00:25
*** deshantm_away is now known as deshantm00:25
*** bengrue has quit IRC00:26
*** nati_ has joined #openstack00:28
*** ncode has joined #openstack00:28
*** nati has quit IRC00:29
*** sam_itx has joined #openstack00:29
*** adiantum has joined #openstack00:29
*** bengrue has joined #openstack00:31
*** berto- has quit IRC00:32
*** nati_ has quit IRC00:35
*** dabo has joined #openstack00:36
*** ccc11 has joined #openstack00:38
*** adiantum has quit IRC00:39
*** nati has joined #openstack00:40
*** JordanRinke is now known as JordanRinke-AFK00:42
*** po has quit IRC00:43
*** troytoman is now known as troytoman-away00:44
*** adiantum has joined #openstack00:45
*** lorin1 has joined #openstack00:50
*** shentonfreude has quit IRC00:50
*** cpg|away is now known as cpg00:53
*** wilmoore has joined #openstack00:56
*** KnuckleSangwich has joined #openstack01:00
*** adiantum has quit IRC01:02
*** adiantum has joined #openstack01:03
notmynamedysinger1: not in swift. you could perhaps disable the account in your auth system01:05
dysinger1swauth  - k01:05
dysinger1there's command line to add & delete01:05
*** rsampaio has joined #openstack01:07
*** lorin1 has quit IRC01:07
*** lorin1 has joined #openstack01:08
*** adiantum has quit IRC01:10
*** rnorwood has quit IRC01:13
*** rnorwood has joined #openstack01:15
*** adiantum has joined #openstack01:17
*** skane has joined #openstack01:18
*** skane_ has joined #openstack01:19
*** catbot has joined #openstack01:20
catbotHave anybody ever booted win7 image successfully on Nova?01:21
uvirtbotNew bug: #837778 in nova "man pages are missing for all the programs" [Undecided,New]
catbotI attempted to boot a win7 on nova, but failed on the blue screen.01:22
*** thickskin has left #openstack01:22
*** skane has quit IRC01:22
catbotAccording to the error message, it appeared the filesystem is corrupted01:23
*** skane_ has quit IRC01:23
vishycatbot: I have booted win server 200801:24
*** adiantum has quit IRC01:24
vishydid you install virio drivers?01:24
*** mattray has joined #openstack01:24
catbotMy win7 image is 8GB size, I guess nova has a size limit on image?01:25
catbotNo, I did not install any 3rd drivers for win701:26
*** aliguori has quit IRC01:27
vishycatbot: in that case it won't boot01:27
*** bonzay has joined #openstack01:28
vishyyou need virio block drivers in the image01:28
tjoyvishy: is that howto the state of the rt?01:29
catbotIn the manual of openstack, I learned the suggestion to install virio for win2008.01:29
*** adiantum has joined #openstack01:29
catbotI not sure if the win7 need the virio as well01:29
vishycatbot: yes it does01:30
*** thickskin has joined #openstack01:30
vishytjoy: the instuctions for converting the image at the end are probably wrong01:30
catbotthanks vishy, I will try again with installing virio01:30
*** rods has quit IRC01:30
vishytjoy: you can just upload the image into glance01:30
*** woleium_ has joined #openstack01:31
*** vernhart has quit IRC01:31
*** woleium has quit IRC01:33
*** woleium_ is now known as woleium01:33
*** HowardRoark has quit IRC01:35
*** adiantum has quit IRC01:35
*** adiantum has joined #openstack01:41
*** nati has quit IRC01:42
*** hadrian has quit IRC01:42
*** Ryan_Lane has quit IRC01:43
*** mfer has joined #openstack01:43
*** bengrue has quit IRC01:45
*** adiantum has quit IRC01:48
*** lorin1 has quit IRC01:49
*** vladimir3p has quit IRC01:49
*** maplebed_ has joined #openstack01:52
*** adiantum has joined #openstack01:53
*** tryggvil has quit IRC01:53
*** lorin1 has joined #openstack01:54
*** woleium_ has joined #openstack01:54
*** huslage has quit IRC01:55
*** maplebed has quit IRC01:56
*** mfer has quit IRC01:57
*** woleium has quit IRC01:57
*** woleium_ is now known as woleium01:57
*** HowardRoark has joined #openstack01:59
*** adiantum has quit IRC02:00
catbotare there any ways to remove/delete a uploaded image?02:01
*** winston-d has quit IRC02:02
catbotall of documents just tell how to upload or publish a image, but no documents tell how to remove a image02:03
*** winston-d has joined #openstack02:04
*** adiantum has joined #openstack02:13
*** ahasenack has quit IRC02:18
*** woleium has quit IRC02:18
*** dragondm has quit IRC02:23
*** dirakx1 has quit IRC02:27
*** mdomsch has joined #openstack02:29
*** cpg is now known as cpg|away02:42
*** axisys has joined #openstack02:42
*** mattray has quit IRC02:42
*** dendro-afk is now known as dendrobates02:44
*** infinite-scale has quit IRC02:46
*** infinite-scale has joined #openstack02:47
*** ChrisAM1 is now known as ChrisAM02:47
*** DavorC has left #openstack02:49
*** woleium has joined #openstack02:54
vishyor delete through glance02:57
vishycatbot: ^^02:57
*** bradm has joined #openstack02:59
*** osier has joined #openstack02:59
*** nelson____ has quit IRC03:01
*** nelson____ has joined #openstack03:02
*** HowardRoark has quit IRC03:03
*** HowardRoark has joined #openstack03:03
catbotthanks vishy, looks like euca-deregister's working well03:08
*** tsuzuki_ has joined #openstack03:09
*** jtanner has quit IRC03:16
*** lorin1 has quit IRC03:17
*** michael_ has joined #openstack03:22
*** infinite-scale has quit IRC03:27
*** dysinger1 has quit IRC03:30
*** martine_ has quit IRC03:32
*** rnorwood has quit IRC03:36
*** adjohn has quit IRC03:37
*** HugoKuo_ has joined #openstack03:37
*** cpg|away is now known as cpg03:38
*** HugoKuo_ has quit IRC03:43
*** ccc11 has quit IRC03:46
*** ccc11 has joined #openstack03:46
uvirtbotNew bug: #837817 in glance "bin/glance should handle cache being disabled" [Undecided,New]
*** rnorwood has joined #openstack03:49
*** catbot has quit IRC03:49
*** rnorwood has quit IRC03:56
*** rnorwood has joined #openstack04:09
*** shentonfreude has joined #openstack04:09
*** rms-ict has quit IRC04:13
*** tcampbell has quit IRC04:13
*** syah has joined #openstack04:23
tjoySo, trying to install the grid-dynamics packages on sl i386:04:24
tjoyBK_man: you said you'd be interested in seeing this04:25
tjoythat's scientificlinux 6.1 i38604:25
*** ccc11 has left #openstack04:26
*** ccc11 has joined #openstack04:30
tjoyi remember why i couldn't do this before04:31
tjoyall the deps are x86_6404:31
*** ccc11 has left #openstack04:31
*** nati has joined #openstack04:32
tjoywell, at least half of them04:34
*** rms-ict has joined #openstack04:37
*** anoopl has joined #openstack04:48
*** zul has quit IRC04:49
*** zul has joined #openstack04:50
*** wilmoore has quit IRC04:55
anooplI get this error while trying to create a nova network04:56
anooplnova: --bridge_interface is required to create a network.04:56
anooplcan someone help me04:56
*** Razique has quit IRC05:05
*** HowardRo_ has joined #openstack05:13
*** HowardRoark has quit IRC05:15
*** HowardRo_ has quit IRC05:15
*** nati has quit IRC05:25
*** miclorb has quit IRC05:25
*** obino has joined #openstack05:26
*** nati has joined #openstack05:29
*** jedi4ever has joined #openstack05:35
*** nati has quit IRC05:36
*** jedi4ever has quit IRC05:38
*** jedi4ever has joined #openstack05:38
*** shaibn has joined #openstack05:40
*** rocambol1 has joined #openstack05:44
*** anoopl has quit IRC05:49
*** guigui has joined #openstack05:49
vishyanoopl: set the flag?05:49
vishytry using --bridge_interface when you create the network?05:49
*** anoopl has joined #openstack05:53
*** GeoDud has quit IRC05:54
*** nati has joined #openstack05:58
*** miclorb has joined #openstack05:59
*** zul has quit IRC05:59
*** scas has quit IRC06:00
*** scas has joined #openstack06:00
*** anoopl has quit IRC06:00
*** zul has joined #openstack06:00
*** nati has quit IRC06:00
*** perestrelka has joined #openstack06:02
*** dysinger has joined #openstack06:07
*** rms-ict has quit IRC06:23
*** dysinger1 has joined #openstack06:37
*** shaibn has quit IRC06:43
*** koolhead17 has joined #openstack06:52
*** reidrac has joined #openstack06:55
*** vernhart has joined #openstack06:57
*** miclorb has quit IRC07:03
*** ccustine has quit IRC07:04
*** lionel has quit IRC07:04
*** siwos has joined #openstack07:04
*** lionel has joined #openstack07:04
*** adjohn has joined #openstack07:05
*** dysinger has quit IRC07:06
*** npmapn has joined #openstack07:08
*** Capashen has joined #openstack07:13
*** rms-ict_ has joined #openstack07:14
*** S_A has joined #openstack07:17
*** S_A has quit IRC07:25
uvirtbotNew bug: #837867 in nova "Add fixed IP fails" [Undecided,New]
*** S_A has joined #openstack07:33
S_Ahi all07:34
S_AI have installed Nova on ubuntu 10.10 system07:35
*** nacx has joined #openstack07:35
S_Aas of now when i  do "euca-run-instances -k test -t m1.tiny ami-tty"07:35
S_Aall of the instances are in shutdown state07:35
S_AI am stuck here07:36
S_Aany suggestions07:36
S_Atjoy: any suggestions ?07:36
tjoyi read you the first time07:36
tjoylook in the various logfiles to see what's going on.07:36
tjoynova's logfiles and the logfiles of whichever hypervisor you used07:37
S_Atjoy: "2011-08-31 13:07:53,578 INFO nova.compute.manager [-] Found instance 'instance-00000001' in DB but no VM. State=5, so setting state to shutoff."07:39
*** adjohn has quit IRC07:40
S_Atjoy: I am a first timer. so no idea. this is the message from /var/log/nova/nova-compute.log07:40
S_Atjoy: do i need to have some VM also07:40
S_AAm i missing something07:40
tjoythat's the point of the whole operation isn't it? to run VMs/07:41
S_Atjoy: how to run VM07:42
S_AI am following this guide till now
*** miclorb has joined #openstack07:44
*** caribou has joined #openstack07:46
*** vdo has joined #openstack07:46
*** javiF has joined #openstack07:46
*** vdo has quit IRC07:48
*** vdo has joined #openstack07:48
titaniumrainwho handles out IP address to an instance? dnsmasq or libvirt? libvirt injects network configuration such as Mac address. but I did not see anywhere IP address is injected to VM.07:49
titaniumrainany hint is much appreciated...07:50
*** katkee has joined #openstack07:52
*** dh__ has quit IRC07:55
siwosdnsmasq gives addresses only when you use FladDHCPManager07:55
siwosFlatManager only injects static ips into vm-s and keeps track of them in nova database07:55
*** adiantum has quit IRC07:55
siwosif you use FlatManager, you also need to specify FlatInjected=True07:56
S_AAnyone has experience in this ""07:57
S_Ai am trying to use this preconfigured vmdk07:58
S_Abut i am not getting how to configure virtualbox for network access07:58
*** nerens has joined #openstack08:00
*** miclorb has quit IRC08:00
*** adiantum has joined #openstack08:01
titaniumrainsiwos: thanks for getting back to me.08:03
siwosn/p ;-)08:04
titaniumrainsiwos: dnsmasq is a mystery to me, actually08:04
titaniumrainswios: it uses dhcpbridge to update db08:04
siwosmight be ;-)08:04
*** Xenith has quit IRC08:04
titaniumrainswios: but judging from what you said, it is libvirt that handles out IP address to VM?08:05
titaniumrainif vlanmanage is useD?08:05
*** Xenith has joined #openstack08:05
siwosI guess it's nova-compute which injects ip-s into vm-s08:06
siwosprovided you use FlatManager08:06
*** ccc11 has joined #openstack08:07
titaniumrainswios: i had a look at the code. nova-compute uses libvirt/ to spawn an instance. at the spawning stage, i.e. create_image method, it injects the ip address into VM08:07
*** Oneiroi_ has joined #openstack08:07
*** adiantum has quit IRC08:07
*** ccc111 has joined #openstack08:08
siwosaaah - ok08:08
titaniumrainswios: what i really don't understand is, what's the role of dnsmasq? or in other words, what's the difference between flatdhcp and vlan manager if dnsmasq handles out ip addresses only if we are using flatdhcp manager08:09
*** tudamp has joined #openstack08:09
*** Oneiroi has quit IRC08:10
*** st4n has joined #openstack08:12
*** ccc11 has quit IRC08:12
*** adiantum has joined #openstack08:12
st4nHi, I'm trying to configure nova+keystone. I have nova-trunk and a copy of git repo for keystone. Actually, nova does not redirect to keystone for auth requests, although I pasted api file in nova.conf08:13
st4nnova client return a 40108:13
st4nanyone can help?08:14
*** dobber has joined #openstack08:14
*** jakedahn has quit IRC08:15
*** pguth66 has joined #openstack08:16
*** adiantum has quit IRC08:19
*** eviscares has joined #openstack08:19
eviscaresGood Morning.08:19
eviscaresI would like to suggest a minor change in the Swift 1.4.3 Documentation. Whom should I bug with this?08:20
siwostitaniumrain: I guess I am not that deep in the technology yet ;-)08:20
*** adiantum has joined #openstack08:24
nerensst4n: hi, I'm also trying to intergrate nova & keystone… I've installed nova milestone via apt-get, the instrucations says we must symlink the nova/keystone directory with /keystone/keystone… what did you do for this step?08:26
titaniumrainsiwos: we have to read a bit more to understand the things behind08:27
*** rms-ict has joined #openstack08:29
st4nnerens: I did not do that, I'm using ubuntu nova trunk packages08:30
*** caribou has quit IRC08:31
nerensst4n: ok, so where did you put the keystone files?08:31
*** cpg is now known as cpg|away08:31
st4nnerens: I just installed keystone from git tree, launched keystone, and added the api_paste_config line into nova.conf08:32
*** rms-ict_ has quit IRC08:33
*** seoman__ has joined #openstack08:33
nerensst4n: so on the keystone website they say: # link keystone into the nova root dir08:36
nerensln -s keystone/keystone nova/keystone08:36
nerensst4n: is this step not needed then when using nova from packages?08:36
nerensst4n: regarding your issue, have you configured your nova endpoints in keystone?08:37
*** adiantum has quit IRC08:42
*** darraghb has joined #openstack08:46
*** caribou has joined #openstack08:46
*** adiantum has joined #openstack08:48
st4nnerens: yes, nova endpoints are configured in keystone. I assume nova trunk include necessary keystone links now since keystone blueprints have been finalized08:50
nerensst4n: hopefully someone who know's more about this can give us a hand :)08:55
*** po has joined #openstack08:55
*** adiantum has quit IRC08:55
*** jeffjapan has quit IRC08:58
*** adiantum has joined #openstack09:02
st4nnerens: I wish I could go into the source code to see what's going around, but no time actually :)09:03
*** dirkx__ has joined #openstack09:04
nerensst4n: I've got the time.. just don't have the python knowledge to know what's really going on :(09:04
*** daysmen has joined #openstack09:05
*** adiantum has quit IRC09:06
doudest4n: I install keystone with nova. To do that, I install from trunk ppa packages on my ubuntu and fetch git source of keystone in a folder. I configure keystone endpoints and I change the api-paste.ini file of nova to use keystone (the keystone config is commented in the file)09:07
doudest4n: I start keystone, and I can use OSAPI09:08
*** javiF has quit IRC09:09
*** adiantum has joined #openstack09:11
*** dendrobates is now known as dendro-afk09:14
st4ndoude: I did that, but it does not work for me...09:15
*** ccc111 has quit IRC09:18
*** adiantum has quit IRC09:18
*** adiantum has joined #openstack09:25
titaniumrainsoren: do u know that if it is libvirt that injects the IP address to instance?09:25
*** ccc11 has joined #openstack09:27
*** ccc111 has joined #openstack09:27
*** ccc11 has quit IRC09:32
*** adiantum has quit IRC09:33
*** aurigus has joined #openstack09:37
*** dirkx__ has quit IRC09:38
*** irahgel has joined #openstack09:39
sorentitaniumrain: Sorry, what?09:42
*** sunny_ has joined #openstack09:42
*** tsuzuki_ has quit IRC09:42
titaniumrainsoren: i am trying to figure out how a virtual machine gets its IP address. when booting an image, libvirt handles out mac address to an instance. then i assume that that instance will broadcast its mac address to get an ip. my question is that how dnsmasq knows which ip address to handle to that exact instance?09:43
doudest4n: Do you correctly configure the api paste file ?09:43
sunny_I got this error CRITICAL nova [-] 'module' object has no attribute 'init_host'09:43
sorentitaniumrain: The network host assigns it.09:46
sorentitaniumrain: And tells dnsmasq.09:46
*** hbosso has quit IRC09:48
sunny_I set vlanManager. my nova.conf file has
*** tryggvil_ has joined #openstack09:53
titaniumrainsoren: i have been reading the code. do you know at which method that network host tells dnsmasq?09:53
sorentitaniumrain: It updates the dnsmasq's lease file and sends dnsmasq a SIGHUP.09:54
titaniumrainsoren: dnsmasq has add, old, delete, init. but i don't know at which point network host invokes add event09:54
sorentitaniumrain: Those are events dnsmasq *emits*.09:55
sorentitaniumrain: Not ones it can receive.09:55
sorentitaniumrain: tells dnsmasq about changes to the MAC->IP mapping.09:56
titaniumrainsoren: rite. so, the correct flow is  1. mac address generated and then binded with ip address from network manager, 2. network manage updates dnsmasq, 3. libvirt assigns mac address to instance, 4. instance requests its ip from dnsmasq, 5 finally vm gets its ip?09:59
sorenNot sure about the order of 2 and 3, but basically yes.10:00
titaniumrainsoren: :) it is nova-dhcpbridge that triggers my question.10:01
titaniumrainsoren: dhcpbridge always uses rpc cast to ask network manager to update ip-mac information in the database10:02
titaniumrainsoren: it would be, potentially, a heavy burden to both rabbit and network manager if millions of vms running, wouldn't it?10:03
titaniumrainsoren: from the perspective of research, millions shall be reasonable.10:03
*** shrii has joined #openstack10:03
shriihey all10:04
sorenOne network manager woulnd't manager a million VMs.10:04
*** dirkx__ has joined #openstack10:04
shriii have one issue with my openstack cloud.10:04
shriisudenly its haveing 100% load and network get jaam.10:05
shriiit get down.10:05
titaniumrainsoren: yes, indeed it can not10:05
*** mies has quit IRC10:06
titaniumrainsoren: how about let dhcpbridge to access DB directly when add, old, delete event emitted by dnsmasq?10:06
*** mies has joined #openstack10:06
titaniumrainsoren: it would be quicker comparing to sending out rabbit msgs?10:07
sorentitaniumrain: Why do you think the database would be any better at handling this sort of load?10:07
titaniumrainsoren: currently route to update is 1. dhcpbridge sends out a msg, 2, rabbit routes it to network manager, 3. network manager updates DB10:08
titaniumrainsoren: if dhcpbridge access to DB directly10:08
titaniumrainsoren: it will save 1.) trip time to rabbit, 2.) network manager time to process the message (it may have a lot of backlog to process)10:09
sorenSeriously... We need *less* things talkting to a central datastore. Not more.10:09
sorenBacklogs are *fine*.10:09
sorenQueueing is *goo*.10:09
soren*Good*, even. :)10:10
shriitech expert are there ?10:11
shriireally time waste,10:11
*** titaniumrain_ has joined #openstack10:11
titaniumrain_soren: sorry, browser crashed10:11
titaniumrain_soren: i definitely agree with you that central DB shall not be accessed by different parties10:12
titaniumrain_soern: i was wondering how the performance of dhcp can be accelerated apart of accessing the DB directly.10:13
*** titaniumrain has quit IRC10:13
titaniumrain_soren:    :)10:13
*** titaniumrain_ has quit IRC10:14
*** titaniumrain has joined #openstack10:14
titaniumrainsoren: sorry, browser crashed. i definitely agree with you that central DB shall not be accessed by different parties. was wondering how the performance of dhcp can be accelerated apart of accessing the DB directly10:15
sorenPerformance of dhcp?10:16
sorenWhy would you want to accelerete the performance of dhcp? Is it slow?10:16
titaniumrainsoren: it will show network controller. should network controller be given the task to maintain dhcp information? judging from the functionality of network node, yes...10:18
*** javiF has joined #openstack10:19
titaniumrainif dhcpbridge can share some burden of network controller, it would be interesting, would it?10:20
sorenI'm confused.10:21
sorenDo you want to improve dhcp performance or something else?10:21
larissashrii: Error: "titaniumrain" is not a valid command.10:21
thickskinhi guys.10:21
thickskinis it possible make freebsd vm with virtio?10:22
titaniumrainsoren: it should be improving the performance of network controller since all dhcp messages are delivered to network controller.10:23
*** jj0hns0n has quit IRC10:29
*** jj0hns0n has joined #openstack10:29
*** ccc111 has quit IRC10:33
*** zigo has joined #openstack10:51
*** markvoelker has joined #openstack10:52
*** po has quit IRC10:54
titaniumrainsoren: thanks. :)11:02
*** dirkx__ has quit IRC11:02
*** seoman__ has quit IRC11:09
*** seoman__ has joined #openstack11:10
*** Razique has joined #openstack11:10
*** PeteDaGuru has joined #openstack11:10
*** ccc11 has joined #openstack11:15
tomh_help /list11:24
Raziquetomh_: yes ? :D11:26
tomh_my bad :-(11:26
Raziquehehe :)11:32
*** mfer has joined #openstack11:33
*** mnot has quit IRC11:33
*** vdo has quit IRC11:42
*** ncode has quit IRC11:44
*** nerdstein has joined #openstack11:44
*** zigo has quit IRC11:48
*** zigo has joined #openstack11:49
*** zigo-_- has joined #openstack11:53
*** guigui has quit IRC11:54
*** guigui1 has joined #openstack11:54
*** zigo has quit IRC11:54
*** guigui1 has quit IRC11:57
*** zigo-_- has quit IRC11:57
agyi've just upgraded from 2011.3~d4~20110812.1417-0ubuntu1 to 2011.3~d4-0ubuntu1 and now my user's keypairs are not working. it looks like nova-api now creates/queries using the NOVA_API_KEY instead of NOVA_USERNAME. is this change deliberate or a bug?12:00
agyi.e. the resultant sql query is now the following:12:03
agySELECT key_pairs.created_at [...] FROM key_pairs WHERE key_pairs.user_id = '$NOVA_API_KEY' AND key_pairs.deleted = 0;12:03
*** hingo has joined #openstack12:03
*** ahasenack has joined #openstack12:09
*** toki_kanno has joined #openstack12:11
*** bsza has joined #openstack12:12
*** nati has joined #openstack12:14
*** mfer has quit IRC12:19
uvirtbotNew bug: #838036 in nova "Upgrade of nova breaks user keypairs" [Undecided,New]
*** lts has joined #openstack12:25
*** doude has quit IRC12:29
*** doude has joined #openstack12:29
*** javiF has quit IRC12:31
*** rnorwood has quit IRC12:31
smoserttx, should bug 812553 be fix-released ?12:33
*** mdomsch has quit IRC12:34
smoserits marked as such in Ubuntu, but not in diablo. maybe you wait for full releases to mark things as fix-released rather than milestone12:34
ttxsmoser: yes, that's how we use that field12:34
ttxfixcommitted -> in trunk, fixreleased -> in a release12:35
*** rods has joined #openstack12:35
smoserso really i want to search with ! fix-committed12:35
*** aliguori has joined #openstack12:36
ttxsmoser: for open bugs ? yes.12:36
*** ahasenack has quit IRC12:36
*** ahasenack has joined #openstack12:37
smoserttx, i would set importance on bug 826798 to medium at best. as our beta images will have qcow formated disks available.12:37
smoser(but i can't set that)12:37
ttxsmoser: done12:38
*** jaypipes has quit IRC12:39
*** Razique_ has joined #openstack12:40
*** Razique has quit IRC12:42
*** Razique_ is now known as Razique12:42
*** owen has joined #openstack12:42
*** daysmen has quit IRC12:45
Raziqueis annegentle here ? :)12:50
*** ncode has joined #openstack12:50
*** ncode has joined #openstack12:50
*** lorin1 has joined #openstack12:51
*** 36DAB5NBC has joined #openstack12:54
*** guigui1 has joined #openstack12:56
*** lorin1 has quit IRC12:58
*** po has joined #openstack12:58
*** martine_ has joined #openstack13:00
owenis this a developer, and admin or a user channel ?13:02
*** ccc11 has quit IRC13:04
*** shrii has quit IRC13:05
ttxthis is a user channel.13:05
owenok thanks, I am doing some admin work (I am a developer), evaluating open stack for a large German physics institute, could you suggest a better channel to join, or should I just chat on this channel when I am stuck (I am quiet an experienced UNIX mass deployment guy)13:07
Raziqueowen: hi :) I think that channel suits your needs13:08
*** sandywalsh has quit IRC13:09
ttxowen: you should be fine here -- though this channel is more active on US timezones.13:10
*** mfer has joined #openstack13:11
owenok thanks I will stay here and listen for a while, as I just discovered that the instructions I was following where for an older release and am now upgrading to a newer release.13:11
*** tcampbell has joined #openstack13:11
*** Shentonfreude1 has joined #openstack13:12
*** katkee has quit IRC13:16
*** shentonfreude has quit IRC13:16
*** Shentonfreude1 has quit IRC13:16
owenjust out of curiosity (I dont need to know yet) is anyone using openstack with IPv6, or is it and IPv4 only application stack ?13:17
*** Shentonfreude has joined #openstack13:17
*** ccc11 has joined #openstack13:18
*** nati has quit IRC13:18
*** huslage has joined #openstack13:23
*** sandywalsh has joined #openstack13:24
*** Oneiroi has joined #openstack13:25
*** Oneiroi has joined #openstack13:25
ttxowen: I didn't use it, but it supports Ipv613:26
*** Oneiroi_ has quit IRC13:27
*** hadrian has joined #openstack13:28
*** hingo has quit IRC13:28
owenthanks tty, it seems like. if we start to use open stack we will have to help improve documentation as it seems install guides are very IPv4 bias, :) open source commitments always seem to grow :)13:29
*** gnu111 has joined #openstack13:30
*** nati has joined #openstack13:30
owenBut I evaluate with IPv4 for now as Im still a beginner in IPv613:30
*** jtanner has joined #openstack13:31
*** bcwaldon has joined #openstack13:32
*** janOlder has joined #openstack13:32
*** jfluhmann has joined #openstack13:34
*** kernelfreak has quit IRC13:36
*** S_A has quit IRC13:39
*** ccc11 has left #openstack13:40
*** imsplitbit has joined #openstack13:43
*** msivanes has joined #openstack13:45
nerensHi, I'm getting an error starting nova-api while trying to intergrate with keystone...13:45
nerensImportError: No module named keystone.middleware.auth_token13:45
nerensHas anyone seen this?13:45
*** foxtrotdelta has joined #openstack13:47
*** rnorwood has joined #openstack13:47
*** kbringard has joined #openstack13:48
*** caribou has quit IRC13:50
*** caribou has joined #openstack13:53
*** mattray has joined #openstack13:53
*** bsza has quit IRC13:54
*** bsza has joined #openstack13:55
*** KnuckleSangwich has quit IRC13:55
*** ejat has joined #openstack13:56
*** ejat has joined #openstack13:56
*** nati has quit IRC13:56
gnu111nova-volume issues. I did this: euca-attach-volume vol-00000009 -i i-00000008 -d /dev/vdb. But when I ssh into my instance, fdisk -l don't know dev/vdb. what gives?13:57
btorchowen: openstack-swift works on ipv6 ... the python-eventlet needs to be eventlet-0.9.15 .. the previous version did not support it. I also tried 0.9.16 a while back but something broke on that version so I got stuck with 0.9.1513:57
*** mattray has quit IRC13:57
kbringardgnu111: kvm or xen?13:57
kbringardyou can't force a device on kvm13:58
kbringardcheck dmesg to see where it attached13:58
kbringardgenerally /dev/vdb is the "local disk" specified in the machine type13:58
kbringardalthough, it is still odd that fdisk says it can't find it13:58
gnu111pvdisplay says this:   PV Name               /dev/sda313:59
gnu111  VG Name               nova-volumes13:59
gnu111what should I look in dmesg?13:59
kbringardwell, that's on your volumes node13:59
kbringardyou need to look at dmesg on the instance you tried to attach it to13:59
gnu111ah, i see13:59
kbringardI would also check the nova-volume.log to make sure it actually carved out the lun and exported it properly14:00
*** jfluhmann has quit IRC14:00
gnu1112011-08-31 09:51:41,423 nova.volume.manager: volume volume-00000002: creating lv of size 7G14:00
gnu1112011-08-31 09:51:41,430 nova.utils: Running cmd (subprocess): sudo lvcreate -L 7G -n volume-00000002 nova-volumes14:00
gnu1112011-08-31 09:51:41,856 nova.volume.manager: volume volume-00000002: creating export14:00
kbringardok, so it looks like it should be there14:01
*** sandywalsh has quit IRC14:01
kbringardso i'd ssh into your instance and run dmesg | grep vd14:01
kbringardif you need to paste more than one line, you can do it here: to make sure you don't spam the channel14:01
*** Binbin has joined #openstack14:01
gnu111how can I make sure the attach command worked?14:02
owenbtourch: Thanks for the information on IPv6, I am still learning what IPv6 means to me, but I am on an IPv6 and IPv4 subnet of our site for evaluation purposes, as we shoudl like the scientists / users to be able to have inbound connectivity in the end so IPv6 is an important deciding feature, as some time in the next 3-4 years we will eventually run out of our IPv4 address hoard.14:02
kbringardif you ssh into your instance, and run dmesg | grep vd, or dmesg | tail −30 should probably show it14:03
kbringardyou should see a line about a new device being attached14:03
kbringardand it'll have it's block device attach point14:03
kbringardwhich, I would guess is probably /dev/vdc14:04
*** deshantm_laptop has joined #openstack14:04
kbringardkvm just attaches them in order, afaik14:04
kbringardalso, if you run euca-describe-instances i-0000000814:05
gnu111I only see vda in dmesg which is '/' in the vm.14:05
kbringardit should show the volume as attached14:05
kbringardor, euca-describe-volumes14:05
gnu111euca-describe does not show the volume info :(14:05
kbringardnothing in euca-describe-volumes either?14:06
gnu111I am on cactus did this: euca-attach-volume vol-00000002 -i i-00000041 -d /dev/vdb14:06
*** robbiew has joined #openstack14:06
kbringardoh, then euca-describe-instances i-0000004114:07
kbringarddoesn't show any volumes attached?14:07
gnu111nope. VOLUMEvol-00000002 7novaavailable2011-08-31T13:51:40Z14:07
gnu111Let me try the attach command again.14:07
kbringardhmm, then I'd start checking all the logs14:07
kbringardif there is nothing in the nova-volume.log14:08
*** Mr_T has joined #openstack14:08
kbringardthen I'd look at the nova-compute.log on the compute node the instance you're trying to attach it to is running on14:08
Raziquegnu111: ok I'm working on the doc regarding volumes D14:09
Raziquei'm pretty deep into it those days :D14:09
kbringardah, cool, Razique, you may be able to help better than I14:09
Raziquelaunch iscsiadm -m session on the compute node14:09
*** osier has quit IRC14:09
gnu111Razique: thanks. trying it now.14:09
*** katkee has joined #openstack14:09
kbringardI need more coffee14:10
*** mitchless has joined #openstack14:10
*** j05h has quit IRC14:10
*** mitchless has left #openstack14:10
Raziquekbringard: what time is it on your country ?14:10
gnu111from the compute nodes log: iscsiadm: cannot make connection to No route to host\niscsiadm: cannot make connection to it is using the private ip which I didn't add in the firewall14:10
kbringardit's 8:10am14:10
kbringardRazique: I live in Colorado, USA14:11
Raziquegnu111: is the ip of the server which run nova-volumes ?14:11
*** cp16net has joined #openstack14:11
Raziquekbringard: ah ok :D you woke up pretty early :D14:11
kbringardhehe, actually late ;-)14:11
kbringardI usually start at 7am Mountain14:11
*** nphase has joined #openstack14:11
kbringardbut my kids kept me up last night, so I only slept like 3 hours :-/14:12
gnu111Razique: Yes it is but I have two nic. eth0 and eth1. 172 is eth0. but for the nova-network, I am using the eth1 address.14:12
Raziquegnu111: have you set the iscsi flag in your nova.conf ?14:12
*** ldleworker has joined #openstack14:12
Raziquekbringard: erf14:13
kbringardRazique: it's cool, coffee is magic14:13
*** sandywalsh has joined #openstack14:14
Raziquegnu111: is 10.0 the eth1 network ?14:14
gnu111Razique: no. maybe that's the problem 10.0 is the nova-network for the bridge and dhcp. Shoudl this be the eth1 prefix?14:14
Raziquegnu111: the iscsi_ip_prefix defines the openstack admin network14:16
Raziquebasically the one openiscsi will use to perform sessions14:16
Raziquetry to change it to 172.29.20014:16
*** nphase has quit IRC14:17
*** bpaluch has joined #openstack14:18
*** Capashen has quit IRC14:18
*** bonzay has quit IRC14:19
gnu111Razique: same error: TRACE: Stderr: 'iscsiadm: cannot make connection to No route to host\niscsiadm:14:20
Raziquegnu111: what this IP corresponds to exactly ?14:20
*** LiamMac has joined #openstack14:21
gnu111this is the eth0 for the controller where I am running nova-network and nova-volume.14:21
*** Capashen has joined #openstack14:21
*** HowardRoark has joined #openstack14:22
Raziquedid you restarted the nova-compute ?14:22
Razique(after the changes)14:22
gnu111no. will try that.14:22
Raziqueservice nova-compute restart14:22
gnu111ah. I also didn't the add the isci prefix in the nova.conf for the compute node.14:23
*** mdomsch has joined #openstack14:23
*** HowardRoark has quit IRC14:24
Raziquegnu111: it only needs to be add for the compute node I think, but better to have it on both servers :)14:24
gnu111Razique: nope, same error. iscsiadm: cannot make connection to No route to host\niscsiadm:.14:26
gnu111I am in RHEL 6.1 btw.14:26
Raziqueok, try this on the compute node :14:27
Razique$iscsiadm -m discovery -t st -p
gnu111from the compute node, right?14:27
gnu111iscsiadm: cannot make connection to No route to hos14:28
Raziquegnu111: ok, ping it14:28
gnu11164 bytes from icmp_seq=1 ttl=64 time=0.228 ms14:28
gnu111do I need to open another port?14:28
Raziquegnu111: ok now, telnet it on the 3260 TCP14:29
Raziquetelnet 326014:29
gnu111that fails. telnet: connect to address No route to host14:29
Raziquethe weird thing is the "no route to host" message14:31
*** zz_bonzay has joined #openstack14:31
*** zz_bonzay is now known as bonzay14:31
Raziqueit's like both servers cannot communicate14:31
Raziquewould you show me the ifconfig on both server s ?14:31
owenimput [euca-bundle-image -i ${KERNEL_FILE_NAME} --kernel true] output [Invalid cert], seem to have problems following, anyone know where I can go from here ?14:32
*** siwos has quit IRC14:33
*** j05h has joined #openstack14:34
*** vladimir3p has joined #openstack14:35
Raziquegnu111: you said nova-volumes were installed on the controller ?14:35
Raziquegnu111: the ip is not affected to that controller obviously14:36
gnu111I think i tried the eth1 address as well.14:36
gnu111iscsiadm -m discovery -t st -p works from the controller.14:37
*** nphase has joined #openstack14:37
*** nphase has joined #openstack14:37
Raziquegnu111: you need to use the controller ip14:38
Raziquegnu111: To which server that IP is affected to ?14:39
Raziquethe controller ?14:39
nhmany guys from HP around?14:39
gnu111Razique: yes, 37 is the controller.14:40
creihtcan you get me a touchpad? :)14:40
Raziquegnu111: I don't see it in your file, could you run a $ip adds sh ?14:40
nhmcreiht: I know, I missed that deal by like an hour.14:41
Raziquegnu111: ok, it's mapped on bt1014:42
Raziquedoes the iscsi discovery works with that ? :14:42
Raziqueiscsiadm -m discovery -t st -p
Raziquefrom the compute node again14:43
gnu111nope same: iscsiadm: cannot make connection to No route to host14:43
Raziquegnu111: which network mode are you using ?14:44
Raziqueok, would you show me your nova.conf for the compute node ?14:44
gnu111ok. I think I found the issue. I didn't had this rule: iptables -I INPUT 1 -p tcp --dport 3260 -j ACCEPT14:45
gnu111I had it for udp.14:45
Raziquegnu111: ok, try again the telnet ?14:46
gnu111let me try to attach it again. Yes. telnet responds now.14:46
Raziquegnu111: you should better authorize the whole nova network14:47
Raziqueit's always useless to restrict via a firewall, and makes u work a pain sometimes14:47
*** code_franco has joined #openstack14:47
gnu111now a different error; from the compute node: Error: Could not find iSCSI export  for volume volume-0000000214:47
Raziquefrom which command ?14:48
gnu111i did euca-attach. this from the compute log14:48
Raziquewhat iscsiadmn -m session outputs now on the compute node ?14:49
owenAre most people using Ubuntu server to run open stack, or RHEL, which would be the better documented quick install ?14:49
Raziqueowen: I use ubuntu 10.04 server for the whole nova cloud14:50
gnu111iscsiadm: No active sessions.14:50
RaziqueI think it's the most documented since it's a free distro14:50
Raziquegnu111: ok, so basically, the discovery works, but not the session14:50
gnu111owen: I have it working with RHEL in cactus wth FlatDHCP and it is working fine but ubuntu is better documented.14:50
Raziquetry to use the same network into u nova.conf file that the once that you use for the iscsi discovery14:51
owenYes we by default use Scientific linux as we know the maintainers (they are payed by sister labs) but I think I will move over to ubunu if its easier at least to start with.14:51
owenunfortuantely means I have to give up for  the rest of the day as its others jobs to install base OS14:52
owenbut pleased to know some others are using a RHEL compatible version14:52
*** nati has joined #openstack14:53
*** heckj has joined #openstack14:53
gnu111Razique: this looks better: nova.volume.driver: Found iSCSI node /dev/disk/by-path/ (after 1 rescans)14:54
*** amccabe has joined #openstack14:54
Raziqueguigui1: ok, it means the session has been established :) try now dmesg | tail on the instance14:54
*** cereal_bars has joined #openstack14:54
Raziqueyou should see a /dev/vdb device14:54
gnu111cool! I can see the dev/vdb from the instances with fdisk -l! :D14:54
gnu111Disk /dev/vdb doesn't contain a valid partition table14:55
Raziquegnu111: hehe, glad you made it14:55
Raziquenow frisk /dev/vdb14:55
*** deshantm_laptop has quit IRC14:55
Raziqueand put the filesystem you'd like on it14:55
gnu111Razique: thanks for the help. I really appreciate it. There are so many details, it is easy to loose track.14:55
Raziquegnu111: you welcome, for your info, i'm working on the nova-volumes section, and finishing a script which will allow the auto backup of all the volumes14:56
*** Mr_T has left #openstack14:56
gnu111that will be helpful.14:56
Raziqueanother will allow to reattach the volumes to the instances in case your whole nova architecture restarts14:56
uvirtbotNew bug: #838154 in nova "LXC instance without injected key doesn't get created" [Undecided,New]
Raziquethe nova-volumes turn to become into a stalled state after a restart, which could be a pain to fix14:57
*** jedi4ever_ has joined #openstack14:57
gnu111ok. got my dev/vdb1 mounted! woo hoo!!14:58
*** jaypipes has joined #openstack14:58
*** jedi4ever has quit IRC14:58
*** jedi4ever_ is now known as jedi4ever14:58
*** rsampaio has quit IRC14:59
*** rsampaio has joined #openstack15:01
gnu111what's the difference between nova-volume and EBS?15:02
*** rsampaio has quit IRC15:06
*** reidrac has quit IRC15:06
*** dragondm has joined #openstack15:07
kbringardgnu111: nova-volume is the OpenStack implementation of EBS, as far as I'm aware15:08
*** ig has quit IRC15:09
gnu111Can I use a nfs mount in nova-volume? For the test, I just create a lvm in the controller for now. But we have some shared storage which probably be used for the volume.15:11
*** deshantm_laptop has joined #openstack15:11
*** andy-hk has joined #openstack15:13
Raziquegnu111: you totally can :) be aware that NFS perf ain't that cool15:14
*** HowardRoark has joined #openstack15:14
RaziqueI use myself an iscsi storage for nova-volumes15:14
*** dgags has joined #openstack15:14
*** shaibn has joined #openstack15:14
*** bcwaldon has quit IRC15:14
kbringardso you're iscsi exporting and iscsi export, nice15:14
*** rsampaio has joined #openstack15:15
Raziquekbringard: not, it's even worse15:15
*** bcwaldon has joined #openstack15:15
Raziquethe controller is a vm itself, which runs on a kvm server15:15
Raziquethat VM is stored on a SAN15:15
Raziqueso I can live-migrate my nova-controller in case something goes wrong15:16
Raziquegnu111: a quick hint, use the flag ,errors=remount-ro15:16
Raziqueso if you have an iscsi loss of connectivity, (that happens,) you ensure that data don't get corrupted (that happens too)15:17
gnu111cool, thanks!15:17
*** rsampaio_ has joined #openstack15:18
RaziqueI remember myself trying to recover data due to LVM corruption, it's the worst thing that happened to me since I use nova15:18
gnu111Razique: did you trying booting off from these volumes?15:18
Raziquegnu111: the EBS booting has been recently implemented into nova, haven't tried yet15:19
*** rsampaio has quit IRC15:19
RaziqueI generally mount the ebs vol on /home15:19
Raziqueso the main data is on the vol. not the instance15:19
*** mattray has joined #openstack15:23
*** mattray has quit IRC15:23
*** rsampaio_ has quit IRC15:27
*** javiF has joined #openstack15:28
*** rsampaio has joined #openstack15:30
*** seoman__ has quit IRC15:31
*** nerens has quit IRC15:34
*** jedi4ever has quit IRC15:39
*** misheska has joined #openstack15:40
*** Gordonz has joined #openstack15:40
*** jedi4ever has joined #openstack15:40
*** bonzay has quit IRC15:43
*** seoman__ has joined #openstack15:44
*** guigui1 has quit IRC15:44
*** phschwartz has joined #openstack15:44
*** dobber has quit IRC15:44
*** deshantm_laptop has quit IRC15:45
*** zz_bonzay has joined #openstack15:46
*** zz_bonzay is now known as bonzay15:46
heckjgnu111:EBS mounts as a filesystem only - it's a block device through ISCSI protocol. You can have an instance export NFS if you need the shared san storage, but your performance is likely to be pretty poor15:46
*** nouse has joined #openstack15:46
*** nphase has quit IRC15:48
*** nphase has joined #openstack15:48
*** nphase has joined #openstack15:48
*** nphase has quit IRC15:50
*** Capashen has quit IRC15:51
jtannerwhat flag/param/arg for nova defines the range passed to dnsmasq in flatdhcp?15:52
jtannerall my configs are set to give out 192.168.10.x but dnsmasq is running and trying to give out 10.0.0.x15:52
heckjjtanner: what did you run when you created your network with "nova-manage network create"? That's where the data gets set...15:54
jtannerheckj, nova network create 1 25515:55
heckjIt uses those flags, but defaults to - so if you did that prior to setting the flags in nova.conf, the database will have incorrect values15:55
jtannerheckj, nova network create 1 25515:55
heckjdo you have a -flat-dhcp-start flag?15:56
jtannerheckj, i do not15:56
heckjjtanner: I suspect that might be what you want to add15:56
*** nati has quit IRC15:56
* heckj looks for specific example15:56
jtannerheckj, what table would store those addresses?15:56
*** llang629 has joined #openstack15:57
jtannerfixed_ips has the correct range/ips15:57
*** nati has joined #openstack15:57
heckjjtanner: here's the flags for FlatDHCP:
heckjjtanner: specifically "--flat_network_dhcp_start"15:58
heckjjtanner: don't recall the schema - would need to go trolling in it to recall/find out.15:58
jtanneraha, it's in the networks table15:58
heckjthe data in there look OK?15:58
jtannerit has my correct range, but there's a in one of the fields ... trying to figure out which15:59
jtanneralways a pita on mysql command line15:59
heckjthat's probably the DHCP start - default is
jtannerok, dhcp_start is
jtanneryou are right16:00
*** rocambol1 has quit IRC16:01
*** nati has quit IRC16:01
jtannerheckj, if i edit the conf and restart nova network, will it update that field or do i need to do that by hand?16:01
*** huslage has quit IRC16:02
*** shaibn has quit IRC16:03
heckjjtanner: I'm afraid it doesn't update the data - that has to be done by hand :-(16:03
jtannerk, will do16:03
*** kernelfreak has joined #openstack16:03
*** mdomsch has quit IRC16:05
*** Razique has quit IRC16:06
jtannerheckj, actually it seems to have updated the field properly16:06
heckjjtanner: nice! I didn't think it did!16:06
jtannerusing cactus on the stackops distro16:07
*** jedi4ever has quit IRC16:07
*** caribou has quit IRC16:07
jtannerdnsmasq is looking better too16:07
*** skane has joined #openstack16:08
*** troys has joined #openstack16:10
jtannerheckj, everything is working (instances pull IPs), thanks for your help!16:13
heckjjtanner: Glad to assist!16:14
heckjjtanner: When diablo is out and ready to roll, check out the settings for multi_host - I think it's an incredible improvement!16:14
WormManit's sure.... fun :)16:14
jtannerwhat is the difference?16:15
* WormMan continues poking D4 with a pointy stick16:15
heckjWormMan - you using it now? I installed it last week on a test cluster16:15
heckjjtanner: nova-network runs on all the compute node hosts, so you don't end up having all incoming&outgoing traffic routing through a single bottleneck node (nova-network instance)16:15
jtanneri have diablo-3 running too, i'm just using the stackops/cactus setup to help me understand the basics and fix my diablo instance16:15
WormManheckj: working on setting up filtering so I can use the same gateway IP on every private network node, we're likely to have a ton of low powered compute nodes, so the 1 IP per compute node is a bit troublesome16:16
jtannerheckj, ah, understood16:16
*** ecarlin has joined #openstack16:16
*** st4n has quit IRC16:16
heckjWormMan: yeah, a downside - but I think a great interim step until Quantum is really in form and here (which is looking pretty soon). I've seen a demo of quantum with Open Vswitch, but haven't done it myself as yet16:17
*** ecarlin has quit IRC16:17
WormManwe need to implement in the next couple weeks, so I'm trying not to add too many new 'features'16:17
heckjWormMan: gotcha - how many nodes are putting into the cluster?16:18
WormManour eventual number looks like 120 or so, but that's a bit further out16:18
WormManthat will shrink as the old crappy hardware is replaced with new, high density hardware16:19
heckjWormMan: yeah, you'll be eating IP addresses like candy. VLan Networking or DHCP?16:19
*** tudamp has left #openstack16:19
heckjYeah, the multi_host will eat an IP/node for every project you have on it - that hurts.16:19
*** troys has quit IRC16:20
WormManwe still need 1 IP on the public side for NAT, but that's less annoying16:20
heckjQuantum can't come soon enough...16:20
*** katkee has quit IRC16:22
WormManlooks like I only need about 4 ebtables rules and then 2 more per VLAN16:22
*** mdomsch has joined #openstack16:22
*** maplebed has joined #openstack16:22
*** maplebed has quit IRC16:23
*** maplebed has joined #openstack16:23
*** dmshelton has joined #openstack16:25
*** nati has joined #openstack16:26
*** jdurgin has joined #openstack16:28
kbringardheckj: amen brother16:30
*** nyeates has joined #openstack16:36
*** joearnold has joined #openstack16:38
*** vernhart has quit IRC16:40
*** nati has quit IRC16:41
*** nati has joined #openstack16:41
*** lvaughn_ has quit IRC16:41
*** lvaughn has joined #openstack16:42
*** lvaughn_ has joined #openstack16:43
*** ahasenack has quit IRC16:45
*** nati has quit IRC16:46
*** lvaughn has quit IRC16:46
*** seoman__ has quit IRC16:47
*** owen has quit IRC16:48
*** jedi4ever has joined #openstack16:49
*** bradhedlund has quit IRC16:55
*** bradhedlund has joined #openstack16:55
*** nati has joined #openstack16:57
*** cpg|away is now known as cpg17:00
*** obino has quit IRC17:01
*** llang629 has left #openstack17:05
*** foods is now known as adam_g17:05
*** eviscares has quit IRC17:08
*** berto- has joined #openstack17:09
*** eviscares has joined #openstack17:09
*** Ryan_Lane has joined #openstack17:10
*** berto- has quit IRC17:11
*** tjikkun has quit IRC17:12
*** tjikkun has joined #openstack17:13
*** ccorrigan has quit IRC17:13
*** ahasenack has joined #openstack17:14
*** vladimir3p has quit IRC17:16
*** rms-ict has quit IRC17:17
*** berto- has joined #openstack17:18
*** k has joined #openstack17:18
*** k is now known as Guest3256817:19
*** adjohn has joined #openstack17:19
*** irahgel has left #openstack17:20
Guest32568What do open stack deployments use for image management for Host OS images, storage nodes etc...   ?17:20
*** mattrobinson has joined #openstack17:21
*** vladimir3p has joined #openstack17:21
uvirtbotNew bug: #838251 in nova "OSAPI create server and get server details response doesn't contain list of security groups" [Undecided,In progress]
*** tcampbell has quit IRC17:22
*** darraghb has quit IRC17:22
*** javiF has quit IRC17:28
*** obino has joined #openstack17:30
*** obino has quit IRC17:32
*** obino has joined #openstack17:32
*** bcwaldon has quit IRC17:32
*** nati has quit IRC17:33
*** bradhedlund has quit IRC17:33
*** nati has joined #openstack17:34
*** bradhedlund has joined #openstack17:34
*** nati has quit IRC17:38
*** Hektor has joined #openstack17:39
HektorI need some advice please :D17:39
HektorHi koolhead17:40
HektorWe just bought 4 SuperMicro servers and want to put openstack on it. Our primary use would be to set up VMs for our customers.17:40
HektorWhat we would like to know is how to start. We found tons of manuals but no luck finding what we need17:41
tjoydid you look at
tjoythere's a quick-start guide and an installation manual17:42
Hektorwe read it multiple times. We even wanted to try stackops17:42
*** cpg is now known as cpg|away17:42
*** nacx has quit IRC17:42
tjoyso where are you stuck?17:42
HektorWell, thre are some manuals for installing openstack on 2 servers, and we have 4.17:44
tjoyrepeat the steps for the second server (the compute node) 3 times17:44
Hektorso u recommend installing controler + 3 computing nodes right?17:44
tjoyyou will not want a second control node until you have a special need for it17:45
tjoywhich you probably won't17:45
HektorI see. We read that stackops needs network node. whats network node. Do we need that as well?17:46
nyeatesStackops, from what i recall, can also take into acct multiple servers - it has you distribute around what you want each servers role to be17:46
tjoyso the network controller is a service that runs between the compute nodes and the outside network17:46
*** vernhart has joined #openstack17:46
tjoymost people run it on the control box for simplicity17:46
nyeatesand one server can do multiple things btw17:46
HektorSo network node would be waste17:47
HektorAlso, they have volume node for images. Is that something we need or it can be implemented in computing node?17:48
tjoyso compute, volume, network, api, etc, are individual programs17:48
nyeatesA seperate network node, may be a waste. You still need its services to run on some box though I beleive.17:48
*** ejat has quit IRC17:48
tjoythey can run wherever you want17:48
*** gnu111 has left #openstack17:48
Hektorhow do u mean : wherever I want?17:49
tjoywherever you choose to put them, they will run.17:49
tjoywhether that's effective is another matter17:49
tjoyif you want to cram them all in one server, fine17:49
tjoyif you want one server per service, fine17:49
HektorWould u recommend it?17:50
tjoyrecommend what17:50
nyeatesI would definetely recommend stackops install - it steps you through some of this visually17:50
WormMansigh, on the plus side openstack is open source so I can look at the code... on the minus side, now I have to look at the code :)17:50
*** winston-d has quit IRC17:51
tjoyHektor: most people put all the services they want on one node, and then compute on any more nodes17:51
HektorWould u recommend storing images on computing nodes instead of volume node like in stackops17:51
*** DavorC has joined #openstack17:51
heckjHektor: I prefer storing the images in Swift - but that means having either a swift-all-in-one tucked away for just that, or a full-out-swift cluster (5 hosts)17:52
HektorI see. so the best thing for me is to put all services on one node and rest to use for computing?17:52
tjoyto begin with, yes17:53
*** nati has joined #openstack17:53
tjoyif you decide you need something else, it will be in the course of setting it up like you said.17:53
*** nickon has joined #openstack17:53
Hektorgreat. I was starting to think that 4 servers is not enough17:53
tjoyif you bought too many you can give some to me17:53
Hektorsure :D17:54
tjoyi won't mind17:54
Hektornow I have 2 spare computing servers17:54
HektorOk. Next thing is : what we need for the network side. Switches , routers etc17:55
Hektor? :D17:55
*** cpg|away is now known as cpg17:55
tjoyhow fast are your disks?17:55
*** dysinger1 has quit IRC17:55
tjoydid you spring for 10kRPM ones?  SSDs?17:55
tjoydo you have a SAN?17:55
Hektorno standard 2TB wd disks17:56
tjoy7200RPM then?17:56
tjoygigabit switches are fine17:56
tjoyyou may want to bond ports on the control machine17:56
tjoyand put in extra NICs17:56
tjoybut you don't have to17:56
HektorSorry for my English, but how do u mean bond ports on the control machine?17:57
tjoyturn 2 1-gigabit ports into 1 2-gigabit port17:57
tjoygoogle for 'port bonding'17:57
Hektoraha ok. I can do that17:57
tjoyor if you can put 10gE on the control machine and have a switch that will accomodate that, then good for you17:58
heckjHektor: If you have the capability, use a switch that supports VLAN's and use the VLAN networking option of OpenStack.17:58
tjoythat's really pushing it17:58
tjoy'supports vlans' is sort of a funny way to say it17:59
Hektorhows that?17:59
tjoyHektor: in vlan mode, every port is an 802.1q port, and every packet is tagged17:59
tjoythe NIC is the switch17:59
*** Guest32568 has quit IRC18:00
Hektorok. what else do I need for this "cloud" to work (except magic)18:00
tjoyyou should be good to start wrecking things now18:00
tjoyhave fun18:01
*** joesavak has joined #openstack18:01
HektorI will. Thank u very much :D18:01
*** jsavak has joined #openstack18:03
DavorCIf I was to install a new OpenStack cluster, would people recommend using Cactus or a Diablo-4 milestone?18:05
*** joesavak has quit IRC18:06
tjoyit depends on what features you want18:08
*** rms-ict has joined #openstack18:08
*** sandywalsh_ has joined #openstack18:08
*** Alowishus has left #openstack18:08
*** jfluhmann has joined #openstack18:09
*** skane has quit IRC18:10
*** sandywalsh has quit IRC18:11
*** grapex has joined #openstack18:12
*** bcwaldon has joined #openstack18:14
*** bradhedlund has quit IRC18:15
*** Hektor has quit IRC18:15
*** bradhedlund has joined #openstack18:15
*** jamesturnbull has quit IRC18:18
*** adjohn has quit IRC18:21
DavorCtjoy: reliability and stability would be good, I would like to put this into production use within a few months18:21
tjoyi think diablo is scheduled for release in september18:22
*** jaypipes has quit IRC18:22
*** adjohn has joined #openstack18:23
DavorCtjoy: I saw Sept 22 as the release date on Launchpad.18:24
tjoyif you can get D4 working in testing, you'll probably be up to speed by the time it drops18:25
tjoythen you can reinstall everything18:25
DavorCI would prefer not to have to reinstall everything, it would be nice if the upgrade "just worked".18:26
tjoythat would be nice, wouldn't it18:26
*** cereal_b_ has joined #openstack18:26
DavorCDo you think that's unlikely?18:26
tjoyon general principles, yes, but i don't have any special knowledge of openstack to say yes or no18:27
*** bengrue has joined #openstack18:28
tjoythat's a good question for someone who's on the project18:28
*** cereal_bars has quit IRC18:29
tjoyi'd say if you install the upcoming feature-freeze release, you might avoid some certain issues18:29
tjoybut that's no reason not to start working with diablo before that18:29
tjoyi'm sure the fundamentals will stay the same18:29
*** adjohn has quit IRC18:33
*** Eyk^off has joined #openstack18:40
*** tryggvil_ has quit IRC18:41
kbringardis there a flag I need to set to get swift-proxy to log?18:41
kbringardI have it installed and (I thought?) configured, but it's just dying when I start it, and i'm not getting any output anywhere that I can find18:42
kbringardI'm following this doc:
*** dgags has quit IRC18:42
*** OutBackDingo has quit IRC18:45
*** AhmedSoliman has joined #openstack18:45
creihtkbringard: it should log to syslog by default18:51
creihtyou can also try running18:51
*** dirkx__ has joined #openstack18:51
creihtswift-proxy-server /etc/swift/proxy-server.conf18:51
creihtThat should run it in the console18:51
*** seoman__ has joined #openstack18:52
kbringardah, OK, nice, thanks creiht18:52
kbringardI'll check it out and see if I can figure out wtf I'm doing stupid18:52
*** mattray has joined #openstack18:53
creihtkbringard: if it truns out that the proxy fails, but isn't logging a message will you file a bug?18:53
*** ejat has joined #openstack18:53
*** ejat has joined #openstack18:53
creihtwe tried to make it so that it catches all of those, but it may still miss something on daemon startup18:53
*** nati has quit IRC18:54
kbringardfor some reason it pasted all on one line though… sorry18:55
*** jamesturnbull has joined #openstack18:56
*** seoman__ has quit IRC18:56
creihtkbringard: yeah the tracebacks are one line in the logs18:56
kbringardthat's my conf, copied almost straight from the docs :-/18:58
creihthah... looks like the error was too long for the log line18:58
creihtbut it is probably conf related18:58
kbringardyea, I figured, which is why I pasted you my conf18:58
kbringardlike I said, I basically just copied it from the doc, so I'm sure I'm missing something important18:59
creihtkbringard: can you run the proxy from the command line like I posted earlier? and then paste the resulting traceback18:59
creihtto see if we can get the full traceback18:59
kbringardsure, one sec19:00
creihtkbringard: what version of swift are you using?19:01
kbringardno trace there :-/19:01
creihthah... it must have posted it to logs19:01
*** GeoDud has joined #openstack19:01
*** adjohn has joined #openstack19:01
creihttempauth wasn't in 1.219:01
kbringardah so19:01
creihtif you are going to continue to user 1.219:01
creihterm 1.219:01
kbringarddeb lucid main19:02
creihtwill give you 1.2 specific docs19:02
kbringardthat's the repo I pulled from19:02
* creiht sighs19:02
kbringardI'm happy to use a newer version19:02
kbringardin fact, I'd probably prefer it19:02
*** syah_ has joined #openstack19:02
creihtI thought we fixed the docs to point to the right repo19:03
*** syah has quit IRC19:03
creihtif you use19:03
kbringardadd-apt-repository ppa:swift-core/ppa19:03
kbringardthat's what I did (from the doc)19:04
creihtthat should be swift-core/release19:04
kbringardOK, cool19:04
kbringardlet me update it and try again19:04
creihtsorry about that :/19:04
kbringardno worries19:04
creihtannegentle: -^ ?19:04
kbringardhmmm, I don't seem to have permissions to update the doc, so I'll let someone else do it :-D19:04
creihtkbringard: yeah I'll make sure we get it fixed :)19:05
kbringardcool, no worries, I'll run through my stuff again and let you know in a few19:05
*** grapex has quit IRC19:05
*** mdomsch has quit IRC19:06
*** heckj has quit IRC19:08
*** aliguori has quit IRC19:08
*** grapex has joined #openstack19:09
*** aliguori has joined #openstack19:09
*** mdomsch has joined #openstack19:12 ... omg, lightbulbs19:14
creihtjtanner: the current docs always follow trunk... We have pointers to the older versions as well19:14
creihtthus the /1.2 /1.3 for swift19:15
*** tcampbell has joined #openstack19:15
*** Eyk^off is now known as Eyk19:16
notmynamecreiht: kbringard: ppa doc fix merged into trunk19:21
kbringardw00t, thanks19:22
*** phschwartz has quit IRC19:24
*** phschwartz has joined #openstack19:26
*** cpg is now known as cpg|away19:26
kbringardah, that worked much better, the proxy is running now19:27
kbringardah, but, there looks to be a dependency problem now19:27
kbringardwith auth19:28
*** syah_ has left #openstack19:28
WormManhrm, ok, with the new multi-host stuff, when an instance requests meta-data it gets forwarded to the API server and natted(otherwise it can't get back to the instance) of course, API is giving a 404 since it doesn't recognize the IP19:30
*** cereal_b_ has quit IRC19:30
WormManwhat's the canonical way to solve this?19:31
WormManahh, the API will take x-forwarded-for so I could setup a small proxy I guess19:32
annegentlejtanner: where are lightbulbs? :)19:32
jtannerin my head19:32
jtanneri was trying to use cactus docs with diablo-319:33
annegentlejtanner: ah! there you go!19:33
annegentlejust put the existence of trunk in a blog entry today, sorry for the confusion19:34
*** tjikkun has quit IRC19:34
*** berto- has quit IRC19:35
jtannerannegentle, i only blame myself19:36
*** worstadmin has joined #openstack19:37
kbringardcreiht: any idea on that deps issue with swift-auth?19:38
kbringardshould I file a bug, or is that something y'all are aware of?19:38
*** huslage has joined #openstack19:38
kbringardor am I doing something dumb?19:38
*** katkee has joined #openstack19:38
*** cereal_bars has joined #openstack19:39
*** tryggvil_ has joined #openstack19:39
*** worstadmin has quit IRC19:40
*** AhmedSoliman has quit IRC19:41
*** nerens has joined #openstack19:43
creihtkbringard: I would try removing all the swift-* packages maybe?19:44
*** tcampbell_ has joined #openstack19:44
kbringardcreiht: OK, but this is on a fresh machine, just preseeded19:45
kbringardI installed swift, proxy and then tried to do auth19:45
*** tcampbell has quit IRC19:45
*** tcampbell_ is now known as tcampbell19:45
creihtI don't know as much about how the packaging is set up19:45
*** nickon has quit IRC19:45
kbringardoh, you know what19:46
kbringardI think it is my bad19:46
kbringardyea, this is my fault, sorry19:46
creihthehe ok19:46
*** tjikkun has joined #openstack19:48
DavorCAny Grid Dynamics folks here?19:54
*** seoman__ has joined #openstack20:01
kbringarddid swift-auth get changed to swift-account?20:03
btorchI think swift-auth was a 1.2 thing . I believe20:04
kbringardah, it did, stupid me reading old docs20:04
*** cereal_bars has quit IRC20:16
*** jaypipes has joined #openstack20:18
*** katkee has quit IRC20:24
*** mies has quit IRC20:25
*** mdomsch has quit IRC20:28
*** nati has joined #openstack20:29
*** tryggvil_ has quit IRC20:31
*** nerdstein has quit IRC20:32
*** nati has quit IRC20:34
*** jano has joined #openstack20:37
*** javiF has joined #openstack20:38
*** med_out is now known as medberry20:38
*** janOlder has quit IRC20:39
*** Razique has joined #openstack20:39
*** jsavak has quit IRC20:39
RaziqueHI all20:40
Raziquehas someone ever resize his nova-volumes' disks ?20:40
uvirtbotNew bug: #838372 in nova "ImageService should not map unknown attributes to metadata" [Undecided,In progress]
*** nyeates has quit IRC20:42
*** bradhedlund has quit IRC20:43
*** bradhedlund has joined #openstack20:44
*** joesavak has joined #openstack20:44
*** jano has quit IRC20:45
*** huslage has quit IRC20:51
*** javiF has quit IRC20:51
*** Eyk is now known as Eyk^off20:52
*** berto- has joined #openstack20:54
*** huslage has joined #openstack20:56
*** bradhedlund has quit IRC21:00
*** bradhedlund has joined #openstack21:00
*** joesavak has quit IRC21:01
uvirtbotNew bug: #838386 in nova "Test suite requires kombu and carrot to be installed" [Undecided,New]
WormManwoo, it's looking good for my horrific combination of iptables and split horizon routing to be able to only give 1 IP to all the nova-network nodes21:13
*** bcwaldon has quit IRC21:15
*** huslage has quit IRC21:24
*** nerdstein has joined #openstack21:24
*** cpg|away is now known as cpg21:25
*** dirkx__ has quit IRC21:26
uvirtbotNew bug: #838399 in nova "xenserver agent calls getting lost" [Undecided,New]
*** OutBackDingo has joined #openstack21:27
*** pothos_ has joined #openstack21:27
*** pothos has quit IRC21:29
*** seoman__ has quit IRC21:29
*** pothos_ is now known as pothos21:30
*** lts has quit IRC21:32
*** jaypipes has quit IRC21:33
*** Razique has quit IRC21:37
*** cp16net has quit IRC21:38
*** rms-ict_ has joined #openstack21:40
*** msivanes has quit IRC21:40
*** rms-ict has quit IRC21:41
*** nerdstein has quit IRC21:42
*** foxtrotdelta has quit IRC21:42
*** bengrue has quit IRC21:44
*** bengrue has joined #openstack21:44
*** imsplitbit has quit IRC21:46
*** sunny_ has quit IRC21:54
*** kbringard has quit IRC21:57
*** mfer has quit IRC22:00
uvirtbotNew bug: #838419 in nova "nova-compute crashes when applying a security group rule" [High,New]
*** LiamMac has quit IRC22:05
*** code_franco has quit IRC22:06
*** code_franco has joined #openstack22:07
jtanneris it just me, or is nova-manage broken for griddynamics cactus repo? "Possible wrong number of arguments supplied", log shows that it passes unexpected keyword argument timeout22:09
*** misheska has quit IRC22:09
*** vernhart1 has joined #openstack22:12
*** vernhart has quit IRC22:12
*** woleium has joined #openstack22:15
*** tryggvil_ has joined #openstack22:17
*** rnorwood has quit IRC22:19
*** ldleworker has quit IRC22:19
*** code_franco has quit IRC22:21
*** HowardRoark has quit IRC22:21
*** PeteDaGuru has left #openstack22:21
*** bradhedlund has quit IRC22:22
*** bradhedlund has joined #openstack22:22
*** jaypipes has joined #openstack22:23
*** medberry is now known as med_out22:23
*** ncode has quit IRC22:26
*** nerdstein has joined #openstack22:26
DavorCjtanner: where, and calling which function? I had to *add* a 'timeout' argument to the "wait" method in eventlet/green/subprocess.py22:27
*** robbiew has quit IRC22:28
*** po has quit IRC22:28
*** jj0hns0n has quit IRC22:29
*** jj0hns0n_ has joined #openstack22:29
*** amccabe has quit IRC22:33
vishyjtanner: it is a bug in centos/redhat library22:34
*** negronjl has quit IRC22:36
*** amccabe has joined #openstack22:38
*** tomh has joined #openstack22:38
*** tomh_ has quit IRC22:39
*** nerdstein has quit IRC22:41
*** tcampbell has quit IRC22:41
*** negronjl has joined #openstack22:43
*** ejat has quit IRC22:44
*** amccabe has quit IRC22:44
*** vladimir3p_ has joined #openstack22:46
*** vladimir3p has quit IRC22:48
*** mattray has quit IRC22:48
*** cdbs has quit IRC22:50
*** nerdstein has joined #openstack22:51
*** jfluhmann has quit IRC22:57
*** lorin1 has joined #openstack22:58
*** HowardRoark has joined #openstack23:00
*** jaypipes has quit IRC23:00
*** ncode has joined #openstack23:00
*** jaypipes has joined #openstack23:02
*** lorin1 has quit IRC23:02
*** wilmoore has joined #openstack23:03
*** worstadmin has joined #openstack23:05
*** titaniumrain has quit IRC23:08
*** viraptor has quit IRC23:09
*** bsza has quit IRC23:10
*** worstadmin has quit IRC23:10
*** nerens has quit IRC23:10
*** worstadmin has joined #openstack23:10
*** andy-hk has quit IRC23:12
*** primeministerp1 has left #openstack23:22
*** cpg is now known as cpg|away23:22
*** rsampaio has quit IRC23:30
*** rsampaio has joined #openstack23:32
*** DavorC has quit IRC23:39
*** miclorb has joined #openstack23:42
*** nerdstein has quit IRC23:44
*** miclorb has quit IRC23:44
*** jeffjapan has joined #openstack23:47
*** grapex has quit IRC23:50
*** npmapn has quit IRC23:50
*** bobracecarbob has joined #openstack23:51
*** wilmoore has quit IRC23:51
*** bengrue has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!