Saturday, 2014-01-25

SpamapSdkehn: I have to agree. Seattle is a much nicer experience.00:01
SpamapSSunnyvale feels like a trip to the suburbs. :-P00:02
dkehnclarkb: I think I've been there in a past life, it is in the middle of nowwhere00:02
dkehnSpamapS: there is always Ft. Collins, lol00:02
dkehnspeaking of in the middle of nowhere00:02
*** cd-undercloud has joined #tripleo00:04
cd-undercloud************** overcloud complete status=1 ************00:04
*** cd-undercloud has quit IRC00:04
SpamapSdkehn: actually they have better beer than sunnyvale :)00:05
SpamapSinteresting00:05
dkehnSpamapS: Seattle or Ft. Collins00:05
SpamapSso we got our nodes stuck in "REBUILD"00:05
SpamapSdkehn: Ft. Collins.00:05
dkehnSpamapS: yes, but so does Seattle00:05
SpamapSYeah, bellies will be happier in Seattle for sure.00:06
cody-somervilleAre we talking about Sprint location?00:08
*** boris-42 has quit IRC00:19
dkehncody-somerville: yepper00:27
*** sballe has quit IRC00:30
*** epim has joined #tripleo00:31
*** epim has quit IRC00:31
NobodyCamSpamapS: look at your first comment on https://review.openstack.org/#/c/66461/1/elements/nova-ironic/os-refresh-config/configure.d/81-nat-metadata00:34
*** UtahDave has quit IRC00:34
NobodyCamFail early, fail often is the device?00:34
*** CaptTofu has joined #tripleo00:40
SpamapSNobodyCam: my point there is to not use --type raw.. like.. ever. :)00:40
NobodyCam:) ack ok :)00:41
SpamapSNobodyCam: netdevice will cause an earlier failure if you try to put something that is clearly not a netdevice in for that field.00:41
SpamapSNobodyCam: oh also while looking at that, '' is not an option, so you should not have a --key-default00:42
SpamapSNobodyCam: '' will produce an error below00:42
NobodyCamok will test and have a patch up soonish00:43
NobodyCamjust have to work out how to address ghe's comment :-p00:44
*** jayg is now known as jayg|g0n300:53
*** dkehn has quit IRC00:56
*** dkehn has joined #tripleo00:58
SpamapSinteresting.. our boxes are wedged in REBUILD00:59
SpamapSERROR: Instance 61436e6d-864a-44ab-904a-dd77e57a616e in task_state rebuild_spawning. Cannot stop while the instance is in this state. (HTTP 400) (Request-ID: req-9d688ba2-1d0d-462d-b632-580644524989)01:00
SpamapSugh01:09
SpamapSdoesn't seem to be much we can do if a REBUILD gets wedged like this01:09
SpamapScan't stop.. can't force rebuild to start over01:09
*** CaptTofu has quit IRC01:15
*** CaptTofu has joined #tripleo01:18
lifelessSpamapS: task state is cleared if you restart nova-compute, I think.01:21
lifelessSpamapS: but even so - bug filing time ;)01:21
SpamapSyeah01:21
SpamapSI started with this https://bugs.launchpad.net/tripleo/+bug/127259501:21
lifelessI'll look later :) gotta run again, or will miss movie01:22
SpamapSlifeless: actually even weirder, if you nova delete them .. they go back to active01:22
SpamapSbut they are _not_ active :-P01:22
SpamapSok .. not easy to get those deleted01:39
*** cwolferh has quit IRC01:42
SpamapShttps://bugs.launchpad.net/nova/+bug/1272600 filed in Nova01:45
*** greghaynes has quit IRC01:46
*** greghaynes has joined #tripleo01:55
*** CaptTofu has quit IRC01:56
*** flashgordon is now known as jog001:57
*** amrit has quit IRC02:21
*** cd-undercloud has joined #tripleo02:48
cd-undercloud************** overcloud complete status=1 ************02:48
*** cd-undercloud has quit IRC02:48
*** beekneemech has quit IRC03:21
*** bnemec has joined #tripleo03:25
*** taps has quit IRC03:44
*** CaptTofu has joined #tripleo03:52
*** cd-undercloud has joined #tripleo03:55
cd-undercloud************** overcloud complete status=1 ************03:55
*** cd-undercloud has quit IRC03:55
*** coolsvap has quit IRC04:08
*** rwsu has quit IRC04:25
*** ramishra has joined #tripleo04:48
*** cd-undercloud has joined #tripleo05:04
cd-undercloud************** overcloud complete status=1 ************05:04
*** cd-undercloud has quit IRC05:04
*** greghaynes has quit IRC05:05
*** e0ne has quit IRC05:07
*** greghaynes has joined #tripleo05:10
*** CaptTofu has quit IRC05:12
*** CaptTofu has joined #tripleo05:13
*** ramishra has quit IRC05:14
*** AaronGr is now known as AaronGr_Zzz05:16
*** ramishra has joined #tripleo05:17
*** CaptTofu has quit IRC05:18
*** ramishra_ has joined #tripleo05:23
*** ramishra has quit IRC05:23
*** ramishra_ has quit IRC05:32
*** jprovazn has quit IRC05:50
*** jprovazn has joined #tripleo05:51
*** ramishra has joined #tripleo05:54
*** ramishra_ has joined #tripleo06:03
*** tzumainn has quit IRC06:05
SpamapSweird.. things seem totally screwed on undercloud nova now06:06
*** ramishra_ has quit IRC06:06
SpamapS| fault                                | {u'message': u'{"error": {"message": "object of type \'NoneType\' has no len()", "code": 400, "title": "Bad Request"}}', u'code': 500, u'details': u'  File "/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/compute/manager.py", line 259, in decorated_function |06:06
*** ramishra has quit IRC06:06
SpamapSseems like there is some kind of cleanup going on06:06
*** ramishra has joined #tripleo06:06
mordredSpamapS: doesn't seem like clean to me06:10
SpamapSyeah its weird06:10
SpamapSit is getting some kind of unauthorized error while trying to start instances06:10
*** ramishra has quit IRC06:11
*** cd-undercloud has joined #tripleo06:11
cd-undercloud************** overcloud complete status=1 ************06:11
*** cd-undercloud has quit IRC06:11
SpamapSmordred: what is frustrating is that nova now won't delete these instances06:15
*** edmund has quit IRC06:16
mordredSpamapS: wow06:29
mordredSpamapS: I believe I've heard of this06:29
mordredbut I don't know what's up06:29
SpamapSjust tried upgrading neutron to latest.. no help06:29
*** coolsvap has joined #tripleo06:34
*** coolsvap has quit IRC06:34
*** ramishra has joined #tripleo06:35
*** akuznetsov has joined #tripleo06:39
lifelessSpamapS: oh hai06:40
SpamapSlifeless: any insight into how to unbreak nova?06:43
lifelessSpamapS: whats up06:43
SpamapSI'm trying to find where neutron or keystone are raising this NoneType has no len error.06:43
lifelesslogging in now to poke06:44
SpamapSneither is logging it06:44
lifelessinstances have no ports06:44
SpamapSright, they're getting refused auth by keystone06:45
SpamapSor rather, nova is06:45
SpamapSat /tokens06:45
lifelesshere's your problem06:46
lifeless  File "/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/virt/baremetal/db/sqlalchemy/api.py", line 152, in bm_node_get_by_instance_uuid06:46
lifeless    raise exception.InstanceNotFound(instance_id=instance_uuid)06:46
lifelessInstanceNotFound: Instance 84c6090b-bf42-4c6a-b2ff-afb22b5ff156 could not be found.06:46
SpamapS| 84c6090b-bf42-4c6a-b2ff-afb22b5ff156 | overcloud-NovaCompute0-scls5xj54oym    | ERROR  | deleting   | NOSTATE     |                       |06:47
SpamapSit never got assigned to a bm node06:47
lifelessjust looking up the bug for you06:47
SpamapSis this where if we fail to get a bm node, we can't do "neutron things" ?06:48
lifelessit should be in https://bugs.launchpad.net/nova/+bugs?field.tag=baremetal06:48
lifelessthere may be multiple issues06:48
lifelessbut this one06:48
lifelessis bug 117758406:48
lifelessI think06:48
lifelesswell, same umbrella06:49
SpamapSso is the fix to assign it manually to a bm node and try to delete it again?06:50
lifelessits the reverse of https://bugs.launchpad.net/nova/+bug/122634206:50
lifelessfiling a bug06:50
lifelessnah, lets fix it.06:50
SpamapSlifeless: this started after we updated to latest nova06:50
SpamapSlifeless: I thought it was the REBUILD getting them wedged but now I think maybe it is just that deleting bm nodes might be badly broken.06:51
SpamapSor maybe even provisioning them06:51
SpamapSsince I had errors before I tried to delete06:51
SpamapSsame neutronclient failing to auth errors06:51
*** ramishra has quit IRC06:52
*** ramishra has joined #tripleo06:52
lifelessI think we're in a cascade fail situation06:53
lifelessone thing masking / causing another06:53
lifelessanyhow06:53
SpamapSagreed :(06:54
SpamapSlifeless: could be an old keystone bug06:54
lifelesshttps://bugs.launchpad.net/nova/+bug/127262306:54
lifelessoh this is in init_host06:55
lifelessfailing to startup06:55
lifelessI'm sure I filed this before06:55
lifelessah yes06:55
lifelessbug 116861006:55
lifelessdifferent trace06:55
lifelesssame cause I think06:55
lifelessnew title on the bug I filed06:56
SpamapSlifeless: ok, I'm out of juice.. going to sleep now.06:58
lifelessk06:58
lifelessI'll fix and monkey that on06:58
SpamapSstopped tripleo-cd06:58
SpamapSgoing to buy some champagne for the next time it status=0's twice in a row06:58
lifelessSpamapS: right, past that, onto te unauthorized07:04
lifelessSpamapS: what else did you alter?07:06
rpodolyakamorning tripleo07:11
*** ramishra has quit IRC07:13
*** CaptTofu has joined #tripleo07:14
*** ramishra has joined #tripleo07:14
lifelessSpamapS: http://lists.openstack.org/pipermail/openstack/2013-November/003478.html07:15
*** CaptTofu has quit IRC07:18
*** ramishra has quit IRC07:19
*** ramishra has joined #tripleo07:26
*** akuznetsov has quit IRC07:37
lifelesstried git keystoneclient, no change07:48
lifelesstried git neutronclient, no change07:50
*** ramishra has quit IRC07:50
SpamapSlifeless: git keystone maybe?07:51
*** ramishra has joined #tripleo07:51
lifelessSpamapS: maybe07:51
* SpamapS should really be asleep :-P07:51
lifelessSpamapS: what had you tried ?07:51
SpamapSlifeless: I'm looking at len() calls in keystone at the commit we have now..07:51
SpamapSlifeless: oh I should have said, I tried git neutron07:51
SpamapSnothing like dragging ye olde undercloud into the future ;)07:51
lifelessoh god, it's going to be a disaster07:52
SpamapShm with keystone the token backend len() calls are all guarded against None..07:53
*** ramishra has quit IRC07:55
lifelesshttps://bugs.launchpad.net/python-keystoneclient/+bug/1257541 looked suspiciously relevant07:55
lifelesswhich is why I tried git python-keystoneclient07:55
lifelessoh god07:57
lifelessI'm blind07:57
lifelesshang on while I make neutronclient better07:57
SpamapS:)07:57
lifelessI wish python had SSA07:57
SpamapSno idea what that is.. :)07:58
lifelesssingle static assignment07:58
lifelesslook at the variable body in neutronclient/client.py line 23207:59
*** ramishra has joined #tripleo07:59
lifelessSpamapS: https://review.openstack.org/6911008:01
SpamapSlifeless: seems legit..08:01
SpamapSlifeless: ok this time I really am going to go lay my head on a pillow and sleep.08:03
lifeless{'auth': {'tenantName': None, 'passwordCredentials': {'username': None, 'password': None}}}08:05
lifelessthat seems broken08:05
*** ramishra has quit IRC08:06
lifelessI smell thread local breakage again08:07
lifelessIMBW08:08
lifelessahhh08:08
lifelessI bet they dropped the quantum backwards compat code08:08
*** akuznetsov has joined #tripleo08:09
lifelessyup08:09
*** ramishra has joined #tripleo08:09
lifelesstripleo-cd restarted08:11
*** ramishra has quit IRC08:11
lifelesscommit ee50436938eb8122185b08681803d07f16dcd3c108:12
lifelessfrom jog0 :(08:12
lifelessanyhow08:13
lifelessfixed08:13
*** lsmola_ has quit IRC08:17
*** openstack has joined #tripleo08:22
*** ramishra has joined #tripleo08:26
*** ramishra has quit IRC08:30
*** csd has quit IRC09:04
*** e0ne has joined #tripleo09:06
*** CaptTofu has joined #tripleo09:15
*** cd-undercloud has joined #tripleo09:17
cd-undercloud************** overcloud complete status=1 ************09:17
*** cd-undercloud has quit IRC09:17
*** CaptTofu has quit IRC09:19
*** e0ne has quit IRC09:21
*** max_lobur_afk is now known as max_lobur10:02
*** cd-undercloud has joined #tripleo10:24
cd-undercloud************** overcloud complete status=1 ************10:24
*** cd-undercloud has quit IRC10:24
*** viktors has joined #tripleo10:30
*** akuznetsov has quit IRC10:31
*** e0ne has joined #tripleo10:32
*** e0ne has quit IRC10:34
*** boris-42 has joined #tripleo10:49
*** spzala has quit IRC10:56
*** CaptTofu has joined #tripleo11:15
*** CaptTofu has quit IRC11:20
*** ramishra has joined #tripleo11:28
openstackgerritDirk Mueller proposed a change to openstack/tripleo-incubator: Fix openSUSE detection  https://review.openstack.org/6860211:28
*** cd-undercloud has joined #tripleo11:32
cd-undercloud************** overcloud complete status=1 ************11:32
*** cd-undercloud has quit IRC11:32
openstackgerritDirk Mueller proposed a change to openstack/tripleo-incubator: Setup libvirt/kvm permissions on openSUSE  https://review.openstack.org/6879411:32
*** ramishra has quit IRC11:34
*** ramishra has joined #tripleo11:40
*** ramishra_ has joined #tripleo11:43
*** ramishr__ has joined #tripleo11:45
*** ramishra has quit IRC11:45
*** ramishr__ has quit IRC11:45
*** ramishra_ has quit IRC11:46
*** ramishra has joined #tripleo11:46
*** ramishra has quit IRC11:51
*** ramishra has joined #tripleo11:52
*** coolsvap has joined #tripleo11:54
*** ramishra has quit IRC11:57
*** panda has joined #tripleo12:06
*** panda__ has quit IRC12:09
*** coolsvap has quit IRC12:30
*** coolsvap has joined #tripleo12:32
*** cd-undercloud has joined #tripleo12:39
cd-undercloud************** overcloud complete status=1 ************12:39
*** cd-undercloud has quit IRC12:39
*** CaptTofu has joined #tripleo12:53
*** CaptTofu has quit IRC12:57
*** rpodolyaka has quit IRC12:59
*** CaptTofu has joined #tripleo13:04
*** coolsvap has quit IRC13:33
*** akuznetsov has joined #tripleo13:42
*** cd-undercloud has joined #tripleo13:45
cd-undercloud************** overcloud complete status=1 ************13:45
*** cd-undercloud has quit IRC13:45
*** akuznetsov has quit IRC13:47
*** michchap has quit IRC13:58
*** michchap has joined #tripleo13:58
*** michchap has quit IRC14:03
*** CaptTofu has quit IRC14:13
*** CaptTofu has joined #tripleo14:14
*** CaptTofu has quit IRC14:18
*** cd-undercloud has joined #tripleo14:51
cd-undercloud************** overcloud complete status=1 ************14:51
*** cd-undercloud has quit IRC14:51
*** CaptTofu has joined #tripleo14:51
*** ccrouch has quit IRC15:01
*** ccrouch has joined #tripleo15:02
*** ccrouch has quit IRC15:06
SpamapShm15:12
SpamapSgetting NoValidHost for nova boots now15:12
*** viktors has left #tripleo15:29
*** cody-somerville has quit IRC15:30
*** blamar has quit IRC15:45
*** cody-somerville has joined #tripleo15:45
*** akuznetsov has joined #tripleo15:46
*** cd-undercloud has joined #tripleo15:58
cd-undercloud************** overcloud complete status=1 ************15:58
*** cd-undercloud has quit IRC15:58
*** e0ne has joined #tripleo15:59
*** e0ne has quit IRC16:12
*** openstack has joined #tripleo16:19
*** CaptTofu has quit IRC16:26
*** CaptTofu has joined #tripleo16:27
*** CaptTofu has quit IRC16:31
*** max_lobur is now known as max_lobur_afk16:38
*** michchap has joined #tripleo16:40
*** michchap has quit IRC16:45
*** rpodolyaka1 has joined #tripleo16:46
*** cd-undercloud has joined #tripleo17:04
cd-undercloud************** overcloud complete status=1 ************17:04
*** cd-undercloud has quit IRC17:04
*** e0ne has joined #tripleo17:23
*** e0ne has quit IRC17:26
*** michchap has joined #tripleo17:41
*** michchap has quit IRC17:45
greghaynesSpamapS: in filter_scheduler.py?17:47
*** rpodolyaka1 has quit IRC18:04
*** e0ne has joined #tripleo18:06
*** cd-undercloud has joined #tripleo18:10
cd-undercloud************** overcloud complete status=1 ************18:10
*** cd-undercloud has quit IRC18:10
*** noslzzp has quit IRC18:11
lifelessmorning18:34
lifelessnova list18:35
lifeless: f9b7d1a3-e057-4efa-b20c-8dee27234a29] Error from last host: undercloud (node a9402861-306e-4f49-96af-789c1fab5762): [u'Traceback (most recent call last):\n', u'  File "/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/compute/manager.py", line 1058, in _build_instance\n    set_access_ip=set_access_ip)\n', u'  File "/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/compute/manager.py", line 357, in decorated_f18:38
lifelesswheee sorry18:38
*** coolsvap has joined #tripleo18:39
lifelessgreghaynes: SpamapS: http://paste.openstack.org/show/61887/18:39
greghaynesaha, my derpd overcloud is unrelated18:40
lifelessto manually fix this one:18:40
lifelessselect * from bm_nodes where uuid='a9402861-306e-4f49-96af-789c1fab5762'\G18:41
lifelessthen check the instance_uuid18:41
lifelesswhich is usually a deleted instance18:41
lifelessthen update the instance uuid18:41
*** coolsvap has quit IRC18:42
*** michchap has joined #tripleo18:43
SpamapSlifeless: so this is a baremetal bug?18:43
lifelessSpamapS: maybe or maybe a side effect of the other bug18:43
lifelessanyhow I'm going to cross reference the dbs and reconcile18:43
SpamapSlifeless: It's times like these where I think we need to think hard if we can just switch to Ironic, warts and all, so that we can actually move forward with fixing bugs.18:46
SpamapSIIRC there is just one neutron thing left and then we can reasonably think about doing that.18:46
*** michchap has quit IRC18:47
lifelessSpamapS: I think we should as soon as it has the features18:58
lifelessSpamapS: theres chris and ghe and deva ft on it atm18:58
lifelessSpamapS: I'm not worried about warts18:58
SpamapSlifeless: feels a lot like when juju python was "the one that works" but we couldn't actually improve it. :(18:59
*** cody-somerville has quit IRC18:59
lifelessSpamapS: well, there are some differences - we've been improving nova-bm as needed19:00
lifelessSpamapS: so we're not blocked, we have a protocol for signalling that a new feature is needed for ironic too19:00
*** coolsvap has joined #tripleo19:00
SpamapSlifeless: right, but time spent on baremetal is wasted. :-/19:00
SpamapSlifeless: I rewrote the entire juju local provider after juju python was deprecated. Fixed most of the bugs, but felt empty. :-/19:01
lifelessSpamapS: I don't entirely agree that its wasted19:01
lifelesswhats the goal we're optimising for19:01
devanandathere's 2 things still missing from ironic... links coming ...19:01
lifelessdoes it help that goal ?19:01
lifelessI don't mind us doing thing two or even three times19:01
lifelessif we learn something19:02
devanandahttps://review.openstack.org/#/c/66071/19:02
lifelessand help something19:02
lifelessalong the way19:02
devanandahttps://review.openstack.org/#/c/51328/919:02
devanandaneutron binding & nova driver19:02
* lifeless blinks at the config setting changes there19:03
devanandaauto-cnofig foo. blame oslo for that19:04
lifelessno19:04
lifelessstuff shouldn't be in there19:04
lifelessoh19:04
lifelessyou're saying its just rearranged.19:04
devanandayes19:04
lifelessI see - but deploy_ramdisk=<None>19:04
lifelessI just wrote a review comment of19:04
lifelessI don't think we should put this in the config file at all - it was a mistake in nova baremetal, and it would be a mistake to replicate it. flavor works /much/ better.19:04
lifelessand for19:04
lifeless623#net_config_template=$pybasedir/ironic/net-dhcp.ubuntu.template19:05
lifelessDraft saved at 8:03 AM19:05
lifelessWe're not doing file injection, so this isn't needed.19:05
GheRiverohi all!19:05
*** e0ne has quit IRC19:05
devanandalifeless: oh, i thought you were talking abbout the neutron patch's conf changes19:06
lifelessdevananda: I thought I was19:06
devanandalifeless: i agree re: net_config_template. but i dont see a reason to remove teh capability to have a default deploy_ramdisk19:06
lifelessdevananda: because it doesn't make any sense19:07
devanandaneutron patch conf changes: https://review.openstack.org/#/c/66071/11/ironic/common/neutron.py19:07
GheRiveronet_config_template is not used anywhere in the code now19:07
GheRiveroso +1 to remove it19:08
devanandaGheRivero: have time this week to do a cleanup of pxe.py?19:08
GheRiverosure19:09
devanandagreat, thanks much19:09
devanandathere are bunch of TODO you left in there, too :)19:09
devanandaat least we should revisit if they are still valid TODO19:09
GheRiveroI should have removed my name from the TODOs19:09
lifelessdevananda: it doesn't make sense because unlike nova-bm ironic is multi-arch19:10
devanandaeh, anyone can do them. it's more a record of who thought it should be done in the first place19:10
lifelessdevananda: having a default there is a great way to utterly break people that start with amd64 and add an i386 machine19:10
lifelessdevananda: because they will then be going 'wtf' 'wtf' until they learn they had been using a special case feature for $however long19:11
devanandalifeless: ironic is multi-driver19:11
lifelessdevananda: it also forces config file updates when the deploy ramdisk is updated, which is much harder than API calls for deployers19:11
lifelessdevananda: yes, but there is one 'pxe' default there, but pxe can deploy to multiple architectures19:12
lifelessdevananda: no?19:12
devanandaeg. for deploying to windows, the pxe driver will need to run slightly different deploy process19:13
devanandanot jst different deply ramdisk19:13
lifelessdevananda: I'm not talking about operating system19:13
lifelessdevananda: same os, same deploy driver, different cpu architecture19:13
devanandasure19:13
devanandaso your point is valid and not the complete answer19:14
devanandaa default ramdisk has the potential to confuse pppl, i agree19:14
devanandamind filing a bug? I've got to run19:14
devanandait's saturday ;)19:14
lifelesssunday :)19:14
lifelessfiling19:14
devanandaty19:15
SpamapSsunday bloody sunday19:15
*** cd-undercloud has joined #tripleo19:17
cd-undercloud************** overcloud complete status=1 ************19:17
*** cd-undercloud has quit IRC19:17
lifelessnow, stack popped, where was I19:22
lifelessthats right, reconciling19:22
lifelessselect uuid, instance_uuid from bm_nodes where instance_uuid not in (select uuid from nova.instances where not deleted);19:23
lifeless^ useful in future19:23
lifeless| f9b7d1a3-e057-4efa-b20c-8dee27234a29 | test2                                  | BUILD  | deleting   | NOSTATE     | ctlplane=10.10.16.171 |19:25
lifelesshmm19:25
lifelessthat might have found a bad machine19:25
lifelessanyhow19:25
lifelesslets see what happens now19:25
*** coolsvap has quit IRC19:29
*** michchap has joined #tripleo19:43
*** michchap has quit IRC19:43
*** michchap has joined #tripleo19:43
*** jcooley_ has joined #tripleo19:47
*** michchap has quit IRC19:48
*** jcooley_ has quit IRC19:55
*** jcooley_ has joined #tripleo19:55
*** jcooley_ has quit IRC20:00
*** jcooley_ has joined #tripleo20:00
*** jcooley_ has quit IRC20:04
*** jcooley_ has joined #tripleo20:05
*** panda_ has joined #tripleo20:17
*** panda has quit IRC20:20
*** jcooley_ has quit IRC20:21
*** jcooley_ has joined #tripleo20:21
*** cd-undercloud has joined #tripleo20:24
cd-undercloud************** overcloud complete status=1 ************20:24
*** cd-undercloud has quit IRC20:24
lifelessok, instances stuck in spawning now20:25
lifelessgreghaynes: you said you're having a filter problem?20:26
lifelessgreghaynes: do yo want to debug that ?20:26
*** jcooley_ has quit IRC20:27
greghaynesI have to afk for a bit but when I return yesplz :)20:28
lifelessping me, if I'm here I'll be delighted to help20:28
greghayneswill do20:28
*** panda_ has quit IRC20:29
*** e0ne has joined #tripleo20:31
lifelessyay mellanox flexboot20:33
lifelesslets see if I managed to hit the interactive jaackpot20:34
lifelessSpamapS: tcpdumping br-ctlplane shows 20:35:16.903058 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 78:e7:d1:23:93:dd, length 38120:36
lifeless20:35:18.880251 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 78:e7:d1:23:93:dd, length 38120:36
lifelesse.g. no response from neutron20:36
lifelessI've restarted the dhcp agent20:36
lifelessSpamapS: and am now in the ipxe command line manualyl triggering dhcp to see wtf is20:39
lifelessup20:39
lifelessnext step is tcpdump in the netns20:39
*** cd-undercloud has joined #tripleo20:40
cd-undercloud************** overcloud complete status=1 ************20:40
*** cd-undercloud has quit IRC20:40
lifelessthe tap device there isn't seeing any traffic20:40
lifelesswhich suggests a wrong or misconfigured openvswitch config20:41
*** michchap has joined #tripleo20:43
lifelesstraffic isn't hitting br-int AFAICT20:43
*** michchap has quit IRC20:44
lifelessSpamapS: ah, I think you broke us with git neutron20:44
*** michchap has joined #tripleo20:44
*** panda_ has joined #tripleo20:44
lifelessSpamapS: rolled neutron back, booting proceeding accordingly20:46
*** michchap has quit IRC20:48
lifeless20:50
lifelessTrying to load: pxelinux.cfg/01-78-e7-d1-23-93-dd                   ok20:50
lifelessLoading /tftpboot/7e032507-93f9-434f-bbba-1bbb8fd02bb9/deploy_kernel...20:50
lifelessfurther along20:50
*** cd-undercloud has joined #tripleo20:50
cd-undercloud************** overcloud complete status=1 ************20:50
*** cd-undercloud has quit IRC20:50
lifelessand dd is off and running20:58
lifelessoh nuts20:58
lifeless2014-01-25 20:54:47,600.600 13789 DEBUG nova.openstack.common.processutils [req-eae1f4fd-e4fc-4b85-94a4-06700e5cf4c3 None None] Running cmd (subprocess): sudo nova-rootwrap /etc/nova/rootwrap.conf mkfs -t ext3 -F -L ephemeral0 /dev/disk/by-path/ip-10.10.16.171:3260-iscsi-iqn-7e032507-93f9-434f-bbba-1bbb8fd02bb9-lun-1-part1 execute /opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/openstack/common/processutils.py:14720:58
lifelesslost the ext4 patch along the way20:59
lifelessSpamapS: ^ we have to do it with config now...20:59
clarkbabout that20:59
lifelessclarkb: yo21:00
clarkbI don't know that the argument was made, but doesn't ext3 have a much greater potential for losing your data in a variety of situations common to cloud environments?21:00
lifelessclarkb: not really21:00
lifelessclarkb: 3 and 4 both journal21:00
clarkbhrm I thought that came up with upstart not nicely syncing on shutdown21:00
lifelessclarkb: if you care about data, you should set ordered to both21:00
lifelessclarkb: *4* had early bugs that were terrible when combined with powerloss or hard off situations21:01
clarkbI remember that21:01
*** akuznetsov has quit IRC21:02
*** akuznetsov has joined #tripleo21:03
clarkbI find the default of ext3 to be annoying because all of our rackspace nodes come with ext3 filesystems. Would be nice to see it changed21:03
*** akuznetsov has quit IRC21:03
lifelessyup21:03
lifelessclarkb: I feel your pain21:09
clarkbI know21:09
lifelessclarkb: you can online convert to ext421:09
lifelessext4 so much faster :)21:11
clarkband you can have more dirs in dir :)21:11
lifelesshmm?21:11
clarkbext3 has a limit of 32k subdirectorys in one directory. ext4 doubles that21:11
lifeless.. isn't a hardlink anymore?21:11
lifelessoh, just a different compile time setting :>21:12
lifelessand yay, one node active21:12
clarkbapparnetly the numbers are related to performance things. It became important on the logs server21:12
lifelesstwo nodes active21:12
clarkbwe worked around it by remounting ext4 which didn't do a conversion but did bump the limit21:13
lifelessso each subdir has a child .. which is a hardlink to it's parent21:13
lifelessthe actual thing that changed is how many hardlinks a single inode can have21:13
clarkbah21:13
lifelessSpamapS: ok, tripleo-cd stopped, overcloud deleted, tripleo-cd started21:14
clarkbbut all of the elasticsearch data is on ext3 and so are the git repos. I am sure there would be noticeable impacts if we were ext4 for all of that stuff21:14
lifelessfsck is faster if you have extents21:15
lifelessbut you need to convert (tune2fs enable extents)  and then rewrite everything to migrate the metadata for files to fix thaat21:16
lifelesslate allocation is (AIUI) entirely in kernel memory, so remounting gets you that21:16
lifelessbut writing the thing is again faster if you have extents21:16
lifelessright, time to go do some shopping21:18
lifelessbbiaw21:18
*** michchap has joined #tripleo21:44
*** michchap has quit IRC21:44
*** michchap has joined #tripleo21:45
SpamapSlifeless: we landed the config patch but have not refreshed nova config yet :-P21:49
*** michchap has quit IRC21:50
lifelessdone by hand21:52
SpamapSlifeless: \o/ :)21:52
SpamapSWould be great if we could somehow get ironic to facilitate 'nova console-log' :)21:55
SpamapSAssuming ipmi serial console is a thing we can leave on without angering the gods.21:55
lifelessthere is a BUG21:57
SpamapSsshd conn refused21:58
SpamapSlifeless: you looking at the consoles?21:58
*** michchap has joined #tripleo21:58
lifelessno21:58
lifeless2014-01-25 21:48:34,679.679 4148 INFO nova.virt.baremetal.deploy_helper [-] deployment to node 11 done21:58
SpamapSlifeless: used fallback data source22:00
SpamapSlifeless: so ssh keys not installed22:01
lifelessrestart nova-api ?22:01
lifelessor log in on console and check metadata access?22:01
SpamapSyeah trying to do that22:01
SpamapSno getty on tty122:01
SpamapSor on the virtual serial22:02
SpamapSlifeless: worked fine for the compute node22:03
SpamapSor not22:04
SpamapSweird22:04
SpamapSok saturday beckons22:04
SpamapSlifeless: things certainly still a little ????? around here :-P22:04
* SpamapS goes back to afk for a while22:05
lifelessok, so I'll debug metadata server later22:06
*** boris-42 has quit IRC22:16
*** cd-undercloud has joined #tripleo22:21
cd-undercloud************** overcloud complete status=1 ************22:21
*** cd-undercloud has quit IRC22:21
*** ccrouch has joined #tripleo22:37
*** panda has joined #tripleo22:56
*** panda_ has quit IRC22:59
*** cd-undercloud has joined #tripleo23:27
cd-undercloud************** overcloud complete status=1 ************23:27
*** cd-undercloud has quit IRC23:27

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!