Wednesday, 2014-05-28

*** matsuhashi has joined #openstack-ceilometer00:01
*** shakamunyi has quit IRC00:18
*** ddieterly has joined #openstack-ceilometer00:25
*** _cjones_ has quit IRC00:42
*** _cjones_ has joined #openstack-ceilometer00:42
*** fnaval has joined #openstack-ceilometer00:45
*** _cjones_ has quit IRC00:47
*** liusheng has quit IRC00:55
*** _cjones_ has joined #openstack-ceilometer01:13
*** _cjones_ has quit IRC01:18
*** _cjones_ has joined #openstack-ceilometer01:23
*** liusheng has joined #openstack-ceilometer01:23
*** _cjones_ has quit IRC01:30
*** thomasem has quit IRC01:36
*** zqfan has joined #openstack-ceilometer01:37
*** thomasem has joined #openstack-ceilometer01:46
*** thomasem has quit IRC01:46
*** nosnos has joined #openstack-ceilometer01:48
*** thomasem has joined #openstack-ceilometer01:49
*** thomasem_ has joined #openstack-ceilometer01:52
*** ddieterly has quit IRC01:52
*** ddieterly has joined #openstack-ceilometer01:53
*** thomasem has quit IRC01:53
*** changbl has joined #openstack-ceilometer01:57
*** fnaval has quit IRC02:01
*** fnaval has joined #openstack-ceilometer02:02
*** rbowen has quit IRC02:09
*** nati_uen_ has quit IRC02:13
*** raymondr has joined #openstack-ceilometer02:16
*** raymondr has quit IRC02:20
*** thomasem_ has quit IRC02:27
zqfanhi, I came across a problem: AttributeError: 'StaticFileMiddleware' object has no attribute 'application', when I run tox -epy27 in latest ceilometer master branch, I already clear all old pyc files with `find ./ceilometer -name *.pyc | xargs -i rm -f {}`, and compare my py27 env with jenkins success test env with `pip freeze`, and nothing different(except ceilometer head)02:42
*** thomasem has joined #openstack-ceilometer02:50
*** thomasem has quit IRC02:51
*** shakamunyi has joined #openstack-ceilometer03:19
*** thomasem has joined #openstack-ceilometer03:22
*** shakamunyi has quit IRC03:24
*** fnaval has quit IRC03:24
*** _cjones_ has joined #openstack-ceilometer03:26
*** _cjones_ has quit IRC03:31
*** fnaval has joined #openstack-ceilometer03:32
*** fnaval has quit IRC03:33
*** fnaval has joined #openstack-ceilometer03:33
liushengHi folks, when I search around ceilometer-notification-agent, I fond that if I delete a non-exist floatingip, neutron will emit a "delete.start" notification, is it reasonable?03:43
*** thomasem has quit IRC03:44
*** nosnos has quit IRC03:47
*** shakamunyi has joined #openstack-ceilometer03:52
*** `jpg has joined #openstack-ceilometer04:00
*** matsuhashi has quit IRC04:15
*** `jpg has quit IRC04:16
*** raymondr has joined #openstack-ceilometer04:17
*** raymondr has quit IRC04:21
*** psharma has joined #openstack-ceilometer04:22
*** matsuhashi has joined #openstack-ceilometer04:26
*** nosnos has joined #openstack-ceilometer04:27
*** shakamunyi has quit IRC04:27
zqfanI think it is not reasonable, neutron should return 404 directly04:30
*** raymondr_ has joined #openstack-ceilometer04:40
*** shakamunyi has joined #openstack-ceilometer04:52
*** shakamunyi has quit IRC04:54
*** raymondr_ has quit IRC05:05
*** shakamunyi has joined #openstack-ceilometer05:12
*** ddieterly has quit IRC05:22
*** ddieterly has joined #openstack-ceilometer05:22
*** alexpilotti has quit IRC05:26
*** _cjones_ has joined #openstack-ceilometer05:28
*** _cjones_ has quit IRC05:32
*** shakamunyi has quit IRC05:40
*** shakamunyi has joined #openstack-ceilometer05:41
*** ildikov has quit IRC05:43
*** renlt has joined #openstack-ceilometer05:43
*** fnaval has quit IRC05:52
*** alexpilotti has joined #openstack-ceilometer05:54
*** alexpilotti has quit IRC05:59
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ceilometer: Imported Translations from Transifex
*** _nadya_ has joined #openstack-ceilometer06:06
*** alexpilotti has joined #openstack-ceilometer06:08
*** ildikov has joined #openstack-ceilometer06:13
*** _nadya_ has quit IRC06:28
*** _cjones_ has joined #openstack-ceilometer06:28
*** _nadya_ has joined #openstack-ceilometer06:30
*** _cjones_ has quit IRC06:33
*** _nadya_ has quit IRC06:38
*** boris-42 has quit IRC06:45
*** inc0 has joined #openstack-ceilometer06:49
*** boris-42 has joined #openstack-ceilometer06:50
*** renlt has quit IRC06:51
*** anand_ts has joined #openstack-ceilometer06:51
anand_tshi all, a silly doubt, I have set an alarm for a cirros instance cpu_util, whenever threshold goes above 20 for 3 consecutive cycles of 600s it should trigger an alarm. but may I know where and how I get the alert.?06:54
*** `jpg has joined #openstack-ceilometer06:58
*** ddieterly has quit IRC07:06
*** ddieterly has joined #openstack-ceilometer07:06
*** _nadya_ has joined #openstack-ceilometer07:07
*** Ruetobas has quit IRC07:14
*** underyx is now known as Underyx07:15
*** inc0 has quit IRC07:21
*** Ruetobas has joined #openstack-ceilometer07:22
*** _nadya_ has quit IRC07:26
*** Ruetobas has quit IRC07:27
*** Ruetobas has joined #openstack-ceilometer07:27
*** _cjones_ has joined #openstack-ceilometer07:29
*** idegtiarov has joined #openstack-ceilometer07:31
openstackgerritliusheng proposed a change to openstack/ceilometer: Improve the timestamp validation of ceilometer API
*** _cjones_ has quit IRC07:34
anand_tshello, what does insufficient data mean here
*** melodous_away has joined #openstack-ceilometer07:41
*** melodous has quit IRC07:43
*** Underyx is now known as underyx07:51
zqfan<anand_ts> insufficient data means there is no enough data to judge the alarm's state08:07
zqfan<anand_ts> for i.e. when you just create a new alarm, it is insufficient; when ceilometer is broken or the depended resource is broken, then no data is record, the state is insufficient; when everything is fine, just no new record is stored, then it is still insufficient (for i.e. alarm on instance meter but there is no vm)08:11
*** `jpg has quit IRC08:14
*** matsuhashi has quit IRC08:14
*** matsuhashi has joined #openstack-ceilometer08:15
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: [HBase] Implement events on HBase   In this change events are written into HBase and read from it.
*** matsuhashi has quit IRC08:17
*** matsuhashi has joined #openstack-ceilometer08:17
*** eglynn has joined #openstack-ceilometer08:17
*** IvanBerezovskiy has joined #openstack-ceilometer08:27
*** inc0 has joined #openstack-ceilometer08:28
*** nacim has joined #openstack-ceilometer08:32
*** melodous_away has quit IRC08:36
*** r0j4z0 has quit IRC08:36
*** shakamunyi has quit IRC08:43
*** `jpg has joined #openstack-ceilometer08:43
*** ddieterly has quit IRC08:49
*** ddieterly has joined #openstack-ceilometer08:50
*** pnascimento has joined #openstack-ceilometer08:50
*** isviridov|away is now known as isviridov08:53
anand_tszqfan, thanks for reply, here I have a cirros instance running fine, all no error in ceilometer logs. still why the state is insufficient. Does it mean that I can't able to check ceilometer if the instance is a cirros instance?08:53
eglynnanand_ts: the fact that the instance is booted from a cirros image is unlikely to be the cause08:59
eglynnanand_ts: is there a ceilometer compute agent running on the same host as the instance?09:00
eglynnanand_ts: can you check the logs of that compute agent?09:00
anand_tsyes, ceilometer-compute agent is running on my compute node09:00
eglynnanand_ts: can you run $ ceilometer statistics -m cpu_util -p 600 -q "resource=$INSTANCE_ID"09:01
anand_ts ceilometer-compute log09:01
anand_tsoksy sure eglynn09:01
anand_tseglynn. output of statistics here
*** ildikov_ has joined #openstack-ceilometer09:03
eglynnanand_ts: ok so that's telling you that the statistics are available, which is good :)09:03
*** ildikov has quit IRC09:04
eglynnanand_ts: so, next thing is to check if the ceilometer-alarm-evaluator service is running09:04
anand_tsyes it is running09:04
eglynnanand_ts: next check the evaluator's logs09:04
anand_tsceilometer-alarm-evaluator in controller node is it. I have only compute agent in compute nodes09:05
eglynnanand_ts: sure the alarm-evaluator doesn't have to run on the compute node09:05
eglynnanand_ts: ... controller node is fine for that09:05
anand_ts  no error in evaluator09:06
openstackgerritIonut Artarisi proposed a change to openstack/ceilometer: reconnect to mongodb on connection failure
eglynnanand_ts:  alarm f526313e-a76a-4064-bea9-e8e589e9ff06 transitioning to ok because Transition to ok due to 1 samples inside threshold, most recent: 9.0317195325509:06
eglynnanand_ts: $ ceilometer alarm-show -a f526313e-a76a-4064-bea9-e8e589e9ff0609:07
eglynnanand_ts: ... ^^^ is that the alarm you're interested in?09:07
anand_tsyeah, alarm show here
anand_tsstatus is showing insufficient.09:09
eglynnanand_ts: hmmm, that's odd given that the evaluator already determined that it should transition to OK09:09
eglynnanand_ts: $ ceilometer alarm-history -a f526313e-a76a-4064-bea9-e8e589e9ff0609:09
eglynnanand_ts: what version are you using?09:10
anand_tshistory here
eglynnanand_ts: 1.0.8 is the ceilometer client version, what service-side version?09:11
anand_tseglynn, I'm sorry. how to check that09:12
eglynnanand_ts: a-ha, that what I call a "flapping alarm" :)09:12
anand_tsohh!!! flapping alarm means?09:12
eglynnanand_ts: ... i.e. transitioning between ok/alarm and insufficient data09:12
eglynnanand_ts: ... because of a mismatch between the alarm period and the cadence of the incoming metric09:12
anand_tsokay, may I know why this happen here,09:13
eglynnanand_ts: so your orignal question was:09:13
eglynnI have set an alarm for a cirros instance cpu_util, whenever threshold goes above 20 for 3 consecutive cycles of *600s* it should trigger an alarm. but may I know where and how I get the alert.?09:13
anand_tsyeah, exactly09:13
eglynnbut the actual alarm period is 60s09:13
anand_tsbut beacuse I didn't get any alert I modified the threshold09:14
eglynnso if you check the /etc/ceilometer/pipeline.yaml, I bet you see the interval for the cpu_pipeline to be 600s09:14
eglynna-ha, I see09:14
anand_tsyeah. so by default it is set to 10 minutes right09:14
eglynnso now you've a mismatch between the time-window for the alarm and the periodicity of the incoming metric on which the alarm is based09:15
eglynn... so you'll need to change one or the other09:15
anand_tsOkay, also If I did that properly and if the thresholds exeeds how can I see the alert?09:15
eglynn... the alarm period must be a whole number multiple (e.g. 1, 2, 3) of the corresponding interval09:15
anand_tsSure. got it09:16
eglynn... well the metric is currently trending below threshold09:16
eglynn... that means the state will settle on OK once you've adjested the period/interval09:17
anand_tsCan I set threshold like 5. just to see how the alarm is generated?09:17
eglynn... whereas your ok action is set to nothing09:17
eglynnok_actions                | []09:17
eglynn... so you won't see any notification, because you've asked not to be told09:17
eglynnsay you adjust the threshold to be lower09:18
anand_tsyeah, but how to set ok_actions?09:18
eglynne.g. $ ceilometer alarm-update -a $ALARM_ID --threshold 509:18
eglynnjust use alarm-update as above09:18
eglynn(to set the ok_action)09:19
eglynnnote that the alarm_action you've set is [u'log://']09:19
eglynnthat means that the alarm-notifier will simply log the state transition09:19
eglynn... if you want a callout on a webhook instead09:19
eglynn... then set the action to
eglynn(subsitute sane values for your endpoint)09:20
eglynnanand_ts: ^^^ make sense?09:20
anand_tswill try that.09:20
eglynnanand_ts: cool :)09:20
anand_tsthank you so much eglynn, that definetely make sense09:21
eglynnanand_ts: np!09:21
*** nacim has quit IRC09:21
*** pnascimento has quit IRC09:24
*** pnascimento has joined #openstack-ceilometer09:27
openstackgerritZhiQiang Fan proposed a change to openstack/ceilometer: Reduce new token when call client object
openstackgerritMartin Geisler proposed a change to openstack/ceilometer: Remove unnecessary coding lines
*** _nadya_ has joined #openstack-ceilometer09:32
openstackgerritMartin Geisler proposed a change to openstack/ceilometer: Remove unnecessary coding lines
openstackgerritMartin Geisler proposed a change to openstack/ceilometer: Use Emacs-friendly file variable to set file encoding
*** Alexei_987 has joined #openstack-ceilometer09:39
*** shakamunyi has joined #openstack-ceilometer09:40
openstackgerritliusheng proposed a change to openstack/ceilometer: Improve the timestamp validation of ceilometer API
*** shakamunyi has quit IRC09:45
*** nacim has joined #openstack-ceilometer09:48
pnascimentogood morning stackers, where is documentation about "ceilometer-send-sample" command?09:50
*** `jpg has quit IRC09:51
*** boris-42 has quit IRC09:52
eglynnpnascimento: good morrow to you sir!10:01
eglynnpnascimento: ... I'm not sure there is much formal documentation10:01
eglynnpnascimento: ... (for that console script)10:02
eglynnpnascimento: ... however the main thing to know would be the command line options, right?10:02
eglynnpnascimento: ... which are clear from the code10:02
eglynnpnascimento: ...
pnascimentoeglynn: I was looking to output of "ceilometer-send-sample -h" and I was wondering if it existed more that10:04
pnascimento*then that10:04
eglynnpnascimento: ... I'm open to correction, but I'm not aware of any further formal documentation for that command10:05
* eglynn is think it should be straight-forward enough in any case, once the options are known ...10:05
pnascimentoeglynn: seems to me, that your thoughts are right ;) I was a question of curiosity... I was searching on Google and didn't find anything about the subject, just that...10:07
*** boris-42 has joined #openstack-ceilometer10:14
*** nacim has quit IRC10:25
*** r0j4z0 has joined #openstack-ceilometer10:26
*** _cjones_ has joined #openstack-ceilometer10:32
anand_tseglynn, webhook thing seems little confusing when I read, is there any other option like sms, email to nofity  when the threshold goes high? do you have any documents or links ?10:32
eglynnanand_ts: sadly, no, not at this time10:32
eglynnanand_ts: ... we're waiting on such user-consumably notifications to be provided via a Marconi SNS-style extension10:33
eglynnanand_ts: ... this was discussed at the last two summits, so hopefully the feature will become available during Juno10:33
anand_tseglynn, okay . good to hear this!!10:34
eglynnanand_ts: ... though there is the added complication that Marconi did not graduate as a integrated project on the last attempt10:34
eglynnanand_ts: ... that said, I'd be fairly confident that (a) Marconi will apply to the TC again during Juno and (b) it'll be better placed to graduate at that time10:35
anand_tsohh I see.10:35
eglynnanand_ts: ... their graduation request was withdrawn last time due to some push-back on the viability of their sql-a driver via mongodb10:36
anand_tsit should get the preference. without metering there is no cloud service10:36
*** _cjones_ has quit IRC10:37
eglynnanand_ts: ... well Marconi is the queuing service, whereas Ceilometer is already graduated as the metering service10:37
*** CAP1089 has joined #openstack-ceilometer10:38
anand_tsOhh!! okay, not aware of Marconi service10:38
*** r0j4z0 has quit IRC10:38
*** CAP1089 is now known as r0j4z010:38
anand_tseglynn, also through dashboard I'm not getting individual instance utilizations but project as whole ?10:43
eglynnanand_ts: IIRC there's a "group by" dropdown box on that dashboard10:44
eglynnanand_ts: ... default is probably "group by project ID"10:44
anand_tsI can see project and some ...10:45
eglynnanand_ts: ... check the other options ... e.g. "group by resource/user/none10:45
anand_tseglynn, I can see like this., . no other options10:47
eglynnanand_ts: so IIRC '--' is meant to mean 'None'10:48
*** matsuhashi has quit IRC10:48
anand_tsBut I have another projects also. let me check again10:48
eglynnanand_ts: note also that the metric you've chosen is cpu (cumulative CPU time) not cpu_util (as per your alarm above)10:48
eglynnanand_ts: the other project won't show up unless they also owned instances during that period10:50
eglynn*running instances10:50
anand_tsokay, this is the cpu_util graph .10:51
openstackgerritArtur Svechnikov proposed a change to openstack/python-ceilometerclient: Add methods to resource classes
eglynnanand_ts: sawtooth!10:54
eglynnanand_ts: ... cron job running?10:54
eglynnanand_ts: (on the instance)10:54
anand_tsnope, instance is juz cirros.10:54
eglynnanand_ts: ... have you tried setting the group-by to '--'10:55
anand_tsnot done anything on that instance10:55
anand_tsNow I get a better graph with instance uuid with -- . pic here
eglynnanand_ts: ok so in reality the CPU util is not actually spiking I suspect, as is clear from
eglynnanand_ts: what happening in is a sampling mismatch from the horizon side I suspect11:00
*** ildikov has joined #openstack-ceilometer11:00
anand_tseglynn. great, yes. it is stable , still value is between 7-9 as shown in graph fro horizon11:00
*** ildikov_ has quit IRC11:01
eglynnanand_ts: ... so ceilometer has a datapoint approx every 600s, whereas horizon uses some arbitrary period to populate that graph11:01
anand_tsyes it is, So which is the correct value we can take?11:02
eglynnanand_ts: ... nah, scratch that thought actually11:02
eglynnanand_ts: ... I just realized you've choosed the Avg aggregation11:02
openstackgerritZhiQiang Fan proposed a change to openstack/ceilometer: Enable filter alarms by their type
anand_tsavg in horizon11:02
eglynnanand_ts: ... I was thinking terms of sum, so that if two datapoints happen to fall into the same period, the sum would spike11:02
eglynnanand_ts: ... but clearly not the case here11:03
anand_tsokay, so in short , the results from cli and horizon are the same11:03
eglynnanand_ts: ... so I dunno why there is a semi-regular ~2% spike pattern in cpu_util, maybe there are some periodic tasks running on the VM11:04
anand_tseglynn, no idea on that. cirros vm I launched yesterday and not done anything inside vm11:06
*** `jpg has joined #openstack-ceilometer11:06
eglynnanand_ts: could be a quirk of some service/daemon running on the instance11:07
eglynnanand_ts: or could be a quirk of the cpu time accounting used by libvirt11:07
*** _nadya_ has quit IRC11:07
eglynnanand_ts: ... probably not big enough to worry unduely over11:07
anand_tseglynn, maybe, my entire setup is inside vmware ESXi11:07
anand_tseglynn. so this resource usage from horizon is diiferent and alarm generated from CLI is different is it?11:08
eglynnanand_ts: nope, the alarm evaluator, the CLI, and the horizon dashboard all use the same statistics API11:09
eglynnanand_ts: ... what's different would the period11:09
eglynnanand_ts: ... your alarm: period == 60s11:10
eglynnanand_ts: ... CLI: whatever you pass to the -p option11:10
anand_tseglynn: okay, now I got cleared in many aspects of ceilometer.11:10
eglynnanand_ts: ... horizon: IIRC period == duration / 40011:10
eglynnlsmola: ^^^ am I mis-remembering that period setting logic in horizon?11:11
* lsmola reads back11:12
*** _nadya_ has joined #openstack-ceilometer11:12
lsmolaeglynn: yeah, it always shows about 400 samples in the chart11:12
eglynnlsmola: cool, thanks for the confirmation11:13
lsmolaeglynn: np11:13
pnascimentoStackers, anyone knows why "ceilometer-send-sample" submit the sample on database but not show it on "ceilometer meter-list" ??11:14
eglynnpnascimento: did you mean sample-list?11:15
*** yfujioka has joined #openstack-ceilometer11:20
*** `jpg has quit IRC11:21
pnascimentoeglynn: neither "ceilometer sample-list -m <meter>" returns samples11:22
pnascimentoeglynn: both commands print an empty table11:23
eglynnpnascimento: but you've confirmed that the data has indeed landed in the DB?11:23
eglynnpnascimento: ... in that case, smells like a bug11:23
pnascimentoeglynn: but query the database through database client I can see that the samples were submitted ...11:23
eglynnpnascimento: ... can you file a bug on launchpad?11:23
pnascimentoeglynn: cloud it be because this version of Ceilometer is from Havana, maybe I should update it to IceHouse11:24
pnascimentoeglynn: what do your think?11:24
pnascimentoeglynn: s/cloud/could/11:25
eglynnpnascimento: ... yes Havana is quite old, in the sense that there will be no more routine releases from the stable/havana branch11:25
eglynnpnascimento: ... so you're better off upgrading to Icehouse11:25
eglynnpnascimento: ... then re-confirm the bug there and submit a report to LP if you could, thanks!11:25
pnascimentoeglynn: of course, if this "problem" persists I will submit a bug on LP ;)11:27
eglynnpnascimento: thank you sir!11:27
pnascimentoeglynn: no problem :) leave the "sir" :P no need for formalities :)11:32
*** promulo has quit IRC11:44
openstackgerritZhiQiang Fan proposed a change to openstack/ceilometer: Improve genconfig and ignore the generated file
*** _nadya_ has quit IRC11:47
*** pnascimento has quit IRC11:47
*** cmart has joined #openstack-ceilometer11:48
*** _nadya_ has joined #openstack-ceilometer11:52
*** _nadya_ has quit IRC11:56
*** _nadya_ has joined #openstack-ceilometer12:04
*** nacim has joined #openstack-ceilometer12:10
eglynnanyone wanna lay bets on the outcome of the K naming vote?12:14
* eglynn 's money is on Klang or Kepler12:14
jd__"30 Oct 2013" :D12:15
*** psharma has quit IRC12:15
jd__Kleber is also a candidate12:15
jd__though I'd bet on Kepler – but I wonder where Klang comes from?12:16
eglynnjd__: hmmm, Kremlin, /me likeee ... are you suggesting our beloved TC is becoming a tad Politburo-like? ;)12:16
eglynnjd__: /me is none too impressed with Klang in the flesh ...
*** julim has joined #openstack-ceilometer12:21
gibiI have to vote for Kossuth based on my nationality but it is fact that Kepler would be nice :)12:21
jd__eglynn: ah I don't think they're supposed to run naming from the whole country :D12:22
eglynnjd__: ... yeah, there may be some bending of the rules going on to find enough candidates for the vote12:23
jd__eglynn: baaaah so there's :D12:24
jd__there's a few ones in Brittany, I'd pick one ;)12:25
jd__but Emilien would prefer East and something like Klang :/ I guess it'll be a war12:25
eglynnwow! decidedly non-French-sounding some of those ...12:27
eglynnKrautwiller, Kurtzenhouse, Kleinfrankenheim, Koenigsmacker12:28
eglynn... distinctly Germanic-smelling ;)12:28
jd__yeah these are from the East departments as you noticed12:30
jd__the one starting with zipcode 22, 29, 56, are in Birttany :)12:31
jd__so it's likely more linked to you as I think they are somehow Gaelic or something12:32
jd__*just sayin'*12:32
eglynnjd__: a-ha, I get ya ... l'espirit pan-Celtique ;)12:33
*** jdob has joined #openstack-ceilometer12:37
*** asvechnikov has joined #openstack-ceilometer12:37
*** cdent has joined #openstack-ceilometer12:40
*** erecio has joined #openstack-ceilometer12:41
*** gordc has joined #openstack-ceilometer12:44
*** cdent has quit IRC12:44
*** thomasem has joined #openstack-ceilometer12:46
*** nosnos has quit IRC12:50
*** anand_ts has quit IRC12:51
*** prad has joined #openstack-ceilometer12:52
*** rbowen has joined #openstack-ceilometer12:54
*** fnaval has joined #openstack-ceilometer12:54
*** nacim has quit IRC12:56
*** _nadya_ has quit IRC13:02
*** admin0 has joined #openstack-ceilometer13:02
cmartI have a question regarding using postgresql with ceilometer..13:09
admin0go ahead cmart :)13:09
cmartI've changed CEILOMETER_BACKEND lib/ceilometer  to postgresql and then ran ./stack13:10
cmartbut after it finishes,  my ceilometer.conf ends with connection = mysql://root:intel123@
DinaBelovacmart, afair I mentioned yesterday that rewrites the conf files)13:11
DinaBelovasoo use unstack+rejoin)13:11
thervecmart, You shouldn't edit lib/ceilometer13:11
therveYou ought to be able to configure everything with local.conf13:11
cmartI also did it by adding  the following to my localrc: CEILOMETER_BACKEND=postgresql13:11
cmartand it didn't work13:12
DinaBelovacmart, oh, sorry, did not mention your lib/ceilometer stuff13:12
*** _nadya_ has joined #openstack-ceilometer13:13
*** _nadya_ has quit IRC13:13
admin0how do I link my codes to a blueprint ?   put blueprint (url of blueprint ) when submitting a review ?13:13
thervecmart, You probably want to explain why it didn't work.13:14
therveadmin0, In the commit message yes13:14
admin0i want to submit my codes for this:
cmartadmin0: check
cmartthey have some commit message examples that may help you :)13:16
openstackgerritLadislav Smola proposed a change to openstack/ceilometer: Automatic discovery of TripleO Overcloud hardware
admin0Implements: blueprint BLUEPRINT  .. in my case for ,, is it implements blueprint : URL  or  blueprint graphite-publisher  or   blueprint: ceilometer-graphite-publisher ?13:17
*** ddieterly has quit IRC13:18
openstackgerritA change was merged to openstack/ceilometer: refactor sql backend to improve write speed
cmarttherve: these were the steps I followed: 1) unstack. 2) change localrc and add CEILOMETER_BACKEND=postgresql. 3) run 4) after finishes, check the ceilometer.conf 4.1) debug the code and see what the engine_facade returns..13:18
eglynnadmin0: we're now following the blueprint specs process13:19
eglynnadmin0: ... so you'll need to propose in gerrit a detailed spec for the openstack/ceilometer-specs repo13:20
eglynnadmin0: ... base don this template
cmartthe thing is that I need to differentiate whenever the engine is mysql or postgresql (I have different implementations for both of them)13:20
*** jaypipes has quit IRC13:20
thervecmart, Was ceilometer.conf changed?13:21
cmartto  connection = mysql://root:intel123@
therveSo no? :)13:22
cmartI was working with mongo, that's why I'm saying that it changed :)13:22
cmartAm I doing smth wrong?13:23
*** yfujioka has quit IRC13:23
*** jaypipes has joined #openstack-ceilometer13:23
cmartwhat's the difference (at ceilo.conf level) when you have mysql or postgresql? Is it any?13:24
therveMaybe in devstack you need to do enable_service postgresql and disable mysql13:24
cmartI didn't know that that was needed.. :S13:26
therveI haven't tried, but it seems reasonable13:28
*** cdent has joined #openstack-ceilometer13:29
cmartguess it's worth to try it, right?13:29
cmartdo you know what are the exact service's names?13:30
cmarteglynn: Hi! Sorry to bother you.. do you know what are the steps needed to change the db backend to postgresql?13:30
*** Alexei_9871 has joined #openstack-ceilometer13:31
eglynncmart: ... setting it in the lib/ceilometer no workee?13:31
cmartnope :S.therve told me to disable mysql service and to enable postgresql.. but I'm not sure what are the service's names13:32
openstackgerritgordon chung proposed a change to openstack/ceilometer: fix indices in event tables
therveeglynn, You shouldn't need to change anything in lib13:33
thervecmart, mysql/postgresql13:33
openstackgerritgordon chung proposed a change to openstack/ceilometer: fix event model
cmartI'm also seeing at this line: that the database connection is configured base on a "database_connection_url' constant13:34
*** prad_ has joined #openstack-ceilometer13:34
*** fnaval has quit IRC13:36
cmartildikov: thanks! I will try that (that would change the entire db backend from the stack, right?)13:36
eglynncmart: what do you want to acheive, spin up a devstack with the DB backend preconfigured as postgresql?13:36
ildikovcmart: I think so, do you need another db for other services than Ceilometer?13:37
*** prad has quit IRC13:37
*** prad_ is now known as prad13:37
cmarteglynn: remember the filtering by meter stuff? That the guys from sqlalchemy told me to use json_extract_path_text from postgresql to do the "LIKE" within a json column type?13:38
eglynncmart: in that case, you need (a) to enable_service postgresql and (b) to ensure CEILOMETER_BACKEND=postgresql13:38
eglynn... (b) is probably settable via the local.conf13:38
cmartwell, I have one ugly implementation for mysql, but I want to use the json_extract_path_text for the postgresql13:38
eglynn... haven't tested it tho' so YMMV13:38
*** alexpilotti has quit IRC13:39
eglynncmart: "one ugly implementation for mysql" ==> are you special-casing for mysql versus postgres in the ceilometer sqlalchemy driver code?13:39
cmartyes.. and I know that you guys will told me not to do that :( but the thing is that my code is really, reaally ugly.. That's why I'm looking to do it better (at least for postgresql)13:41
cmartplease don't hit me13:41
*** Ruetobas has quit IRC13:41
*** pnascimento has joined #openstack-ceilometer13:43
cmartthis is the stuff that I'm doing:
eglynncmart: I'm not gonna lie to you, you would like see some "push-back" on including such special casing in impl_sqlalchemy13:44
cmartI'll be glad to have feeback from you guys13:45
gordcvrovachev: for the tempest patch, do you need multiple collector workers enabled or does it work with the one worker?13:45
admin0eglynn: the blueprint , it has to be submitted via git now ?13:47
admin0i am a bit unclear on the steps i need to take13:47
eglynnadmin0: spec submitted via gerrit13:47
*** Ruetobas has joined #openstack-ceilometer13:48
*** ddieterly has joined #openstack-ceilometer13:49
*** Ruetobas has quit IRC13:52
*** shakamunyi has joined #openstack-ceilometer13:54
*** nacim has joined #openstack-ceilometer13:55
*** Ruetobas has joined #openstack-ceilometer13:58
*** nacim has quit IRC14:02
*** fnaval has joined #openstack-ceilometer14:02
*** kaisero has joined #openstack-ceilometer14:05
*** fnaval_ has joined #openstack-ceilometer14:07
admin0silly question .. how do I createa  file here:  ?14:09
*** fnaval has quit IRC14:09
vrovachevgordc: I don't know exactly :) I'm testing it14:09
admin0i see gnocchi.rst , but its blueprint url is invalid .. or does the spec i am writing when approved automatically becomes the blueprint when approved14:09
eglynnadmin0: I moved gnocchi.rst to under the juno dir14:10
eglynnadmin0: fixing BP now14:10
admin0ok .. question still remains .. do i need to create a BP first, or this spec automatically becomes the bp ?14:11
gordcvrovachev: cool cool. let me know if it does and i'll enable my devstack multi-worker patch (
*** r0j4z0 has quit IRC14:11
eglynnadmin0: answer ... for fresh BPs, only create the spec ... then PTL will create a placeholder BP in launchpad if and when the spec lands in github14:12
vrovachevgordc: of course :)14:12
*** mengxd has joined #openstack-ceilometer14:12
eglynnadmin0: for pre-existing BPs, the LP BP remains but should be linked (via the 'Specification URL' field) to the spec in git once that lands14:13
*** fnaval_ has quit IRC14:14
*** fnaval has joined #openstack-ceilometer14:18
*** r0j4z0 has joined #openstack-ceilometer14:21
*** nacim has joined #openstack-ceilometer14:22
*** promulo has joined #openstack-ceilometer14:26
r0j4z0one question...14:30
ildikovdhellmann: so I usually run it outside of the virtual env and then it sometimes interferes with DevStack14:31
r0j4z0im trying to get working the ceilometer events14:31
r0j4z0on devstack14:31
ildikovdhellmann: so my experience is one or the other, but I did not try to run it inside the virt env?14:31
r0j4z0i configured 2 notification_driver on nova.conf14:31
r0j4z0notification_driver = ceilometer.compute.nova_notifier14:31
ildikovdhellmann: ... but I would expect that it should work there...14:31
r0j4z0y get the messages on the queue, the ceilometer consumes it14:32
dhellmannildikov: I always use the virtualenv for building the docs to ensure that I'm running the commands the same way they will run on the CI systems14:32
r0j4z0but i cant get any event when i run ceilometer event-list14:32
ildikovdhellmann: ... on my vm I usually reinstall the half of my packages and reclone the half world and then praying that it should work :)14:32
ildikovdhellmann: ... that sounds reasonable, I was just too happy that it works now as I did not want to move it14:33
*** raymondr has joined #openstack-ceilometer14:34
ildikovdhellmann: in my current virtualenv it works fine14:36
dhellmannildikov: the toctree error is being reported by sphinx in a different way than the errors that raise exceptions:
ildikovdhellmann: I tried tox -edocs yesterday it also worked fine and behaved that way as it is described in the bug14:36
dhellmannildikov: "worked fine" meaning it reported the expected failure, or "worked fine" meaning there were no warnings or errors and it did not fail?14:39
ildikovdhellmann: worked fine means that it did not have any import errors, so I could run the build14:40
dhellmannildikov: ok14:40
r0j4z0is like ceilometer are ignoring my event_definitions.yaml14:40
r0j4z0then it dont create the traits or events14:40
ildikovdhellmann: and I mentioned after it behaved the same misleading way, so it does not matter that where it is ran from, with which command, the basic issue seems to be the same14:41
ildikovdhellmann: toctree is for table of contents, right?14:41
dhellmannildikov: let's address one issue at a time. I'm trying to understand why the toctree and glossary warnings do not cause the job to fail14:41
dhellmannildikov: yes14:41
dhellmannildikov: it appears that sphinx is treating those warnings differently from some others14:42
ildikovdhellmann: ah, ok, I had the other issue in mind, but let's focus on this one then14:43
dhellmannildikov: which other issue?14:43
ildikovdhellmann: this error:
dhellmannildikov: I don't see that error in master. Is that from a patch?14:44
ildikovdhellmann: so in case of warnings in the logs I thought that it should not explicitely fail the build14:44
dhellmannildikov: it does not, but it should14:44
ildikovdhellmann: yes, it has been fixed, as I managed to recognize this formatting issue, while the patch was on review14:45
ildikovdhellmann: but earlier we had the Nova table missing from the Measurements section for a while, because of this behavior of the build14:46
*** alexpilotti has joined #openstack-ceilometer14:50
ildikovdhellmann: should this toctree error handling be changed in sphinx?14:59
dhellmannildikov: I'm still working on figuring that out14:59
admin0whom do i contact for issues in gerrit .. apparantly i cannot submit a bp( gerrit review) without current contact imformation .. and when i try to set it , i get server error - cannot store contact information15:00
*** mengxd has quit IRC15:02
amalagonDinaBelova: looks like gnocchi-api might be happier now15:04
DinaBelovaamalagon, okay, what did you do? :)15:04
dhellmannildikov: it looks like there's a small configuration change we can make to ceilometer to enable the warning behavior; maybe you've already tried that?15:05
amalagonDinaBelova: it was very silly I had installed gnocchi in my home dir not in /opt/stack/ ...15:05
DinaBelovaamalagon :)15:05
ildikovdhellmann: no, I think not15:05
amalagonDinaBelova: :))15:05
dhellmannildikov: add a "[pbr]" section with "warnerrors = True" to setup.cfg15:05
dhellmannildikov: that causes the build to fail on the toctree error, at least, and I think it should fail for the others, too15:06
dhellmannildikov: you'll have to fix all of the current doc build errors in the same patch where you enable that flag, of course15:06
ildikovdhellmann: ok, then what I wrote to fungi was right :)15:06
amalagonDinaBelova: is unhappy still ( but I *think* that the api is running15:07
DinaBelovaamalagon - try start it manually15:07
DinaBelovaas tests are still tests15:07
DinaBelovaif you'll try it manually it'll be clearer15:07
dhellmannildikov: I missed that :-)15:07
*** nacim has quit IRC15:07
DinaBelovaamalagon - as said it yesterday - gnocchi-api --debug --config-file /etc/gnocchi/gnocchi.conf15:08
ildikovdhellmann: I mean that the "warnerrors" setting seems to be missing15:08
dhellmannildikov: ah, yes, I thought it was on by default but it seems not15:08
ildikovdhellmann: as fungi were talking about it, I thought it also and I was not 100% aware of how to configure it, so I just believed what was easier :)15:09
amalagonDinaBelova: yep, I'm running that command, after putting out some debug messages it that the expected behavior?15:09
openstackgerritFabio Giannetti proposed a change to openstack/ceilometer-specs: Migrate Ceilometer Python Client to Support Keystone V3 APIs
ildikovdhellmann: but I will try this out, hopefully it will solve all the problems15:10
*** pnascimento has quit IRC15:10
*** alexpilotti has quit IRC15:11
dhellmannildikov: it will definitely help; let me know if you run into more issues and I'll try to help again15:11
DinaBelovaamalagon - smth like this looks ok15:11
ildikovdhellmann: thanks for the help15:11
DinaBelovaamalagon - api service is waiting for requests to it15:11
DinaBelovaamalagon - you may try to use curl and ping it15:11
eglynnamalagon: to confirm that gnocchi API is in fact operation, why not use wget or curl to invoke on an API?15:11
eglynnDinaBelova: snap! :)15:12
DinaBelovaeglynn ;)15:12
amalagonDinaBelova: exactly what I'm getting! (re paste) ok, I'll try that15:12
ildikovdhellmann: sure, I will update you if there is any new or unsolved issue with it, thanks again15:12
DinaBelovaamalagon, eglynn - I may confess - I use small python scripts usually here)15:12
DinaBelovawith requests lib)15:12
amalagonah, is it easier to use?15:13
eglynnamalagon: ... assuming keystone is not enabled yet for gnocchi15:15
eglynnamalagon: ... should be relatively straightforward to piece together either curl cmd line or requests usage as DinaBelova suggests15:15
eglynnamalagon: ... even with keystone, you can just reuse a token from another command15:16
*** shakamunyi has quit IRC15:16
eglynnamalagon: ... e.g. glance --debug image-list # <== then copy the X-Auth-Token header15:16
amalagonok I'll try that15:17
eglynnamalagon: ... ignore my keystone comments initially15:17
eglynnamalagon: ... only needed if your call fails with 401 Not Authorized15:18
*** shakamunyi has joined #openstack-ceilometer15:18
*** _cjones_ has joined #openstack-ceilometer15:19
vrovachevgordc: eglynn: dhellmann: After increasing timeout in notification tests: and merge Gordon's patch: notifications tests finished successfully. But Jenkins job for stable/icehouse with mysql backend are still dropping. We need to backport ceilometer changes to icehouse or ask qa team to disable vote this job.15:24
eglynnvrovachev: so the Tempest master runs against the stable/icehouse?15:25
amalagoneglynn: 401 not authorized, but maybe I'm not doing this right? I tried the authentication request with an empty tenant as done in
*** liusheng has quit IRC15:27
eglynnamalagon: are you calling ... $ curl -i '' -X POST ... ?15:27
amalagoneglynn: yeah15:28
eglynnamalagon: ... that's a POST to keystone to create a token15:28
*** liusheng has joined #openstack-ceilometer15:28
vrovacheveglynn: yes, one of the jobs runs against the stable/icehouse:
amalagonbut don't I need to create a token to use in the glance image-list command?15:29
eglynnamalagon: ... you'll want to POST to gnocchi /v1/resource or /v1/entity API15:29
vrovacheveglynn: this job has name: check-tempest-dsvm-full-icehouse15:29
eglynnamalagon: ... just ignore my comments on keystone for the moment15:29
eglynnamalagon: ... *if* you find your POST on the gnocchi API fails with 40115:29
eglynnamalagon: ... that's an indication that a keystone token is required15:30
amalagonah ok15:30
eglynnamalagon: ... in that case, you can easily grab one by copying the token used by another command15:30
* eglynn has confused the issue ...15:30
*** idegtiarov has quit IRC15:30
eglynnvrovachev: ... hmmm, so suppose the Tempest test exercised a new Juno feature?15:31
eglynnvrovachev: ... that didn't exist in Icehouse15:31
eglynnvrovachev: ... how could such a new test be landed on tempest/master?15:31
vrovacheveglynn, yes - this new feature by Gordon helped15:33
vrovachevbut as iceouse had no this one...15:33
vrovachevso we'll have icesouce tempest job failing in the Tempest15:34
vrovacheveglynn - so to have tempest ceilometer tests landed we have two variants15:34
eglynnvrovachev: I'm just confused about the general principal of tempest/master running against stable/icehouse15:34
gordceglynn: same thoughts.15:34
eglynnvrovachev: ... seems to block progress on tempest coverage for stuff that works in master15:35
vrovacheveglynn: they always had previous release checked too...15:35
vrovachevyes, exactly15:35
eglynnvrovachev: ... so is that related to the concept of "branchless tempest"?15:35
vrovachevI guess yes15:35
eglynnvrovachev: ... I would have it more sensible for tempest:stable/icehouse to be run against the services stable/icehouse15:36
vrovacheveglynn: the second variant is to backport features to icehouse, but v1 api drop hlped much here and it's impossible to backport it..15:36
eglynnvrovachev: ... and tempest:master to only run against services master15:36
vrovacheveglynn, we need to raise this question in openstack-qa channel I guess15:37
eglynnvrovachev: ... exactly, backporting is not the solution here in general15:37
vrovachevthey had some reason I guess - not sure which one15:37
*** ityaptin has quit IRC15:37
vrovacheveglynn - will you go with me to openstack-qa channel to communicate with qa folks?15:38
eglynnvrovachev: yep, gimme 5 mins15:38
vrovachevabout this problem we have here15:38
ildikovdhellmann: after adding that pbr section with the 'warnerrors' setting, it looks promising, the first warning made the doc build fail15:45
ildikovdhellmann: so I will fix the issues one by one and hopefully the build will fail in every case from now on15:45
ildikovdhellmann: it was a long time ago, when I was this happy because of a failed build, so thanks for the hint :)15:46
dhellmannildikov: :-)15:47
dhellmannildikov: thank *you* for fixing the problems the rest of us have left in the docs15:47
ildikovdhellmann: np, it has to be some genetic disorder that I deal with the docs that much or just simply a girly thing that I like the nice looking documentation :)15:51
*** Ruetobas has quit IRC16:01
*** Ruetobas has joined #openstack-ceilometer16:03
vrovacheveglynn - hehe, Sean does not like this idea :(16:06
eglynnvrovachev: yeap, the whole concept of branchless tempest seems flawed to me16:06
*** CAP1089 has joined #openstack-ceilometer16:07
eglynnvrovachev: ... it seems to ignore the fact that there is both (a) API fidelity and (b) quality-of-service in the implementation sitting behind the API16:07
*** r0j4z0 has quit IRC16:08
* eglynn scurries off to read up on "branchless tempest"16:08
*** Ruetobas has quit IRC16:08
vrovachevwell, I guess we need to push these folks to have ceilometer tests16:09
vrovachevas said - I see no way to run these tests on icehouse..16:10
vrovachevI'm not sure what is the best solution here16:10
eglynnvrovachev: so I'm just reading the branchless tempest BP spec
eglynnvrovachev: ... and a naughty thought occurred to me16:12
vrovachevwhich one?16:12
eglynnvrovachev: ... what if we treat these tests as falling under "Scenario 1: New Tests for new features"16:12
eglynnvrovachev: ... the new feature being: "Run fast enough for gate testing to be viable" ;)16:13
eglynnvrovachev: ... or just treat sql-alchemy in all versions prior Juno as being essentially an experimental/PoC thing16:14
*** Ruetobas has joined #openstack-ceilometer16:14
dhellmannildikov: I'm often the same way, so I don't think it's a girly thing :-)16:14
vrovachev:D not sure how this might work)16:14
vrovachevbut I love the idea16:14
eglynnvrovachev: ... yeah, prolly not what Sean wants to hear16:14
eglynnvrovachev: .... but I'll try that line of reasoning when we resume the conversation16:15
*** IvanBerezovskiy has left #openstack-ceilometer16:15
openstackgerritIlya Tyaptin proposed a change to openstack/ceilometer: Improve performance of api requests with hbase scan
vrovacheveglynn - if the idea is not to run these tests on the icehouse by some decorator, etc - I'm ok with it16:16
vrovachevand I can do ths, I guess16:16
vrovachevwe'll see what Sean wants to see here too16:16
eglynnvrovachev: yep, the BP isn't specific as to mechanics of the skipping for new feature tests on stable/icehouse16:17
eglynnvrovachev: ... "this test will need to be behind a feature flag (as the feature wasn't available in Icehouse)."16:17
eglynnvrovachev: ... if that "feature flag" maps on a test.attr decorator or some-such, then maybe this is workable16:18
eglynnmaps *onto16:18
vrovachev++ to this - let's ask Sean if it's possible16:18
eglynnvrovachev: ... yep, will do once he returns from lunch16:19
*** CAP1089 is now known as r0j4z016:20
openstackgerritFabio Giannetti proposed a change to openstack/ceilometer-specs: Migrate Ceilometer Python Client to Support Keystone V3 APIs
*** r0j4z0 has quit IRC16:21
DinaBelovaeglynn - personally I believe we need to skip them somehow for icehouse...16:26
DinaBelovaotherwise it's stupid really16:26
eglynnDinaBelova: I totally agree16:26
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Update the copyright date in doc
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ceilometer: Updated from global requirements
DinaBelovaeglynn, well, let's try to persuade Sean Dague)16:26
eglynnDinaBelova: ... yeah I need to "finesse" that message slightly with the qa folks ;)16:27
vrovachevyep, this situation looks strange - I do not believe it was no situation like this (without api impact) in the tempest16:27
DinaBelovavrovachev, definately - I guess this kind of changes should not be so unique..16:28
ildikovit seems odd to me too16:28
ildikov... or maybe I do not like backporting enough... :)16:29
vrovacheveglynn, I need to go now...16:29
vrovachevDinaBelova will be there yo help you)16:29
eglynnvrovachev: cool, thanks for your input on this16:29
vrovachevI also want my patches to be landed16:29
ildikovdhellmann: LOL :)16:30
*** jdob has quit IRC16:33
*** pnascimento has joined #openstack-ceilometer16:37
*** admin0 has quit IRC16:38
pnascimentoHello, I would like to run tests with tox, but how can I only specify only one package or suboackage?16:38
openstackgerritgordon chung proposed a change to openstack/ceilometer: clean up sample index
openstackgerritgordon chung proposed a change to openstack/ceilometer: clean up sample index
amalagonhey eglynn I think I may not understand curl so well - is something like curl -i -X POST {"archives":[1,720]} completely wrong syntax?16:48
eglynnamalagon: ... yep, use the -d option to curl for the POST data16:49
amalagonah, thanks16:49
DinaBelovaamalagon, use python ;) will be simpler ;)16:52
eglynnamalagon: ... yeah somthing along the lines of $ curl -i -X POST -H "Content-Type: application/json" -H "Accept: application/json" -d '{"archives":[1,720]}' ...16:53
eglynnamalagon: ... also DinaBelova speaks wisely on that point :)16:53
cdentthat's saying something not great about the (HTTP) API16:54
DinaBelovacdent - well, we need to pass this data somehow16:54
* cdent nods16:54
cdentit's a sort of sad-but-true16:54
DinaBelovaso that's not the worst thing I've seen16:54
*** _nadya_ has joined #openstack-ceilometer16:56
*** nati_ueno has joined #openstack-ceilometer16:57
*** eglynn has quit IRC17:01
*** ildikov has quit IRC17:09
*** pnascimento has quit IRC17:18
amalagonDinaBelova: the requests module is a beautiful beautiful thing17:21
*** jdob has joined #openstack-ceilometer17:28
*** raymondr has quit IRC17:30
*** raymondr_ has joined #openstack-ceilometer17:31
*** Alexei_9871 has quit IRC17:34
*** inc0 has quit IRC17:42
*** ildikov has joined #openstack-ceilometer17:44
*** _nadya_ has quit IRC17:50
openstackgerritgordon chung proposed a change to openstack/ceilometer: fix indices in event tables
*** _nadya_ has joined #openstack-ceilometer18:16
*** melodous has joined #openstack-ceilometer18:21
*** nati_ueno has quit IRC18:24
*** rbowen has quit IRC18:29
*** rbowen has joined #openstack-ceilometer18:29
*** nati_ueno has joined #openstack-ceilometer18:30
DinaBelovaamalagon ;)18:32
*** niteshselkari has joined #openstack-ceilometer18:40
niteshselkarihi all18:41
niteshselkariI want to know what does disk.write.bytes actually signifies?18:42
niteshselkariIs it  a volume space utilization  of an instance?18:43
*** _nadya_ has quit IRC18:49
*** _nadya_ has joined #openstack-ceilometer18:49
*** _nadya_ has quit IRC18:51
cdentniteshselkari: it's a cumulative count of the number of bytes written to the disk18:53
cdentso not how many bytes are written to the disk _right now_ but how many have been since the counter started counting18:53
niteshselkaricdent: can i find volume space utilized by an instance from this meter?18:54
cdentthere's more info here:
cdentI'm not certain but from that list it doesn't look like it18:57
*** shardy is now known as shardy_afk18:57
cdenti'm just in the process of spinning up some stuff to see if I can find out more18:57
*** eglynn has joined #openstack-ceilometer18:58
niteshselkariso what if i want to find the utilized space by an instance? which meter can give me this information? for example: I have an instance with 5 gb volume out of which 2 gb gets utilized. 3gb is remaining. so which meter can give me space utilized and available space in the volume18:59
cdentthis guy will probably know: eglynn : is it possible to "find volume space utilized by an instance"18:59
cdent(was that legible?)18:59
eglynnniteshselkari, cdent: I'd need to check the resource metadata associated with the volume meters19:03
eglynnniteshselkari, cdent: ... i.e. whether the UUID of the instance to which the volume attached is reported19:04
eglynnniteshselkari, cdent: ... but even assuming it was, it would still take a bit of query finessing to relate the two19:05
niteshselkarieglynn: can we take disk.write.byte meter into account while finding out space utilization?19:05
eglynnniteshselkari: that relates to the rate of I/O in terms of byte written per second19:06
eglynnniteshselkari: ... whereas you want the total current used in attached volumes, correct?19:07
niteshselkarieglynn : yes19:07
eglynn... of *bytes written per second19:07
eglynnniteshselkari: leave it with me and I'll dig around a bit tmrw (... just finishing up for the day)19:08
eglynnniteshselkari: ... actually if you could capture the question on to ensure that it doesn't fall thru' the cracks, that would be great!19:08
eglynnniteshselkari: ... just copy the question URL back here once you've saved it19:09
eglynnniteshselkari: thanks!19:09
cmartHi.. sorry to interrupt.. Could it be that this line ( has an extra "[" right at the beginning? I'm trying to solve an issue and I found that :)19:11
cdentit does look suspicious19:12
*** _nadya_ has joined #openstack-ceilometer19:15
cmartyeah.. I'm having issue with the cli and I found that.. Is there any tests around that code that I can use to check the effects of removing the "["?19:16
cmart*having an issue19:16
cdentsorry cmart: I'm too new to know.19:17
eglynnniteshselkari: thanks19:17
cmartcdent: me too :)19:17
melodoushi guys19:20
melodoussomeone here to give me a hand please?19:21
melodousim trying to understand how to work with events on ceilometer19:21
melodousI run ceilometer event-list i dont get anything19:21
eglynncmart: do you mean the extra opening square bracket in "[[a-zA-Z0-9_.]"?19:22
cmarteglynn: yes19:22
eglynncmart: ... that's valid surely, to match a literal "[" in addition to a alphanumeric, underscore or period?19:22
cdentthat ordering is bad for readability19:23
eglynncmart: e.g. ... re.match('[[a-zA-Z0-9_.]', 'this is a square bracket [')19:23
eglynncmart: ... true as true could be :)19:23
eglynncdent: fair point19:23
cmarteglynn: Oh I see..  I won't touch it then :)19:24
eglynn... of course the fact that it's valid doesn't mean it was also *intentionally* so ;)19:24
cdentalso . is probably wrong19:24
cdent[\[a-zA-Z0-9_\.] is probably more properly explicit19:24
eglynnmelodous: have you read
cdentbut then again without looking at tests, expected, inputs, etc19:25
cmartcdent: that's why I was thinking about tests..19:25
eglynncdent, cmart: chaps, I'm open to the possibility of there being wrongness there ...19:26
eglynncmart: ... so feel free to dig and submit a patch for whatever you view as suspect19:26
cdentI've added a note to my todo list to look at it more closely later, as I'm curious19:26
eglynncmart: ... dig *in19:26
eglynncdent: cool19:26
cmarteglynn: I think that the question will be: "what type of expressions do we need to handle? If an "[" is expected, then we should see what is the best option. Otherwise, we should remove it19:27
eglynncmart: that's fair19:27
cmartfor instance, with the new alarm filtering by state, I would have smth like this: ceilometer alarm-list --q "state=string::insufficient data"19:28
cmartbut the current regex used (right below it's not correct..19:29
DinaBelovaeglynn - so to summarise info about tempest testing19:29
cmartbecause it does not match the entire string after the "::"19:29
DinaBelovaeglynn - our small naughty plan is ok?19:29
eglynncmart: ... cool, there's no assumption what's there already is eteched in stone :)19:29
DinaBelovaeglynn - to skip these tests for the icehouse?19:30
openstackgerritgordon chung proposed a change to openstack/ceilometer: fix indices in event tables
eglynnDinaBelova: ... that was my read from sdague's response on #openstack-qa19:30
cmartcdent: do you have experience with regex? Could you give me a hand on this?19:30
cdentcmart: I do and can, but not right now: I'm in the midst of fighting with devstack but will be around all day (UTC day) tomorrow.19:31
eglynnDinaBelova: ... well I guess the more positive response was from mtreinish19:31
DinaBelovaeglynn, ok19:31
eglynnDinaBelova: ... <mtreinish> eglynn-afk: I think that's probably the right approach. BUt, you'll need to document that somewhere on the ceilometer side, about it being a poc prior to juno.19:31
DinaBelovaok, cool!19:31
eglynnDinaBelova: ... but Sean also seemed agreeable to the idea19:31
cmartcdent: OK!19:32
DinaBelovaso that means Vadim needs to find this magical skip and use it)19:32
*** nati_ueno has quit IRC19:32
eglynnDinaBelova: ... so I'm adding an item to the weekly agenda for tmrw to follow up on the questions like: what to call the flag for enabling the tests that won't run against "slow ceilo"19:33
eglynnDinaBelova: ... I'll circle back with Vadim at the meeting tmrw19:33
cmart"flagceilo" :):)19:33
cmartsounds great in spanish, let me tell you that19:34
* eglynn raps himself on the knuckles for using bizspeak like "circle back" ...19:34
eglynn... we should have the virtual equivalent of a "swear jar" ;)19:34
DinaBelovaeglynn, cools, thank you sir!19:34
*** niteshselkari has quit IRC19:41
openstackgerritA change was merged to openstack/ceilometer: Updated from global requirements
*** eglynn has quit IRC19:55
*** shardy_afk is now known as shardy19:58
openstackgerritgordon chung proposed a change to openstack/ceilometer: clean up sample index
*** nati_ueno has joined #openstack-ceilometer20:06
*** jaypipes has quit IRC20:08
*** nati_ueno has quit IRC20:11
*** cdent_ has joined #openstack-ceilometer20:12
*** erecio has quit IRC20:12
openstackgerritgordon chung proposed a change to openstack/ceilometer: clean up sample index
*** cdent has quit IRC20:14
*** cdent_ is now known as cdent20:14
*** _nadya_ has quit IRC20:14
*** dmsimard has joined #openstack-ceilometer20:19
dmsimardHi. Using both the ceilometer middleware and ceilometer-agent-central for Swift.20:19
dmsimardUsing mongodb as a backend database for ceilometer - it is my understanding that the Ceilometer middleware wants to insert data directly in MongoDB.20:20
dmsimardThis is kind of a problem for me since the MongoDB servers are in a private network but I do expose a ceilometer api endpoint.20:21
dmsimardIs it possible at all to have the ceilometer middleware talk to the API instead of attempting to write to the database directly &20:21
*** r-daneel has joined #openstack-ceilometer20:22
*** cdent has quit IRC20:26
*** cmart has quit IRC20:30
*** jdonalds has joined #openstack-ceilometer20:46
*** nati_ueno has joined #openstack-ceilometer20:50
*** jdob has quit IRC20:52
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
openstackgerritOpenStack Proposal Bot proposed a change to openstack/pycadf: Updated from global requirements
openstackgerritOpenStack Proposal Bot proposed a change to openstack/python-ceilometerclient: Updated from global requirements
*** promulo has quit IRC21:14
openstackgerritPradeep Kilambi proposed a change to openstack/ceilometer-specs: Spec for metering LBaaS
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
*** fnaval has quit IRC21:28
*** vigneshvar has joined #openstack-ceilometer21:31
*** vigneshvar has quit IRC21:32
*** jdonalds has quit IRC21:34
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Update the copyright date in doc
*** fnaval has joined #openstack-ceilometer21:38
*** nati_ueno has quit IRC21:49
*** thomasem has quit IRC21:49
*** ddieterly has quit IRC21:51
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
*** ddieterly has joined #openstack-ceilometer21:55
*** rbowen has quit IRC21:55
*** dhellmann is now known as dhellmann_21:58
*** dhellmann_ is now known as dhellmann21:58
*** dhellmann is now known as dhellmann_21:58
*** promulo has joined #openstack-ceilometer21:58
*** ekarlso has quit IRC22:01
*** nati_ueno has joined #openstack-ceilometer22:07
*** raymondr_ has quit IRC22:09
*** kaisero has quit IRC22:21
*** r-daneel has quit IRC22:26
*** ekarlso has joined #openstack-ceilometer22:27
*** jdonalds has joined #openstack-ceilometer22:28
*** gordc has left #openstack-ceilometer22:30
*** fnaval has quit IRC22:35
*** ddieterly has quit IRC23:11
*** prad has quit IRC23:14
*** raymondr has joined #openstack-ceilometer23:47
*** jaypipes has joined #openstack-ceilometer23:54

Generated by 2.14.0 by Marius Gedminas - find it at!