Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2016-07-21

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:33 Sesso joined #fuel
00:44 Liuqing joined #fuel
00:52 raunak joined #fuel
01:02 shank7485 joined #fuel
01:03 shank7485 left #fuel
01:54 aneliubin joined #fuel
02:18 raunak joined #fuel
02:35 xarses joined #fuel
02:44 raunak joined #fuel
03:56 xarses joined #fuel
04:25 raunak joined #fuel
04:36 Liuqing joined #fuel
04:37 Egyptian[Home] joined #fuel
05:09 xarses joined #fuel
05:50 code-R joined #fuel
06:08 javeriak joined #fuel
06:35 eric_lopez joined #fuel
06:49 bwe_ joined #fuel
06:51 bwe_ Hi guys, just a short question, I installed version 10 of Fuel (as 9 did not work in my environment) and everything worked fine except Horizon, it seems that the CSS is not loaded, did anyone else notice this?
07:04 javeriak joined #fuel
07:12 maestropandy joined #fuel
07:19 DavidRama joined #fuel
07:21 DavidRama joined #fuel
07:28 rtikunov joined #fuel
07:40 athinkingmeat hello, using mitaka with fuel and i need to replace ssl certificate for servises - should i only do it in haproxy or is there any other place?
07:41 Sesso its haproxy from what i remember
07:44 athinkingmeat thanks, Sesso
07:52 Sesso welcome
08:05 blahRus joined #fuel
08:11 DavidRama1 joined #fuel
08:17 Kdecherf is there any known issue about novnc on fuel 9 community ?
08:18 noshankus joined #fuel
08:18 yassine joined #fuel
09:20 krogon_ joined #fuel
09:21 vkulanov joined #fuel
09:38 javeriak joined #fuel
09:39 code-R_ joined #fuel
09:48 javeriak_ joined #fuel
09:51 ekosareva_ joined #fuel
09:55 azvyagintsev joined #fuel
09:56 dmburmistrov joined #fuel
10:09 Liuqing joined #fuel
10:13 asilenkov joined #fuel
10:14 meow-nofer_ joined #fuel
10:27 dmburmistrov joined #fuel
10:28 meow-nofer_ joined #fuel
10:30 asilenkov joined #fuel
10:31 tsufiev joined #fuel
10:37 maestropandy joined #fuel
10:37 maestropandy1 joined #fuel
10:37 Vijayendra joined #fuel
10:50 javeriak joined #fuel
10:54 code-R joined #fuel
10:56 chopmann joined #fuel
10:59 maestropandy joined #fuel
10:59 maestropandy left #fuel
11:30 aglarendil @iberezovskiy ^^
11:31 aglarendil @Kdecherf not those I am aware of
11:51 iberezovskiy Kdecherf, what's the issue you've faced?
11:55 javeriak joined #fuel
12:11 Sketch so i thought that when my ping to the haproxy vip started working after 10 minutes on my new deployment yesterday that it was just slow to come up, but today it's not responding again
12:12 Sketch i also notice that if i ping the non-vip br-ex ip on the controller, i get duplicate responses to every ping, which seems odd
12:14 Sketch i also get duplicates from the vip when it does work
12:23 Sketch is any of that expected behavior?
12:23 aglarendil Sketch: nope, it is not. are you sure that you do not really have a duplicated IP somewhere?
12:24 aglarendil in my experience I faced these issues with glitchy switches
12:24 Sketch i only have 3 nodes, and each one has a separate IP
12:24 eric_lopez joined #fuel
12:25 aglarendil what does your arping say?
12:26 fritchie joined #fuel
12:28 fritchie joined #fuel
12:28 Sketch well, i'm pinging from a different vlan
12:28 Sketch though i do see dups if i ping from the fuel server too
12:28 Sketch i wonder if the bonding isn't working right somehow
12:29 aglarendil which bond type did you enable?
12:29 Sketch balance-rr
12:29 aglarendil well, then you will surely get duplication
12:30 Sketch ok, so that part is normal then...
12:32 Sketch not sure why, though, but if it's just duplicate pings i can live with that.
12:33 Kdecherf iberezovskiy: on my last test, the novnc viewer was unable to connect to the server
12:33 Kdecherf (or the novnc proxy to the target node, i did not check)
12:34 GMAzrael bwe_: there seems to be a horizon problem with compressing the static material. MOS 9 and 10 seem to be affected.
12:34 aglarendil Sketch: it duplicates each and every packet
12:36 bwe_ GMAzrael: hmm, yes, sounds plausible, question is: how do I workaround that, apart from that everything seems to work fine
12:37 Sketch aglarendil: i expected it would do round-robin
12:37 bwe_ even though I like CLIs, using Openstack via CLI is not something I want to do on a regular basis ;-)
12:37 aglarendil well, yeah, I meant, that there is no difference between protocols you are using in a packet header
12:37 GMAzrael either find a way to get the horizon.js included in to /usr/share/openstack-dashboard/static/dashboard/js
12:37 GMAzrael or wait for a solution to be found in https://bugs.launchpad.net/mos/+bug/1603528
12:39 aglarendil GMAzrael: I guess, this one is related to UCA packages. May be it should be assigned to UCA packaging folks
12:39 iberezovskiy Kdecherf, do you use ssl? and how does url to vncconsole look like?
12:42 bwe_ aglarendil: I don't think so, Horizon is broken for me even if I do not select UCA
12:43 GMAzrael all of the javascript files for that directory are missing
12:46 Kdecherf iberezovskiy: i will check on the new lab
12:47 bwe_ hmm, I will compare it with a working Horizon Mitaka install
12:52 fritchie joined #fuel
12:58 Seafire joined #fuel
12:59 fritchie joined #fuel
13:10 romcheg IMPORTANT message for folks that are working on python-fuelclient: please rebase onto this patch https://review.openstack.org/#/c/344886/
13:10 romcheg it fixes a critical issue that marks a lot of failing tests as passing
13:10 romcheg vkulanov, ikalnitsky, please review ^ ASAP
13:12 romcheg dpyzhov: there was an issue in production code. Could you please consider assigning this bug to stable branches?
13:21 neouf joined #fuel
13:21 Seafire Hi guys, I have a few quick questions regarding backup and upgrade: in Fuelv8 backing up the Fuel Master was done by backing up the different docker containers, now in Fuelv9 docker containers are gone so how is the backup process handled? Does v9 have something in place to do backups or is it coming in the next versions? Another point is upgrading the Fuel Master..does a yum update / upgrade on the Fuel master and an apg-get update /
13:31 johnavp19891 joined #fuel
13:33 ddmitriev joined #fuel
13:33 ekosareva_ joined #fuel
13:33 ogelbukh Seafire: you should use fuel-octane for that
13:34 ogelbukh I mean, both for backups and upgrading
13:35 ogelbukh upgrade is a risky operation though, I don't recommend it for any kind of production environment
13:35 Seafire ogelbukh: ok thank you
13:35 ogelbukh without guidance from our engineers at least
13:40 Liuqing joined #fuel
13:40 neouf joined #fuel
13:48 Seafire ogelbukh: do you know if fuel-octane is compatible with Fuel 9? I'm looking at the git repo and the install instruction point to a file which isn't there anymore...
13:50 ogelbukh Seafire: let me check, I think 9.0 support is still in development
13:59 Kdecherf is there any reason to still deploy amqp queues with x-policy-ha: all on fuel even with the fact that this is not supported anymore by rabbitmq 3?
14:00 fritchie joined #fuel
14:04 fritchie joined #fuel
14:05 GMAzrael is 10.0 scheduled to run on 16.04?
14:06 Sketch aglarendil: well, active-backup works much better than balance-rr.  no more flaky network.
14:11 Sketch still have a couple of issues though
14:12 Sketch the links in fuel: "Horizon (HTTP)" ... horizon points to the fuel server itself.  http points to the horizon vip.
14:12 Sketch also, when i create a VM and try to view it's console via horizon, it points me at the fuel server instead of the controller vip.
14:13 Sketch this didn't happen with 8.0
14:14 code-R_ joined #fuel
14:15 Sketch https://bugs.launchpad.net/mos/+bug/1493287
14:16 Sketch i'm guessing this is why, because the dns hostname is set to the fuel server's name by deafult
14:19 neouf joined #fuel
14:20 romcheg ci-team: could you guys please re-trigger this one: https://ci.fuel-infra.org/job/master.python-fuelclient.pkgs.ubuntu.review_fuel_client/979/console
14:20 romcheg bookwar ^
14:43 xarses joined #fuel
14:50 johnavp19891 seafire: I don't know if fuel-octane supports 9.0 but I can tell you the docs are out of date hence why you can't find the octane file.  To build it do... yum install -y git patch python-pip python-paramiko && git clone https://git.openstack.org/openstack/fuel-octane -b stable/9.0 && cd fuel-octane && pip install -e ./
14:52 penguinolog joined #fuel
15:08 Kdecherf hm, when a controller goes down i see that all services become unavailable 'cause of memcache failure (and sometimes rabbitmq)
15:08 Kdecherf does anyone have any advice to acheive a good ha on this part?
15:08 cr0wrx joined #fuel
15:08 ekosareva_ joined #fuel
15:35 aglarendil Kdecherf: which version do you have? 9.0 ?
15:37 Kdecherf aglarendil: 9.0 community
15:38 xarses Kdecherf: neither should be a problem. Please create a bug, document your environment and steps to reproduce
15:41 Egyptian[Home] joined #fuel
15:43 fritchie joined #fuel
15:53 code-R joined #fuel
16:00 raunak joined #fuel
16:01 raunak joined #fuel
16:08 johnavp19891 Are there any functional differences between Mirantis fuel and fuel community or is it just cosmetic?
16:12 mwhahaha Sketch: how did you set the custom host name for your environment that was causing the deployment failure?
16:16 Zer0Byte__ joined #fuel
16:20 Sketch mwhahaha: the TLS endpoint name
16:20 mwhahaha sorry not that one, the one from yesterday
16:21 Sketch the default was public.fuel.tld or something like that, i changed it to the fuel dns name.
16:21 mwhahaha the duplicate host definition thing
16:21 Sketch oh
16:21 mwhahaha did you use fuel node --hostname ?
16:21 Sketch i just used the web ui
16:21 mwhahaha where in the webui? cause changing the display name doesn't change the hostname
16:21 Sketch but i did delete and re-add the same nodes a few times while trying to get things working
16:22 mwhahaha oh there it is
16:22 mwhahaha hiding
16:22 Sketch yeah, in the gear settings popup thing
16:23 Sketch but it only happened once.  not sure how, unless i had misnamed one one time.
16:23 fritchie joined #fuel
16:26 Sketch i don't think i did, but user error is always a possibility...
16:27 Sketch but now things seem to be almost up and running.
16:27 Sketch deployment works, but i get an error when trying to attach the public network to a VM
16:28 mwhahaha did the health check pass after deployment?
16:29 Sketch hmm, good Q.
16:30 Sketch i guess this may take a while.
16:30 mwhahaha it can
16:37 Kdecherf xarses: i will open several issues
16:37 Kdecherf what is the expected behavior regarding keystone when one memcache node dies?
16:38 mwhahaha well I'd assume the expected behavior be that it still keeps working but clearly that's not what's  happening
16:38 mwhahaha of course that's more a question for the keystone folks
16:39 Sketch mwhahaha: so far only the HA tests have failed, presumably because i only have 1 controller.
16:40 mwhahaha yea the ha tests won't pass with one controller
16:40 Sketch it's on the "3-60 min" tests now, we'll see how long those take.
16:54 Sketch hmm
16:54 Sketch Check creation of stack with Wait Condition/Handle resources: Time limit exceeded while waiting for stack status becoming "CREATE_COMPLETE" to finish. Please refer to OpenStack logs for more details.
17:18 fritchie joined #fuel
17:26 fritchie joined #fuel
17:27 aglarendil Kdecherf it should recalculate stuff for memcached sharding and continue working, but there will be something like 1 minute failover time where some of the requests to keystone will be failing
17:32 javeriak joined #fuel
17:37 Sketch hmm
17:38 Sketch last thing in the health test captured log: CommunicationError: Error finding address for https://ch0-stack.domain.com:9494/v0.1/artifacts/murano/v1/77d92a1d-ee55-40c5-9dd1-43ed1ac23dbb/archive: [SSL: UNKNOWN_PROTOCOL] unknown protocol (_ssl.c:765)
17:43 Sketch the log: https://rednsx.org/tmp/log/cluster_2_tests_platform.log
17:43 Sketch i ran it twice, once at 16:3x and once at 17:2x, in between I removed the image/instance/keypair i had set up just to make sure it didn't interfere
17:43 Sketch i'm using TLS self-signed by fuel
17:46 Sketch https://.../:9494 is up and running, i can use curl and get a 401 from it.
17:47 Sesso_ joined #fuel
17:49 Zer0Byte__ joined #fuel
17:49 Sketch oh wait, murano is the next test (which also failed).  the first test has it's own error in the log, it just seems to be looping waiting for completion which never finishes.  not very helpful.
17:50 Sketch seems like it could be a side effect of the network issue that i saw when i tried to bring up a vm?
17:51 Sketch https://bugs.launchpad.net/mos/+bug/1527224
17:51 fritchie joined #fuel
17:51 Sketch seems like the murano error is known and marked wontfix
17:51 elo joined #fuel
17:52 javeriak_ joined #fuel
17:54 elo joined #fuel
17:58 fritchie joined #fuel
18:04 raunak joined #fuel
18:05 fritchie joined #fuel
18:17 mwhahaha so yea i guess that's a known issue
18:23 Sketch https://bugs.launchpad.net/fuel/+bug/1544179
18:23 Sketch same error as the other failure, but this one says it's fixed
18:24 Sketch there's nothing in the log that seems useful, it just seems like it's not creating anything.
18:29 Sketch hrm, just looked in horizon and i have an extra network after running the tests: ostf-platform-service-net-221504906
18:31 mwhahaha yea you may need to manually clean that up
18:31 mwhahaha it's probably a left over from the failed test
18:31 Sketch yeah
18:32 Sketch hmm, can't delete it from horizon
18:41 vkulanov joined #fuel
18:48 Zer0Byte__ joined #fuel
18:55 xarses joined #fuel
18:56 Sketch ah, had to delete the router first, then the network.  but anyway, these tests appear to just be for heat/murano/etc, so they can probably wait. and i suspect they might be failing for the same reason that i can't create an instance with a public network, whatever that is.
18:57 Sketch if i create an instance without the public network, then try to add one later, i get an API error message in the log telling me to submit a bug report.
18:57 Sketch so, something definitely seems to be broken.
18:58 Sketch hmm, maybe.  i'm not 100% sure i'm even doing the right thing here.
18:59 Sketch but my only available networks are admin_floating_net and admin_internal_net
18:59 Sketch so i assume i want to add the floating net since it's on the public network.
19:00 Sketch i did check "assign public network to all nodes" but afaik it doesn't actually assign one.  plus i have to pick a network before i can create an instance.
19:03 Sketch so it's either floating or internal, or both
19:10 raunak joined #fuel
19:50 raunak joined #fuel
19:54 Zer0Byte__ joined #fuel
20:07 fritchie joined #fuel
20:17 DaveJ__ joined #fuel
20:57 Zer0Byte__ joined #fuel
21:05 cr0wrx joined #fuel
21:11 elo joined #fuel
21:40 HeOS joined #fuel
22:07 xarses joined #fuel
22:52 Zer0Byte__ joined #fuel
23:02 Kdecherf aglarendil: even after the return of the failed controller, requests of keystone to the preciously-failed memcached node still fail
23:02 Kdecherf (even after 40 minutes)
23:16 johnavp19891 joined #fuel
23:25 chopmann joined #fuel
23:38 Egyptian[Home] joined #fuel

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary