Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-06-13

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

All times shown according to UTC.

Time Nick Message
00:42 xarses joined #fuel
00:42 rmoe joined #fuel
00:47 aleksandr_null joined #fuel
03:48 dhblaz joined #fuel
03:49 dhblaz joined #fuel
04:26 hinv joined #fuel
04:27 hinv left #fuel
04:37 neophy joined #fuel
04:48 Kupo24z joined #fuel
05:24 al_ex3 joined #fuel
05:51 e0ne joined #fuel
06:03 e0ne joined #fuel
06:18 zhiyan hi, folks, can i get some fuel5 support from here?
06:34 al_ex3 joined #fuel
06:38 e0ne joined #fuel
06:40 e0ne joined #fuel
06:42 bogdando joined #fuel
07:26 odyssey4me joined #fuel
08:16 e0ne joined #fuel
08:16 alex_didenko left #fuel
08:16 e0ne_ joined #fuel
08:20 alex_didenko joined #fuel
08:28 e0ne joined #fuel
08:59 al_ex3 joined #fuel
09:39 artem_panchenko joined #fuel
09:49 b-zone joined #fuel
10:23 e0ne joined #fuel
10:28 e0ne joined #fuel
10:39 e0ne joined #fuel
11:05 DaveJ__ joined #fuel
12:16 bookwar joined #fuel
14:43 jobewan joined #fuel
15:23 albionandrew joined #fuel
15:31 e0ne joined #fuel
15:44 dhblaz joined #fuel
15:52 blahRus joined #fuel
15:55 TVR_ does fuel support multiple IP ranges with different Gateways for the Public IP Range?
15:56 albionandrew MiroslavAnashkin: xarses I have what I believe is a working 5.0 cluster. I can spark up VM’s, ping out to the web from instances etc etc. However the verify network test fails. All other tests pass. Is there something I can do to see whether the verify network test is a false positive?
15:57 gleam tvr, I don't believe so. I think that'll come in 5.1 with advanced networking
15:58 TVR_ OK.. thank you.
15:59 TVR_ Lastly, to get the dashboard to run HTTPS rather than HTTP, is the change as simple as editing /etc/puppet/modules/horizon/templates/openstack-dashboard.conf.erb ?
16:08 xarses TVR_: if you set up https for the dashboard then you should at a minimum convert keystone to https, and consider switching it to PKI tokens
16:08 xarses TVR_: no, you need to provide certs and put them in the right places
16:09 rmoe joined #fuel
16:12 xarses albionandrew: network verification used to not work in neutron after deployment, i wouldn't be surprised if it doesn't work
16:12 albionandrew xarses I wondered. Our 4. prod cluster fails but works fine. Thanks.
16:13 albionandrew xarses: I am running some other tests and trying to spark up a snap shot instance I get Error: No valid host was found.
16:13 albionandrew xarses: any ideas?
16:13 albionandrew same set up as in production.
16:14 xarses you would have to look at the logs to see why it raised the error
16:14 xarses back in a bit
16:14 albionandrew xarses: On the master?
16:16 TVR_ Thanks for the Info..will get on it to test.... thanks again
16:39 angdraug joined #fuel
16:42 e0ne joined #fuel
16:49 xarses joined #fuel
17:03 xarses joined #fuel
17:13 xarses joined #fuel
17:14 xarses joined #fuel
17:19 Kupo24z rmoe: get a chance to take a look at those logs?
17:24 rmoe yes, I have an idea too
17:24 rmoe can you run this command on your source and destination compute nodes
17:24 rmoe virsh sysinfo
17:29 Kupo24z 1s
17:31 mutex joined #fuel
17:34 Kupo24z rmoe: http://pastebin.mozilla.org/5403883
17:43 angdraug joined #fuel
18:18 albionandrew xarses: Re the verify networks tests. Would it help if I launched a bug or anything like that? I confirmed out networking is correct, the 5.0 deployment is a clone/testbed of our 4, we have in production.
18:22 xarses albionandrew: please do and attach logs
18:23 albionandrew Which logs do you want?
18:27 xarses support bundle would be best, otherwise we need the network settings you used (steps to reproduce) and the log from the verify networks tool
18:27 rmoe Kupo24z: I see the issue
18:27 rmoe the libvirt UUID is the same on both hosts
18:27 rmoe so libvirt won't migrate the vm
18:27 albionandrew I can give you the support bundle if you unlock my account? It was suspended when our support expired on 4.0
18:28 rmoe Kupo24z: you can override it by setting host_uuid in /etc/libvirt/libvirtd.conf
18:28 xarses albionandrew: on launchpad
18:28 albionandrew no on support.mirantos
18:28 xarses albionandrew: ya, we should be able to attach the file to launchpad. If you don't want it public i can change the bug to private
18:29 xarses or we can upload the bundle elsewhere
18:29 albionandrew Ok will mark as private
18:29 albionandrew I’ll put on launchpad
18:36 Kupo24z rmoe: I can just enter whatever numbers i want as long as its different?
18:36 rmoe I would guess so, but I'm not 100% sure
18:37 Kupo24z I can see this being a big problem as supermicro is kind of a big vendor
18:37 Kupo24z thanks for figureing it out
18:41 Kupo24z rmoe: do i need to restart anything more than nova-compute on each changed node?
18:43 rmoe I think that should be sufficient
18:44 Kupo24z Hmm no luck
18:44 Kupo24z same error, with a different UUID that in libvirtd.conf
18:46 rmoe it still thinks it's migrating to the same host?
18:47 Kupo24z yea, but i got it working
18:47 Kupo24z i needed to restart libvirt-bin
18:47 rmoe what was the solution?
18:48 rmoe ah, ok
18:48 Kupo24z i guess the solution is to force a 'uuidgen' on node deployment
18:49 Kupo24z ty for your help
18:49 rmoe no problem
18:50 Kupo24z the uuid determined by the motherboard im guessing?
18:54 rmoe I'm looking into it, this is a new problem for me
18:54 rmoe it's obviously dependent on the hardware
18:54 rmoe but I'm nopt sure how
18:58 rmoe it comes from SMBIOS
18:58 rmoe dmidecode will show you that info
18:59 rmoe UUID under 'System Information'
19:00 Kupo24z rmoe: let me know if you need some test hardware, all of our servers are SM and we have multiple generations of their stuff
19:08 e0ne joined #fuel
19:11 Kupo24z Looks like this MB is X10SLE
19:38 dhblaz joined #fuel
19:48 Kupo24z rmoe: do you know if https://bugs.launchpad.net/nova/+bug/1284709 will be in 5.1?
19:50 rmoe it'll be in both 5.0.1 and 5.1
19:50 Kupo24z cool
19:54 albionandrew xarses: Should I be able to do a migration from horizon?
19:54 xarses a live-migration, yes if you selected ceph for ephemeral
19:54 xarses a migration, is a cold migration
19:56 Kupo24z will 5.0.1 have an updated ISO?
20:06 bookwar joined #fuel
20:07 dpyzhov joined #fuel
20:07 albionandrew xarses: thanks. I figured that was the only way to do it.
20:25 albionandrew xarses If I check that box epheral disks will the disks now be as slow as a ceph backed volume?
20:26 xarses ephemeral will be backed by ceph, so yes it will perform the same as ceph
20:27 xarses otherwise it won't set up live-migrations correctly
20:27 albionandrew xarses: Thanks. Also is there a recomendation for mongo db size?
20:28 albionandrew At that moment its taking 40G that seems a lot… until I fill it up I suppose
20:30 xarses It will attempt to take an entire disk
20:31 xarses failing that it has some min-size, I'm not sure what. For a lab it's not really relevant. For production it is very relevant as mongo can become IO hungry
20:31 albionandrew xarses: OK Thanks
20:31 xarses As to some actual effective calculator, we don't have one
20:35 albionandrew xarses: MongoDB just for info - You must allocate at least 10,240 MB
20:39 albionandrew xarses: Can an instance that is volume backed by migrated if you don’t check the “ceph RBD for ephemeral volumes” option
20:40 xarses its possible, but we don't have it configured correctly unless ceph for ephemeral was selected, its a bug. I'm filing a report for it.
20:40 albionandrew xarses: thanks
20:41 xarses albionandrew: at a minimum https://github.com/stackforge/fuel-library/blob/master/deployment/puppet/ceph/manifests/ephemeral.pp has to be done
20:42 xarses actually some of that will actually enable ephemeral
20:42 xarses not just live migrations
20:42 xarses so we'd have to untangle which belongs to which
20:42 xarses you must have the firewall and live_migration_flag
20:43 xarses beyond that I'd have to research
20:45 albionandrew xarses: just out of interest if we push the reset, and make a small change does fuel start from scratch or just change whats changed?
20:47 albionandrew xarses: ON another matter “The stack tests fail” do they only pass if you have already set up stacks in horizon?
20:47 xarses albionandrew: in intergration?
20:47 xarses the bottom section in OSTF?
20:48 xarses I have no clue how to get that to work correctly. I think you need to have murano installed for it
20:48 xarses but again, i have no clue
20:48 xarses I need to file a bug for that aswell
20:49 albionandrew xarses Thanks
20:51 albionandrew xarses: With the reset button its a lot easier to play around and try stuff, so I will just play
20:52 xarses with the reset button, it starts from scratch
20:52 albionandrew xarses: Thanks
20:52 albionandrew Its cool thought great idea
20:52 xarses its the same as pressing delete, execpt it doesn't forget your settings and roles
20:53 christopheraedo Trying fuel-dev under Ubuntu 14.04, but getting yum-downloader problem: http://pastebin.com/LjVpE3it
20:55 xarses angdraug: ^
20:55 christopheraedo xarses: have you seen that before? Not seeing any issues in /opt/fuel-main/build/mirror/centos/yumdownloader.log either
20:56 angdraug xarses: that's not the false positive I keep commenting out
20:57 angdraug it shouldn't come up with any No Match lines
20:57 angdraug do you see which packages are causing this?
20:57 christopheraedo ah case sensitive grep damn you! "No Match for argument curl-7.19.7-37.el6_4"
20:58 christopheraedo (same thing after make deep_clean; make iso)
21:02 angdraug http://fuel-repository.mirantis.com/fwm/5.1/centos/os/x86_64/Packages/curl-7.19.7-37.el6_5.3.x86_64.rpm
21:03 christopheraedo cool thanks :) grabbed it and trying again...
21:05 christopheraedo (and removed offending line from /opt/fuel-main/build/mirror/centos/yumdownloader.log)
21:05 angdraug the problem is that it's a different build version
21:05 angdraug how come yumdownloder decided to look for el6_4 instead of el6_5.3?
21:06 christopheraedo No idea - could it be from doing this on Ubuntu 14.04? I wouldn't think that should matter.
21:06 angdraug shouldn't
21:07 christopheraedo ok that didn't work, it just tried to grab it again, failed, and dropped out.
21:09 angdraug can you grep for curl-7.19.7-37.el6_4 in your build dir? something is referencing it, and it's not requirements-rpm.txts, that one's got no version number
21:09 angdraug something in repodata/ ?
21:10 angdraug wait a minute
21:10 christopheraedo requirements-rpm.txt:curl-7.19.7-37.el6_4
21:10 angdraug https://review.openstack.org/97349
21:10 angdraug update your fuel-main
21:11 christopheraedo ah ok, I was trying with 5.0 tag
21:11 angdraug https://review.openstack.org/99434 then :)
21:12 christopheraedo excellent, thanks, trying again!
21:32 christopheraedo hit this one - at least it's known :)  https://bugs.launchpad.net/fuel/+bug/1312654
21:39 e0ne joined #fuel
21:58 e0ne joined #fuel
22:02 angdraug I thought we had a fix for that :(
22:05 christopheraedo well patching multistrap on 14.04 is a one-line fix, and I think it'll be in a package repo soon
22:08 christopheraedo but now I have a new error :( http://pastebin.com/LViiakm1
22:10 christopheraedo https://bugs.launchpad.net/fuel/+bug/1326476
22:20 bookwar christopheraedo: we run ISO build on Ubuntu 14.04 with miltistrap package taken from 12.04
22:21 bookwar for 14.04 version: the first issue with multistrap has been already resolved upstream, but the second you get is still there
22:23 bookwar first issue = that one-line fix
22:27 christopheraedo got it, thanks
22:35 christopheraedo is there any manual fix for the base-files bug (1326476)?
23:17 Kupo24z rmoe: is it working as intended that live migration doesnt copy the console log?
23:21 rmoe correct
23:21 rmoe it's a bug in libvirt: https://bugzilla.redhat.com/show_bug.cgi?id=994882
23:29 e0ne joined #fuel

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