Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-07-28

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

All times shown according to UTC.

Time Nick Message
02:30 mattgriffin joined #fuel
03:17 mattgriffin joined #fuel
03:51 ArminderS joined #fuel
03:51 xarses joined #fuel
05:43 e0ne joined #fuel
05:52 e0ne joined #fuel
05:58 e0ne joined #fuel
06:06 e0ne joined #fuel
06:11 xarses joined #fuel
06:12 e0ne joined #fuel
06:13 al_ex11 joined #fuel
06:22 e0ne joined #fuel
06:27 e0ne joined #fuel
06:33 e0ne joined #fuel
06:56 hyperbaba joined #fuel
06:58 pasquier-s joined #fuel
07:14 ddmitriev joined #fuel
07:28 e0ne joined #fuel
07:44 Longgeek joined #fuel
08:02 sbartel joined #fuel
08:14 artem_panchenko joined #fuel
08:22 e0ne joined #fuel
08:22 hyperbaba_ joined #fuel
09:12 book` joined #fuel
09:14 e0ne joined #fuel
09:18 andriikolesnikov joined #fuel
09:28 sikor_sxe joined #fuel
09:30 geekinut1h joined #fuel
09:45 sikor_sxe if i have a instance which i cannot delete
09:46 sikor_sxe how can i fix it?
09:46 sikor_sxe nova reset-state does not work
09:55 sc-rm joined #fuel
09:55 e0ne joined #fuel
10:00 sc-rm joined #fuel
10:01 sc-rm Hello, when I deployed openstack everything goes correct, but even thought I selected KVM it still sets the hypervisors to be QEMU, why is that?
10:07 evg_ sikor_sxe: what's about "nova reset-state --active ..."?
10:12 evg_ sc-rm: hello, please check "libvirt_type" parameter in your nova.conf.
10:43 adanin joined #fuel
11:13 sc-rm okay
11:13 sc-rm I’ll check it out
11:15 sikor_sxe evg_: doesn't work either :/
11:24 sikor_sxe i removed the instance from the db and the hd now
11:25 sikor_sxe messy
11:46 e0ne joined #fuel
11:47 sc-rm evg_: in the nova.conf it says libvirt_type=kvm
11:49 sc-rm evg_: this is what I’m talking about http://snag.gy/of7O8.jpg
12:16 sbartel sc-rm: it is known bug about nova : https://bugs.launchpad.net/nova/+bug/1195361
12:18 sc-rm Okay, so what openstack is reporting is wrong and it actually uses kvm
12:18 sbartel yes
12:18 sc-rm thanks :-)
12:58 evg_ sc-rm: there is mb some misunderstanding here, any way qemu is used with ot without kvm.
13:08 al_ex11 joined #fuel
13:15 sc-rm evg_: so basically no matter what I select in the installation phase of open stack it’s going to run on qemu?
13:17 sc-rm evg_: but if it does not have any performance impact I don’t care whethere it runs kvm or qemu?
13:24 sbartel really? I though they share the same connection type /interface
13:25 sbartel and that when we selected kvm, qemu interface was used but with kvm hypervisor
13:25 sbartel am I wrong?
13:27 ArminderS joined #fuel
14:16 evg_ sc-rm: qemu is ised for testing deployment inside virtual env (vitualbox for ex.). On bare metal you _should_  chose kvm as it'll increase perfermance significantly.
14:18 mattgriffin joined #fuel
14:21 evg_ sc-rm: the difference on low lever is just running "qemu bla-bla-bla", or qemu with kvm acceleration "qemu --enable-kvm bla-bla-bla"
14:21 jobewan joined #fuel
14:28 angdraug joined #fuel
14:37 Longgeek_ joined #fuel
14:48 Bomfunk joined #fuel
14:48 Longgeek joined #fuel
14:53 taj joined #fuel
15:09 geekinutah joined #fuel
15:14 xdeller joined #fuel
15:27 blahRus joined #fuel
15:32 TVR_ what is the default policy supposed to be for access & security?
15:33 TVR_ can I add a deny rule?
15:40 xdeller joined #fuel
15:42 geekinutah joined #fuel
15:44 xarses joined #fuel
15:56 ArminderS- joined #fuel
15:58 rmoe joined #fuel
16:44 TVR_ who can help me with access and security to the instances....
16:44 TVR_ it would seem my system is 'stuck' open
16:44 TVR_ for access...
16:45 TVR_ usually, the issue is it is locked down too tight
16:45 TVR_ is there a way to create a deny rule? what is the openstack default?
16:51 jaypipes joined #fuel
17:28 angdraug joined #fuel
17:34 andriikolesnikov joined #fuel
17:43 e0ne joined #fuel
18:07 e0ne joined #fuel
18:23 taj joined #fuel
18:56 e0ne joined #fuel
19:18 taj joined #fuel
19:28 xarses TVR_: the default rule should be deny
19:28 xarses the rules created in the default security group are needed so that any traffic can flow
19:31 TVR_ ok... so it would seem my default is corrupt.... as the instances are being opened wide up
19:32 xarses check to see if there are ingress rules with out protocol or port range
19:32 xarses which would make them any rules
19:32 loth2 TVR_: this is the default rules as they should be http://i.imgur.com/SJoy92J.png
19:33 TVR_ yes... I have that...
19:35 TVR_ is there a way to set the access and security to strict default? IE restart the services?
19:35 TVR_ it seems somehow we have caused our groups to not apply.....
19:36 xarses neutron?
19:37 TVR_ we are using heat templates... and something has borked the system to where when we create a new instance ... even if we apply a policy only allowing ssh (IE it wouldn't work and firwall everything off) it still is wide open
19:37 loth2 check your iptables -L -nv ?
19:37 TVR_ we have twin external networks here
19:37 xarses using neutron?
19:38 TVR_ yes
19:38 TVR_ neutron and gre
19:38 xarses http://openstack.redhat.com/Networking_in_too_much_detail
19:38 TVR_ I added bridges... and added the seccond network, bringing the entire CIDR into the SDN layer
19:39 TVR_ read that...
19:39 xarses are the iptables rules being created on the qrouter?
19:40 TVR_ so we have 3 networks... one internal non routable.... one anycast external where neutron holds the GW and the real router knows a route to it and a unicast external where we have floating IP's
19:40 TVR_ so you mean on the compute node holding the instance... yes?
19:56 e0ne_ joined #fuel
20:08 tatyana joined #fuel
20:13 maletilla_ joined #fuel
20:14 maletilla_ angdraug: what happened with RedHat?
20:15 angdraug maletilla: that's a broad question :)
20:16 maletilla a colleague just told me that they got pissed at u folks.
20:16 maletilla similar shit as with Piston
20:17 angdraug they generally don't like their competitors :)
20:18 maletilla :-)
20:18 angdraug coopetition in open source is a cutthroat business :p
20:19 angdraug if you just google for mirantis and red hat, there's a lot of fun stuff to read
20:19 KodiakF joined #fuel
20:20 KodiakF angdraug:  Thanks again for tipping me off to Fuel -  I'll see if I can sell the approach to mgmt
20:20 angdraug come back here if you have more questions, plenty of folk hang around and can help
20:21 angdraug if you end up patching our puppet manifests to meet your needs, there's #fuel-dev, we welcome external contributors
20:21 e0ne joined #fuel
20:22 maletilla shit, looks like I need a bigger cup of coffee
20:22 angdraug told you )
20:23 maletilla so, basically I'd install Fuel on one server (virtual or physical) and use it to deploy Openstack on bare metal hw?
20:23 * maletilla is a bit old.
20:25 adanin joined #fuel
20:25 angdraug yup, that's the gist
20:28 loth2 assuming your network is configured right :)
20:28 maletilla that can be arranged
20:29 angdraug yeah, make sure you read up on fuel's expectations about the network before laying it out
20:29 loth2 http://docs.mirantis.com/openstack/fuel/fuel-5.0/
20:29 loth2 Specificly http://docs.mirantis.com/openstack/fuel/fuel-5.0/pdf/Mirantis-OpenStack-5.0-PlanningGuide.pdf
20:30 angdraug or same in html: http://docs.mirantis.com/openstack/fuel/fuel-5.0/pre-install-guide.html
20:32 loth2 angdraug: Is it possible to add a purely swift node in Fuel? I see lots of refrences to it in the documentation
20:33 e0ne joined #fuel
20:34 KodiakF Shoot swift - I'd need a big-disk host for that too in addition to cinder - so Fuel will cost me 2 Dell 720hx to start instead of using my blades w/ enterprise SAN backing as I'm doing currently
20:34 angdraug loth2: no, https://blueprints.launchpad.net/fuel/+spec/swift-separate-role
20:34 angdraug KodiakF: just put everything in Ceph :)
20:34 KodiakF Then I'd need 3 Dell  720hx :)
20:35 angdraug nice thing about ceph is, it's the same storage pool for everything: cinder, nova, swift api
20:35 kupo24z Well, problem is with ceph/radosGW you dont have all the newest swift features
20:35 angdraug no, if you don't care about redundancy 1 720hx will suffice
20:35 angdraug just set your replication factor to 1 :)
20:35 KodiakF Wonder if Ceph is "ready" yet - so much talk out there about it being "almost" ready
20:35 angdraug it's at least as ready as swift :p
20:35 KodiakF ah yea I suppose
20:35 KodiakF haha true
20:36 angdraug I'd say starting with "Cuttlefish" and especially "Dumpling" it's pretty solid
20:36 KodiakF Going w/ Ceph will make all the devs here happy they have a boner for it
20:37 KodiakF I've been leery and coming from the Redhat ops role I've been leaning toward Gluster for my storage when we get hardware
20:37 angdraug ...and then Red Hat bought Inktank :p
20:37 xarses You would have one too if you played with it
20:37 xarses ;)
20:38 KodiakF xarses:  I'm sure - and Redhat is working on supporting it so I can't be too worried - and it incubated in my fave hosting company so that's a plus as well
20:40 angdraug on a more serious note, if you go with single node ceph, you'll need to hand-tweak your crushmap and ceph.conf so that there's redundancy across drives in that 720
20:40 angdraug if you just go with repl factor 1 and leave it at that, it'll be worse than raid
20:41 angdraug of course, all depends on how valuable is the data you're going to put there
20:47 kupo24z joined #fuel
20:56 KodiakF I'll have to read up more on Ceph - haven't looked at it in about 18 months.  I'm going to try to score 3x r720xd
21:10 angdraug KodiakF: http://www.mirantis.com/openstack-portal/external-tutorials/ceph-mirantis-openstack-full-transcript/
21:10 KodiakF woot
21:21 kupo24z Hmm, all the images on that page 404 for me
21:21 kupo24z weird
21:22 e0ne joined #fuel
21:24 angdraug hm, might not be just you
21:25 angdraug I'll check with our website people
21:27 kupo24z i have the image URL showing as https://lh3.googleusercontent.com/ysFzKQwehQzb2bhPa7UjNsMJl2KRJD5A_ddRRphWf_WKaKVgrS7zhWUi5vlzxbOcZDnoex8DZNuOTAELbyzc9zFY6PdTJRolCrckPpkLLZSbnR4oFGakuqwejg
21:27 kupo24z <p><img alt="" src="https://lh4.googleusercontent.com/SeukWlxtGM95Js5k8qWGYyh2zM4z6YUbAV3ybzcxP4gR_X5vr1eXJFoCRQ0R7-VKToHaTC2ZiVr3HW9ladfaqUS4EwN1GEabnrTMiR1zmRci8KLH4WnjUqzllw" width="299px;" height="151px;" /></p>
21:27 kupo24z <p dir="ltr">Figure 1 Ceph block storage</p>
21:31 angdraug kupo24z: looks like the google doc these links were referring to got moved, we'll recover it
21:31 angdraug for now all the diagrams can be found here: https://github.com/angdraug/beamer-fuel-ceph
21:33 angdraug and in the linked slideshare presentation, too
22:00 rmoe joined #fuel
22:01 kupo24z joined #fuel
23:42 vidalinux joined #fuel

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