*** rlandy_ is now known as rlandy|out | 00:05 | |
*** ministry is now known as __ministry | 02:35 | |
*** krumelmonster_ is now known as krumelmonster | 05:20 | |
*** krumelmonster_ is now known as krumelmonster | 05:29 | |
xiaolong | hello | 07:22 |
---|---|---|
xiaolong | It's so quiet here | 07:26 |
admin1 | hello | 07:29 |
ironfoot | jamesbenson: interesting, maybe is time for us to re-share our config files? | 07:52 |
ironfoot | I bet they have changed a lot | 07:53 |
*** rlandy|out is now known as rlandy | 10:31 | |
jamesbenson | @ironfoot, definitely! Here are my ceph configs: https://paste.ubuntu.com/p/R9KNvPxrPN/ And my globals: https://paste.ubuntu.com/p/KvQ5TvzhGq/ | 12:52 |
*** rlandy is now known as rlandy|mtg | 13:03 | |
jamesbenson | @ironfoot, also odd is that zonegroup get and zonegroup list show default placement policy, but in horizon, no policy is shown. | 13:17 |
ironfoot | jamesbenson: can't look at this right now, but could you add your ceph.conf (without password) too? | 13:25 |
jamesbenson | @ironfoot: My ceph.conf is minimal, the ceph configs above is everything I modified outside of the default cephadm install | 13:34 |
ironfoot | Ah, I can see that now. Sorry, didn't even have time to scan properly the file | 13:35 |
jamesbenson | no worries | 13:35 |
jamesbenson | Everything I do is scripted, this is completely automated. So if you want to see anything else, just let me know. :-) | 13:37 |
jamesbenson | I was planning on showing you the final product when it worked | 13:37 |
ironfoot | heh nice :) | 13:38 |
jamesbenson | (And giving you credit due) | 13:39 |
ironfoot | we actually hit an issue in our setup, with nautilus, that seems not fixed in pacific: https://tracker.ceph.com/issues/48382 | 13:39 |
ironfoot | but didn't look like the one you were facing | 13:39 |
jamesbenson | yeah, doesn't quiet look the same, but I do notice that we have rgw_keystone_implicit_tenants = true, there it is false. I have tried it both ways though | 13:42 |
mirek186 | hi, quick question are you guys using DPDK in production, is it stable enough, worth spending time implementing it? | 14:26 |
jamesbenson | @mirek186, I'm not using it, so I can't comment :-) | 14:28 |
ironfoot | I started to see dhcp issues on my networks, and looks like disabling/enabling dhcp server in subnets seems to fix the issue, but they keep happening | 14:35 |
ironfoot | Today I noticed the following error on my dhcp agents, and I feel it could be related: | 14:35 |
ironfoot | neutron.agent.linux.ip_lib.AddressNotReady: Failure waiting for address fe80::a9fe:a9fe to become ready: Duplicate address detected | 14:35 |
ironfoot | Isn't that the metadata IP? what could be happening here? | 14:35 |
jamesbenson | we aren't using ipv6 or dhcp. I did find this: https://bugs.launchpad.net/neutron/+bug/1460177 | 14:43 |
frickler | ironfoot: iirc there was an issue with metadata and HA or DVR. which release are you using? | 14:49 |
ironfoot | frickler: Xena, deployed using Kolla | 15:32 |
ironfoot | I couldn't find anything on the web when looking for that exact error. So wasn't expecting this was a bug | 15:36 |
frickler | ironfoot: I also couldn't find the bug I seemed to remember, so maybe you want to create a new one. I'm pretty sure that this isn't the intended behavior | 15:50 |
ironfoot | i can't be the only one having this issue, right? | 15:51 |
ironfoot | I wonder if i can just disable ipv6 | 15:51 |
frickler | ironfoot: I also often wonder if I'm the only one running OpenStack, that seems to be a common perception - which somehow contradicts itself, I know | 15:54 |
ironfoot | with kolla nowadays anybody can run an openstack :) | 15:55 |
*** rlandy|mtg is now known as rlandy | 16:03 | |
jamesbenson | @ironfoot, I found this in my ceph logs: Not possible to open ports <[7480]>. firewalld.service is not available | 16:17 |
jamesbenson | But how would the curls work then? | 16:18 |
jamesbenson | so confused! | 16:18 |
jamesbenson | Also, everything I've done is port 7480 in my ceph deploy and kolla, but my curls are port 6780 | 16:22 |
ironfoot | i believe port 6780 is used by a haproxy (depoloyed by kolla) as a loadbalancer between your rgw services (on port 7480) | 16:24 |
jamesbenson | yes, I've seen that, okay. Thanks. Any ideas on the configs? | 16:25 |
ironfoot | i'm really really sorry, very busy day today, haven't had a chance | 16:26 |
ironfoot | this dhcp issue is killing me | 16:27 |
jamesbenson | fair enough :-). I'll post in ceph and kolla see if anyone else has idea's | 16:30 |
*** rlandy is now known as rlandy|bbl | 22:35 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!