| *** jmlowe has quit IRC | 00:01 | |
| *** jcollin has quit IRC | 00:11 | |
| *** jcollin has joined #openstack-manila | 00:11 | |
| vkmc | gouthamr, wanna hear your feedback later about access lists | 00:20 |
|---|---|---|
| vkmc | in the spec we put openstack share access create and openstack share access delete | 00:20 |
| vkmc | the verb used should be allow or deny | 00:21 |
| vkmc | I completely missed that | 00:21 |
| vkmc | should we update it to allow/deny or we stick with create/delete? | 00:21 |
| vkmc | IMHO allow/deny are more accurate but I wanna hear other people feedback | 00:21 |
| tbarron | allow/deny are historically what we have said but there are no deny rules, deny just deletes an allow rule :) | 01:05 |
| vkmc | well yes | 01:06 |
| vkmc | but semantically | 01:06 |
| tbarron | i'm not suggesting that we change from allow/deny | 01:06 |
| vkmc | if you have allow, you have deny | 01:06 |
| tbarron | but it's confusing to people (like me) who came from packet filter world where there were explicit deny rules too | 01:07 |
| tbarron | basially there were bit maps with two senses, allow and deny | 01:07 |
| tbarron | but we have only allow and implicit deny if there ain't a rule allowing it | 01:08 |
| tbarron | and the verb deny means delete a rule | 01:08 |
| * tbarron is going home now, just wanted to cause a little trouble on the way out | 01:08 | |
| vkmc | hahaha | 01:09 |
| vkmc | there no deny, there is "not allow" | 01:09 |
| *** brinzhang has joined #openstack-manila | 01:50 | |
| *** tinwood has quit IRC | 02:10 | |
| *** tinwood has joined #openstack-manila | 02:12 | |
| *** jcollin has quit IRC | 02:12 | |
| *** brinzhang has quit IRC | 02:59 | |
| *** jcollin has joined #openstack-manila | 03:32 | |
| *** jmlowe has joined #openstack-manila | 05:05 | |
| *** jmlowe has quit IRC | 05:08 | |
| *** jmlowe has joined #openstack-manila | 05:16 | |
| *** brinzhang has joined #openstack-manila | 05:41 | |
| *** jmlowe has quit IRC | 05:55 | |
| *** jmlowe has joined #openstack-manila | 05:57 | |
| *** jmlowe has quit IRC | 06:04 | |
| *** brinzhang has quit IRC | 06:47 | |
| *** maaritamm has joined #openstack-manila | 06:54 | |
| *** maaritamm has quit IRC | 06:55 | |
| *** maaritamm has joined #openstack-manila | 06:56 | |
| *** maaritamm has quit IRC | 07:03 | |
| *** maaritamm has joined #openstack-manila | 07:04 | |
| *** eharney has quit IRC | 07:08 | |
| *** eharney has joined #openstack-manila | 07:21 | |
| *** brinzhang has joined #openstack-manila | 07:27 | |
| *** brinzhang has quit IRC | 07:45 | |
| *** brinzhang has joined #openstack-manila | 07:45 | |
| openstackgerrit | Spyros Trigazis proposed openstack/python-manilaclient master: OSC Fix list for all projects https://review.opendev.org/704934 | 08:28 |
| *** tosky has joined #openstack-manila | 08:46 | |
| *** dviroel has joined #openstack-manila | 10:36 | |
| mnaser | how does the manila team feel about https://review.opendev.org/#/c/704807/ ? | 10:37 |
| *** pcaruana has quit IRC | 10:46 | |
| *** pcaruana has joined #openstack-manila | 11:48 | |
| *** jcollin has quit IRC | 11:48 | |
| *** brinzhang has quit IRC | 11:52 | |
| *** jvisser has joined #openstack-manila | 12:16 | |
| *** vhari has joined #openstack-manila | 12:16 | |
| *** pcaruana has quit IRC | 12:21 | |
| *** jvisser has quit IRC | 12:27 | |
| *** jvisser has joined #openstack-manila | 12:27 | |
| tbarron | mnaser: manila community meeting is a 1500 utc today in #openstack-meeting-alt so I added it to the agenda | 12:34 |
| mnaser | tbarron: cool thanks, i'll be in the channel, feel free to highlight me :) | 12:34 |
| tbarron | mnaser: seems OK to me offhand but worth getting general awareness, etc. | 12:34 |
| mnaser | yeah, i might actually have some manila-image-elements clean ups too, there's a bunch of stuff we can probably get away with -- i'll see how far i can get | 12:35 |
| tbarron | mnaser: excellent, we'll tag you when this comes up in the meeting | 12:43 |
| *** jcollin has joined #openstack-manila | 13:17 | |
| *** jvisser has quit IRC | 13:19 | |
| *** enriquetaso has joined #openstack-manila | 13:27 | |
| *** jvisser has joined #openstack-manila | 13:51 | |
| *** pcaruana has joined #openstack-manila | 14:03 | |
| *** eharney has quit IRC | 14:18 | |
| *** pcaruana has quit IRC | 14:34 | |
| *** jcollin has quit IRC | 14:40 | |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Use StrOpt for instance type https://review.opendev.org/705034 | 14:51 |
| *** andrebeltrami has joined #openstack-manila | 14:54 | |
| openstackgerrit | Maari Tamm proposed openstack/python-manilaclient master: Implement OSC share type commands https://review.opendev.org/701229 | 15:01 |
| *** pcaruana has joined #openstack-manila | 15:02 | |
| *** trident has quit IRC | 15:06 | |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Use StrOpt for instance type https://review.opendev.org/705034 | 15:09 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Introduce noop interface driver https://review.opendev.org/704807 | 15:09 |
| *** trident has joined #openstack-manila | 15:17 | |
| *** eharney has joined #openstack-manila | 15:18 | |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Use StrOpt for instance type https://review.opendev.org/705034 | 15:40 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Introduce noop interface driver https://review.opendev.org/704807 | 15:40 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: generic: Refactor network functions to l3_init https://review.opendev.org/705043 | 15:40 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: clear_outdated_routes: reduce neutron calls https://review.opendev.org/705044 | 15:40 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Use StrOpt for instance type https://review.opendev.org/705034 | 15:55 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: generic: Refactor network functions to l3_init https://review.opendev.org/705043 | 15:55 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: clear_outdated_routes: reduce neutron calls https://review.opendev.org/705044 | 15:55 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Introduce noop interface driver https://review.opendev.org/704807 | 15:55 |
| openstackgerrit | Mohammed Naser proposed openstack/manila master: Refactor route clearing to linux interface https://review.opendev.org/705050 | 15:55 |
| *** jmlowe has joined #openstack-manila | 16:00 | |
| *** jvisser_ has joined #openstack-manila | 16:07 | |
| *** jvisser_ has quit IRC | 16:08 | |
| *** eharney has quit IRC | 16:16 | |
| *** eharney has joined #openstack-manila | 16:16 | |
| *** vhari has quit IRC | 16:59 | |
| *** jmlowe has quit IRC | 17:12 | |
| *** jmlowe has joined #openstack-manila | 17:20 | |
| *** jmlowe has quit IRC | 17:24 | |
| *** jmlowe has joined #openstack-manila | 17:24 | |
| *** jmlowe has quit IRC | 17:48 | |
| *** jmlowe has joined #openstack-manila | 17:51 | |
| *** jmlowe has quit IRC | 18:10 | |
| *** andrebeltrami has quit IRC | 18:13 | |
| *** jmlowe has joined #openstack-manila | 18:21 | |
| *** jmlowe has quit IRC | 18:29 | |
| *** jmlowe has joined #openstack-manila | 18:38 | |
| noonedeadpunk | tbarron: hey, I have a question regarding cephfs_ganesha_server_ip. So it seems that it's used for reaching server via ssh. In case export config lies not in rados (but in local filesystem) and there are several ganesha servers, seems like only one will be properly configured and reached | 19:03 |
| tbarron | noonedeadpunk: maybe we need a list for that config option, we had really only envisioned one ganesha server (or, later running DHSS=true with a dynamically launched ganesha server per tenant, but that wouldn't use this option) | 19:05 |
| *** jmlowe has quit IRC | 19:05 | |
| tbarron | noonedeadpunk: do you have separate file systems behind the ganesha servers? | 19:06 |
| noonedeadpunk | I think they 're gonna be the same... I'm just thinking about some ha | 19:06 |
| noonedeadpunk | but just strated thinking about it actually | 19:07 |
| noonedeadpunk | I thought that both of them should use the same cephfs pool and in terms of failure of one ns server another one was able to pick up stuff | 19:08 |
| noonedeadpunk | hm, launching ganesha per tenant sound also interestingly actually... | 19:09 |
| noonedeadpunk | so in this case ganesha binary should be installed on the manila-share host? | 19:09 |
| tbarron | noonedeadpunk: jeff layton has work to run multiple ganesha daemons active-active with the same filesystem but I dunno that the default shipping ganesha stuff that ships today can do that | 19:10 |
| tbarron | noonedeadpunk: in any case i started this journey when it couldn't, so downstream we run it active-standby under pacemaker-corosync purposes for HA | 19:11 |
| tbarron | noonedeadpunk: only one ganesha is actually running but there are three controllers where it is "queued up" | 19:12 |
| noonedeadpunk | I'm thinking about active/backup at the point with keepalived | 19:12 |
| tbarron | noonedeadpunk: so yeah, we run it on the manila-share host and run manila-share the same way | 19:12 |
| noonedeadpunk | yes, that is exactly what I want to implement at this point | 19:12 |
| noonedeadpunk | oh, ok, so cephfs_ganesha_server_is_remote is false for this usecase? | 19:13 |
| tbarron | noonedeadpunk: yes | 19:13 |
| noonedeadpunk | and with remote it's not so trivial... | 19:14 |
| tbarron | noonedeadpunk: we also want to support the external ceph cluster case where ganesha is running externally as well, and | 19:15 |
| tbarron | the external cluster is responsible for ganesha HA. OpenStack/manila don't own the problem in that case. | 19:15 |
| tbarron | and for that currently the external cluster must present a VIP or some such for the config connection. | 19:16 |
| tbarron | present itself as one thing. | 19:16 |
| tbarron | noonedeadpunk: and then as I mentioned there's the future DHSS=True case, but work needs to be done such that | 19:17 |
| tbarron | the external Ceph cluster plus ganesha has an api so we can spawn up new ganeshas per tenant and then | 19:17 |
| noonedeadpunk | yeah, I see | 19:17 |
| tbarron | network them (we're thinking via kuryr) back into the tenant's own network | 19:18 |
| noonedeadpunk | sounds even more complicated | 19:18 |
| noonedeadpunk | but interestesting | 19:19 |
| tbarron | noonedeadpunk: the work to do it, yes. but if we can do it then life for the deployer will actually be pretty easy | 19:19 |
| openstackgerrit | Maari Tamm proposed openstack/python-manilaclient master: Implement OSC share access rules commands https://review.opendev.org/702924 | 19:22 |
| openstackgerrit | Maari Tamm proposed openstack/python-manilaclient master: Implement OSC share type commands https://review.opendev.org/701229 | 19:30 |
| *** jmlowe has joined #openstack-manila | 19:47 | |
| *** vhari has joined #openstack-manila | 20:15 | |
| dviroel | gouthamr: about your last comment in the spec [create from snapshot..], in order to get this info from db, we'll need to add a new field to shares, do you agree with that? | 20:46 |
| *** jvisser has quit IRC | 20:50 | |
| dviroel | s/shares/share_instances | 20:53 |
| *** eharney has quit IRC | 21:15 | |
| *** jvisser has joined #openstack-manila | 21:23 | |
| *** jvisser has quit IRC | 21:28 | |
| *** jvisser has joined #openstack-manila | 21:28 | |
| *** enriquetaso has quit IRC | 21:33 | |
| *** jvisser has quit IRC | 22:21 | |
| *** jmlowe has quit IRC | 22:43 | |
| *** vhari has quit IRC | 23:24 | |
| *** dviroel has quit IRC | 23:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!