| *** thorst has joined #openstack-powervm | 00:13 | |
| *** thorst has quit IRC | 00:18 | |
| *** thorst has joined #openstack-powervm | 00:18 | |
| *** thorst has quit IRC | 00:22 | |
| *** jwcroppe has joined #openstack-powervm | 00:39 | |
| *** apearson has joined #openstack-powervm | 00:43 | |
| *** jwcroppe has quit IRC | 00:45 | |
| *** tblakeslee has joined #openstack-powervm | 00:58 | |
| *** tblakeslee has quit IRC | 01:02 | |
| *** thorst has joined #openstack-powervm | 01:04 | |
| *** thorst has quit IRC | 01:08 | |
| *** thorst has joined #openstack-powervm | 01:24 | |
| *** thorst has quit IRC | 01:25 | |
| *** thorst has joined #openstack-powervm | 01:25 | |
| *** thorst has quit IRC | 01:30 | |
| *** thorst has joined #openstack-powervm | 02:40 | |
| *** thorst has quit IRC | 02:40 | |
| *** jwcroppe has joined #openstack-powervm | 02:42 | |
| *** jwcroppe has quit IRC | 02:48 | |
| *** kotra03 has joined #openstack-powervm | 02:58 | |
| *** Ashana has joined #openstack-powervm | 03:21 | |
| *** jwcroppe has joined #openstack-powervm | 03:22 | |
| *** Ashana_ has joined #openstack-powervm | 03:24 | |
| *** jwcroppe has quit IRC | 03:24 | |
| *** Ashana has quit IRC | 03:26 | |
| *** Ashana_ has quit IRC | 03:28 | |
| *** Ashana has joined #openstack-powervm | 03:30 | |
| *** Ashana has quit IRC | 03:34 | |
| *** Ashana has joined #openstack-powervm | 03:35 | |
| *** Ashana has quit IRC | 03:40 | |
| *** thorst has joined #openstack-powervm | 03:41 | |
| *** Ashana has joined #openstack-powervm | 03:41 | |
| *** Ashana has quit IRC | 03:46 | |
| *** thorst has quit IRC | 03:46 | |
| *** Ashana has joined #openstack-powervm | 03:47 | |
| *** Ashana has quit IRC | 03:51 | |
| *** Ashana has joined #openstack-powervm | 03:53 | |
| *** Ashana has quit IRC | 03:58 | |
| *** Ashana has joined #openstack-powervm | 03:59 | |
| *** Ashana has quit IRC | 04:04 | |
| *** Ashana has joined #openstack-powervm | 04:05 | |
| *** Ashana has quit IRC | 04:09 | |
| *** Ashana has joined #openstack-powervm | 04:11 | |
| *** Ashana has quit IRC | 04:15 | |
| *** Ashana has joined #openstack-powervm | 04:17 | |
| *** Ashana has quit IRC | 04:21 | |
| *** Ashana has joined #openstack-powervm | 04:22 | |
| *** Ashana has quit IRC | 04:27 | |
| *** kotra03 has quit IRC | 06:56 | |
| *** kotra03 has joined #openstack-powervm | 06:56 | |
| *** arnoldje has quit IRC | 07:10 | |
| openstackgerrit | Ravi Kumar Kota proposed openstack/nova-powervm: [WIP] Handle deletion of tap device from bridge https://review.openstack.org/341438 | 07:37 |
|---|---|---|
| *** k0da has joined #openstack-powervm | 07:58 | |
| *** kotra03 has quit IRC | 08:23 | |
| *** kotra03 has joined #openstack-powervm | 08:25 | |
| *** madhaviy has joined #openstack-powervm | 09:13 | |
| openstackgerrit | Esha Seth proposed openstack/networking-powervm: Change to return non-novalink vs all lpars https://review.openstack.org/341125 | 09:57 |
| *** madhaviy_ has joined #openstack-powervm | 10:59 | |
| *** madhaviy has quit IRC | 11:01 | |
| *** madhaviy_ is now known as madhaviy | 11:01 | |
| *** svenkat has joined #openstack-powervm | 11:29 | |
| *** thorst_ has joined #openstack-powervm | 11:40 | |
| *** jwcroppe has joined #openstack-powervm | 12:01 | |
| *** Ashana has joined #openstack-powervm | 12:04 | |
| *** mdrabe has joined #openstack-powervm | 12:31 | |
| *** madhaviy_ has joined #openstack-powervm | 12:35 | |
| *** madhaviy has quit IRC | 12:36 | |
| *** madhaviy_ is now known as madhaviy | 12:37 | |
| *** seroyer has joined #openstack-powervm | 12:45 | |
| *** kylek3h has joined #openstack-powervm | 12:51 | |
| *** kylek3h has quit IRC | 12:52 | |
| *** kylek3h has joined #openstack-powervm | 12:52 | |
| *** arnoldje has joined #openstack-powervm | 12:57 | |
| *** kylek3h has quit IRC | 12:59 | |
| *** arnoldje has quit IRC | 13:02 | |
| *** edmondsw has joined #openstack-powervm | 13:06 | |
| *** lmtaylor1 has joined #openstack-powervm | 13:15 | |
| *** lmtaylor1 has quit IRC | 13:21 | |
| *** lmtaylor has joined #openstack-powervm | 13:22 | |
| *** tblakeslee has joined #openstack-powervm | 13:24 | |
| *** apearson has quit IRC | 13:30 | |
| *** esberglu has joined #openstack-powervm | 13:32 | |
| *** arnoldje has joined #openstack-powervm | 13:50 | |
| *** seroyer has quit IRC | 13:55 | |
| *** edmondsw has quit IRC | 14:10 | |
| *** seroyer has joined #openstack-powervm | 14:11 | |
| *** efried has joined #openstack-powervm | 14:12 | |
| *** catintheroof has joined #openstack-powervm | 14:16 | |
| *** tjakobs has joined #openstack-powervm | 14:26 | |
| *** burgerk has joined #openstack-powervm | 14:31 | |
| *** tblakeslee has quit IRC | 14:35 | |
| *** apearson has joined #openstack-powervm | 14:42 | |
| *** tblakeslee has joined #openstack-powervm | 14:43 | |
| thorst_ | adreznec: do you know if OSA switched to MariaDB 10.1? | 14:46 |
| adreznec | thorst_: Interestingly enough no, and we actually don't need the work either anymore. The 10.0.26 release of MariaDB added support for ppc64el xenial | 14:53 |
| adreznec | e.g. http://mirror.rackspace.com/mariadb/repo/10.0/ubuntu/pool/main/m/mariadb-10.0/ | 14:53 |
| thorst_ | adreznec: whoa | 14:53 |
| thorst_ | Qing Wu shall be excited. | 14:53 |
| adreznec | I have a patch that adds 10.1 support locally that I'd written up... I might still push it if I have time to write the upgrade flow/test | 14:53 |
| adreznec | But we don't need it | 14:53 |
| *** apearson has quit IRC | 15:00 | |
| *** apearson has joined #openstack-powervm | 15:02 | |
| *** edmondsw has joined #openstack-powervm | 15:05 | |
| *** esberglu has quit IRC | 15:17 | |
| *** esberglu has joined #openstack-powervm | 15:18 | |
| *** kylek3h has joined #openstack-powervm | 15:23 | |
| *** esberglu has quit IRC | 15:26 | |
| *** esberglu has joined #openstack-powervm | 15:27 | |
| *** tblakeslee has quit IRC | 15:29 | |
| *** seroyer_ has joined #openstack-powervm | 15:31 | |
| *** seroyer has quit IRC | 15:31 | |
| *** seroyer_ is now known as seroyer | 15:31 | |
| *** tblakeslee has joined #openstack-powervm | 15:31 | |
| *** jwcroppe_ has joined #openstack-powervm | 15:32 | |
| *** arnoldje has quit IRC | 15:32 | |
| *** arnoldje has joined #openstack-powervm | 15:32 | |
| *** jwcroppe has quit IRC | 15:33 | |
| efried | thorst_, csky: https://review.openstack.org/#/c/321816/ | 15:36 |
| efried | I get that there's just one mapping object for all the volumes. | 15:36 |
| efried | When we add a new volume, do we augment that mapping? | 15:36 |
| efried | Are we talking about adding/removing volumes to/from the VM, or mapping/unmapping volumes to/from the VIOS? | 15:38 |
| thorst_ | efried: depends on the connection type | 15:40 |
| thorst_ | NPIV == no, it is what it is | 15:40 |
| thorst_ | vSCSI == if we've reached a fuse limit, we'll add a new mapping | 15:41 |
| efried | The above change set is NPIV, right? | 15:41 |
| thorst_ | right | 15:41 |
| thorst_ | the issue with NPIV is that there is one mapping for ALL volumes | 15:41 |
| thorst_ | so when you detach a volume - you have no idea if its the last volume | 15:41 |
| thorst_ | so you leave the mapping in the VIOS | 15:41 |
| thorst_ | then when you LPM it, the VIOS says 'there's nothing behind it'...so it moves the map, but doesn't vfcmap it to a physical port | 15:41 |
| efried | When you create the mapping, does this mapping object change? | 15:42 |
| thorst_ | csky's change set detects that on the next attach volume and connects the physical port to the mapping | 15:42 |
| thorst_ | only when we attach an object in this really weird scenario do we update the actual mapping object | 15:42 |
| thorst_ | so a new mapping isn't being created, rather we're just updating an existing one (to add a physical port) | 15:42 |
| efried | If there's a way to add a physical port.... | 15:43 |
| efried | ...isn't there a way to remove one? | 15:43 |
| efried | And shouldn't ya do that if you're removing a volume? | 15:43 |
| efried | And then you can tell how many physical ports are there | 15:44 |
| efried | including when there's zero | 15:44 |
| efried | so you can remove the mapping? | 15:44 |
| thorst_ | well, its one physical port to many volumes | 15:46 |
| thorst_ | the whole problem is you don't know how many volumes are in that mapping | 15:46 |
| thorst_ | because the VIOS doesn't see the volumes | 15:46 |
| thorst_ | we *could* find out if the volume they're asking us about is the last, first, etc.... but its a lot of extra computation | 15:47 |
| thorst_ | what csky has is pretty nice workaround that is performant...but not the most obvious way to solve the problem | 15:47 |
| csky | yeah, my patch does not attempt cleanup for the case where we've removed the last volume. But it does seem to work well for when we find ourselves in this stale situation on a new host. | 15:51 |
| thorst_ | I really like 'dealing with it' | 15:52 |
| thorst_ | cause 'stuff happens' way more than you can predict. | 15:52 |
| thorst_ | :-) | 15:52 |
| efried | I'm not convinced about "a lot of extra computation". We're talking about keeping a counter (per physical port?) on the volume driver, and incrementing/decrementing it whenever we add/remove a volume? | 15:52 |
| csky | We may need/want to fulfill requirements in the future that roll new NPIV volume attaches into a new mapping based on configuration rules (X number per mapping, etc.). At that point... | 15:52 |
| csky | we may want to add more sophistication here | 15:52 |
| *** k0da has quit IRC | 15:53 | |
| efried | I'm not necessarily advocating doing that - just making sure I understand what the alternative would entail. | 15:53 |
| thorst_ | csky: I specifically don't want to add that sophistication | 15:54 |
| thorst_ | lol | 15:54 |
| csky | Yeah, I would think a good counter | 15:54 |
| thorst_ | more stuff to keep track of....little benefit | 15:54 |
| thorst_ | unless we have more benefit than I know of | 15:54 |
| csky | System i has use cases for hundreds of volumes and limitation of 64 per VFC mapping...or something like that. | 15:55 |
| efried | We don't have a "fuse limit" deal for NPIV today? | 15:56 |
| *** apearson has quit IRC | 15:57 | |
| csky | we don't, do we? | 15:58 |
| csky | @thorst_: http://morpheus.rch.stglabs.ibm.com/#/c/3593/ | 15:58 |
| thorst_ | not for NPIV | 15:58 |
| thorst_ | fuse limit is for SCSI | 15:58 |
| thorst_ | you could certainly implement something like that, but the whole point of NPIV is that the system itself has no idea the number of volumes | 15:59 |
| thorst_ | just the LPAR's operating system and the backing storage device | 15:59 |
| thorst_ | the VIOS, the hypervisor, even the LPAR's hardware device have no idea. | 16:00 |
| thorst_ | (I guess the VIOS can figure it out kinda...but rare) | 16:00 |
| csky | It seems like something a management application would like to know...but yeah, maybe not novalink | 16:02 |
| *** apearson has joined #openstack-powervm | 16:05 | |
| *** seroyer has quit IRC | 16:17 | |
| *** seroyer_ has joined #openstack-powervm | 16:18 | |
| *** seroyer_ has quit IRC | 16:26 | |
| *** esberglu has quit IRC | 17:06 | |
| *** arnoldje has quit IRC | 17:11 | |
| *** madhaviy has quit IRC | 17:32 | |
| *** catintheroof has quit IRC | 17:58 | |
| *** k0da has joined #openstack-powervm | 18:04 | |
| *** apearson has quit IRC | 18:05 | |
| *** arnoldje has joined #openstack-powervm | 18:18 | |
| *** k0da has quit IRC | 18:24 | |
| *** tjakobs has quit IRC | 18:25 | |
| *** k0da has joined #openstack-powervm | 18:33 | |
| openstackgerrit | Sridhar Venkat proposed openstack/nova-powervm: Blueprint for nova-powervm SR-IOV VIFs https://review.openstack.org/322203 | 18:34 |
| *** apearson has joined #openstack-powervm | 18:35 | |
| openstackgerrit | Sridhar Venkat proposed openstack/nova-powervm: Blueprint for nova-powervm SR-IOV VIFs https://review.openstack.org/322203 | 18:39 |
| *** kotra03 has quit IRC | 18:52 | |
| *** apearson has quit IRC | 18:58 | |
| *** tblakeslee has quit IRC | 18:59 | |
| *** apearson has joined #openstack-powervm | 18:59 | |
| *** tblakeslee has joined #openstack-powervm | 19:00 | |
| *** k0da has quit IRC | 19:02 | |
| *** apearson has quit IRC | 19:17 | |
| *** catintheroof has joined #openstack-powervm | 19:18 | |
| thorst_ | csky: You're still waiting for nova-powervm merge? | 19:29 |
| csky | yes | 19:30 |
| thorst_ | I'll get on that now | 19:30 |
| thorst_ | this has been live tested? | 19:31 |
| thorst_ | efried: I +2'd but since I started the patch, you should be a second reviewer | 19:32 |
| *** apearson has joined #openstack-powervm | 19:34 | |
| *** lmtaylor has left #openstack-powervm | 19:41 | |
| csky | thorst_: yes, I made sure to live test again with latest nova-powervm and pypowervm changes yesterday. | 19:46 |
| *** k0da has joined #openstack-powervm | 19:53 | |
| *** apearson has quit IRC | 19:56 | |
| *** apearson has joined #openstack-powervm | 19:56 | |
| *** tjakobs has joined #openstack-powervm | 20:01 | |
| efried | thorst_, which one? | 20:06 |
| efried | https://review.openstack.org/#/c/321816 ? | 20:07 |
| efried | csky ^^ | 20:07 |
| csky | efried: yes | 20:10 |
| thorst_ | yep | 20:11 |
| efried | csky, thorst_: is an instance.save needed? | 20:13 |
| csky | The save does eventually happen for the attach | 20:14 |
| thorst_ | at the end of the flow | 20:14 |
| efried | k | 20:15 |
| efried | csky, thorst_: is it possible that the mappings are stale-but-still-on-this-system such that they're on the opposite VIOSes than when they started? | 20:16 |
| thorst_ | efried: don't think so | 20:17 |
| efried | In other words, if we start out with [port1: vios1] [port2: vios2] and we end up with [port2: vios1] [port1, vios2], then this line will still succeed: | 20:17 |
| efried | if all(pvm_vfcm.find_vios_for_wwpn(vios_wraps, pm[0])[0] | 20:17 |
| efried | for pm in npiv_port_maps): | 20:17 |
| csky | I think that is what we want | 20:19 |
| efried | Doesn't that mean the map is wrong? | 20:19 |
| efried | or it's just a map of ports to fabrics | 20:19 |
| efried | (This is what you people get for making me review code I don't understand.) | 20:20 |
| thorst_ | lol | 20:20 |
| thorst_ | yeah, that could happen. Would be totally fine | 20:20 |
| efried | okay. | 20:21 |
| efried | Done. | 20:22 |
| thorst_ | efried: get your hotel again for the mid cycle? | 20:24 |
| thorst_ | or are we bunking together | 20:24 |
| efried | Heh. I didn't actually cancel the old place before calling the new one. | 20:24 |
| thorst_ | good | 20:24 |
| thorst_ | I almost did, but then realized the new one may be booked too | 20:25 |
| efried | I could just invoke the BJJ Globetrotters clause and crash on the mats at one of the gyms in the area. | 20:25 |
| thorst_ | the watcher mid cycle is there the same week | 20:25 |
| efried | You would think we would have taken precedence. | 20:25 |
| thorst_ | meh | 20:25 |
| openstackgerrit | Merged openstack/nova-powervm: Handle stale system metadata (physical WWPNs) https://review.openstack.org/321816 | 20:28 |
| thorst_ | csky: You now need to cherry pick 321816 back to stable/Mitaka | 20:28 |
| csky | sure - thanks | 20:29 |
| *** k0da has quit IRC | 20:42 | |
| csky | thorst_: done | 20:45 |
| *** tjakobs has quit IRC | 20:54 | |
| *** k0da has joined #openstack-powervm | 20:55 | |
| *** erlarese has quit IRC | 20:58 | |
| *** tblakeslee has quit IRC | 20:58 | |
| *** tblakeslee has joined #openstack-powervm | 20:59 | |
| *** csky has quit IRC | 21:02 | |
| *** apearson has quit IRC | 21:06 | |
| *** apearson has joined #openstack-powervm | 21:12 | |
| *** thorst_ has quit IRC | 21:13 | |
| *** thorst_ has joined #openstack-powervm | 21:15 | |
| *** burgerk has quit IRC | 21:17 | |
| *** thorst_ has quit IRC | 21:19 | |
| *** seroyer has joined #openstack-powervm | 21:21 | |
| *** svenkat has quit IRC | 21:32 | |
| *** thorst_ has joined #openstack-powervm | 21:38 | |
| *** catintheroof has quit IRC | 21:40 | |
| *** thorst_ has quit IRC | 21:42 | |
| *** seroyer has quit IRC | 21:45 | |
| *** seroyer has joined #openstack-powervm | 21:47 | |
| *** seroyer has quit IRC | 21:47 | |
| *** edmondsw has quit IRC | 21:52 | |
| *** tblakeslee has quit IRC | 22:02 | |
| *** openstackgerrit has quit IRC | 22:03 | |
| *** openstackgerrit has joined #openstack-powervm | 22:03 | |
| *** k0da has quit IRC | 22:09 | |
| *** Ashana has quit IRC | 22:18 | |
| *** apearson has quit IRC | 22:19 | |
| *** efried has quit IRC | 22:22 | |
| *** efried has joined #openstack-powervm | 22:23 | |
| *** mdrabe has quit IRC | 22:38 | |
| *** esberglu has joined #openstack-powervm | 22:50 | |
| *** arnoldje has quit IRC | 23:02 | |
| *** apearson has joined #openstack-powervm | 23:08 | |
| *** thorst_ has joined #openstack-powervm | 23:14 | |
| *** thorst_ has quit IRC | 23:19 | |
| *** thorst_ has joined #openstack-powervm | 23:20 | |
| *** jwcroppe_ has quit IRC | 23:23 | |
| *** thorst_ has quit IRC | 23:24 | |
| *** jwcroppe has joined #openstack-powervm | 23:24 | |
| *** jwcroppe_ has joined #openstack-powervm | 23:25 | |
| *** jwcroppe has quit IRC | 23:29 | |
| *** jwcroppe_ has quit IRC | 23:30 | |
| *** thorst_ has joined #openstack-powervm | 23:42 | |
| *** seroyer has joined #openstack-powervm | 23:46 | |
| *** thorst_ has quit IRC | 23:51 | |
| *** seroyer has quit IRC | 23:51 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!