| opendevreview | Amir Hossein Ahmadi proposed openstack/kolla-ansible master: swift: Reintroduce the Swift deployment role. https://review.opendev.org/c/openstack/kolla-ansible/+/953576 | 06:29 |
|---|---|---|
| opendevreview | Piotr Milewski proposed openstack/kolla-ansible master: haproxy: add support for disabling basic auth per service https://review.opendev.org/c/openstack/kolla-ansible/+/958861 | 08:18 |
| opendevreview | Merged openstack/kolla-ansible master: docs: fix mariadb client binary https://review.opendev.org/c/openstack/kolla-ansible/+/973556 | 09:15 |
| opendevreview | Merged openstack/kayobe master: CI: Add bandit testing https://review.opendev.org/c/openstack/kayobe/+/933438 | 09:22 |
| opendevreview | Pierre Riteau proposed openstack/kolla-ansible stable/2025.2: docs: fix mariadb client binary https://review.opendev.org/c/openstack/kolla-ansible/+/975769 | 09:43 |
| Moha_ | If we don't set the `_allocation_ratio` for RAM/CPU, the default values would be 3x and 16x respectively for RAM and CPU? | 10:09 |
| *** jhorstmann is now known as Guest1566 | 10:26 | |
| opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: CI: add service-check-containers to ovn scenario trigger https://review.opendev.org/c/openstack/kolla-ansible/+/975770 | 10:33 |
| opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 10:34 |
| opendevreview | Merged openstack/kolla-ansible stable/2025.2: docs: fix mariadb client binary https://review.opendev.org/c/openstack/kolla-ansible/+/975769 | 10:37 |
| bbezak | Moha_: https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.initial_cpu_allocation_ratio | 10:43 |
| bbezak | Depends on openstack release | 10:43 |
| opendevreview | Matt Crees proposed openstack/kolla-ansible master: Fix mount leak with NFS https://review.opendev.org/c/openstack/kolla-ansible/+/975771 | 10:43 |
| bbezak | It changed in Antelope to cpu ratio to 4 I believe | 10:45 |
| bbezak | In zed it was 16 - https://docs.openstack.org/nova/zed/configuration/config.html#DEFAULT.initial_cpu_allocation_ratio | 10:45 |
| bbezak | And ram allocation ratio changed from 1.5 to 1 | 10:45 |
| opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: CI: add service-check-containers to ovn scenario trigger https://review.opendev.org/c/openstack/kolla-ansible/+/975770 | 10:56 |
| opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 10:56 |
| opendevreview | Amir Hossein Ahmadi proposed openstack/kolla-ansible master: swift: Reintroduce the Swift deployment role. https://review.opendev.org/c/openstack/kolla-ansible/+/953576 | 11:02 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 11:04 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: CI: add service-check-containers to ovn scenario trigger https://review.opendev.org/c/openstack/kolla-ansible/+/975770 | 11:16 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 11:16 |
| Moha_ | bbezak: Ah, thanks | 11:19 |
| Moha_ | Does this downward trend of the “ratio” in different versions mean that, for example, you no longer recommend coefficients like 16x for CPU? | 11:19 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 11:19 |
| frickler | Moha_: I don't think there ever was a recommendation from the kolla team? anyway, the best we could say probably is "it depends on your use case" | 11:36 |
| Moha_ | I meant the 'Deafult'. `<deploy-venv>/share/kolla-ansible/doc/source/admin/advanced-configuration.rst` | 11:54 |
| Moha_ | frickler: I see, +1 | 11:55 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: CI: add service-check-containers to ovn scenario trigger https://review.opendev.org/c/openstack/kolla-ansible/+/975770 | 12:11 |
| opendevreview | Doug Szumski proposed openstack/kolla-ansible master: Partially revert "Adapt notifying handlers message" https://review.opendev.org/c/openstack/kolla-ansible/+/975192 | 12:12 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 12:13 |
| opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: CI: add service-check-containers to ovn scenario trigger https://review.opendev.org/c/openstack/kolla-ansible/+/975770 | 12:17 |
| opendevreview | Doug Szumski proposed openstack/kolla-ansible master: Partially revert "Adapt notifying handlers message" https://review.opendev.org/c/openstack/kolla-ansible/+/975192 | 12:17 |
| opendevreview | Bertrand Lanson proposed openstack/kolla-ansible master: DNM: Making iterated container-check idempotent https://review.opendev.org/c/openstack/kolla-ansible/+/973803 | 12:17 |
| frickler | Moha_: that is just an example for how to override these values in a kolla deployment, that doesn't imply the shown values are recommended in any way. if you have an idea how to make that more clear in the wording of that doc, that would be great | 13:09 |
| Moha_ | frickler: Seems meaningful | 13:31 |
| opendevreview | Piotr Milewski proposed openstack/kolla-ansible master: Add Redis/Valkey support for [cache] and fix Neutron template https://review.opendev.org/c/openstack/kolla-ansible/+/971970 | 14:35 |
| tafkamax | Hmm | 14:55 |
| tafkamax | I am trying to deploy a new cluster using stable/2025.2 | 14:55 |
| tafkamax | and i am stuck at... (full message at <https://matrix.org/oftc/media/v1/media/download/AXFT58jEYF2skiVxYipNALd2NBzaHoMQQQ_PHnLSywCBxy4y0V1MU9Fnsz5-cHSri4DhNhNqgoKD-UI1InDv0elCecdmaD6AAG1hdHJpeC5vcmcvcWFiY3lKZGl5RUN4WmpzUVlwTUlldG5u>) | 14:55 |
| tafkamax | i deployed another similar setup in december or so | 14:56 |
| tafkamax | only real difference is that I pulled in all patches since then | 14:56 |
| tafkamax | so i am on 21.0.0-27 | 14:56 |
| tafkamax | "msg": "Task failed: Module failed: Could not find versioned identity endpoints when attempting to authenticate. Please check that your auth_url is correct. Service Unavailable (HTTP 503)"} | 14:57 |
| tafkamax | have there been any patches related to this since? | 14:57 |
| tafkamax | should I downgrade to just 21.0.0 tag and retry | 14:57 |
| blanson[m] | Moha_: if it can be of any help, we do ram_allocation_ratio: 1 and cpu_allocation_ratio: 2, anything above get complicated to guarantee SLA with densely populated clusters | 14:58 |
| blanson[m] | but as frickler said, it most likely depends on your use-case. This is ours for public hosting | 14:59 |
| tafkamax | hmm might be related to certificate... | 15:01 |
| Vii | tafkamax: add to the end -vvvv | 15:41 |
| Vii | verbose mode | 15:41 |
| tafkamax | ok, I currently changed my package to 21.0.0 tag release | 15:52 |
| tafkamax | I destroyed the previous cluster | 15:52 |
| tafkamax | lets see if it works | 15:52 |
| tafkamax | and if it works I will try to bump and recreate again | 15:52 |
| tafkamax | only change i can find is this: https://opendev.org/openstack/kolla-ansible/commit/d21951467185a7e97d8d2a8b9ae878c41ef9b626 | 15:53 |
| Moha_ | blanson[m]: Yes; ratios like 1× and 2× are generally safe, but at that point it hardly qualifies as cloud by definition. | 15:54 |
| Moha_ | I mean, one of the core pillars of cloud is a shared resource pool, where unused capacity paid for by one tenant can be dynamically consumed by another. | 15:55 |
| blanson[m] | yes but clients won't be too happy if you tell them the reason their VM is laggy is because they're currently sharing a CPU with 16 other VMs, of which 2 are mining bitcoin :) | 15:57 |
| blanson[m] | and ram get's even uglier when hypervisors are too dense and start swapping guests pages | 15:58 |
| blanson[m] | gets* | 15:58 |
| blanson[m] | it's a problem of how dense vs how much performance/stability you're selling | 15:59 |
| Moha_ | That’s correct; I think it would be better if this were stated transparently in the customer’s user panel, so they clearly understand the trade‑off they are accepting - price versus stability. | 16:01 |
| blanson[m] | we do "low-cost" hosting aswell, where density is around x1.7 RAM and 6-8x CPU, but you can't sell those to companies looking to use your public cloud unfortunately | 16:03 |
| opendevreview | Amir Hossein Ahmadi proposed openstack/kolla-ansible master: Fix ulimit defaults for Debian family container engines https://review.opendev.org/c/openstack/kolla-ansible/+/971801 | 16:05 |
| tafkamax | hmm still same issue, will try again tomorrow with verbose mode | 16:22 |
| opendevreview | Amir Hossein Ahmadi proposed openstack/kolla-ansible master: Fix ulimit defaults for Debian family container engines https://review.opendev.org/c/openstack/kolla-ansible/+/971801 | 18:51 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!