October 2024 - openstack-discuss - lists.openstack.org

Upstream Oslo IRC Meetings Are Back.
by Daniel Bengtsson 12 Nov '24

12 Nov '24

Hello everyone, I hope you are all doing well! I sent an email a while ago to reestablish the Upstream Oslo IRC meetings. With the removal of Eventlet, this makes even more sense. I have received two positive responses, and we also mentioned it during the PTG last week. I propose to hold these meetings every other Monday, starting today, at 4:00 PM UTC. Does this work for everyone? Thank you for your feedback! -- Daniel Bengtsson Software engineer

6 18

#eventlet-removal - OFTC IRC meeting time slot poll - Express your choice!
by Herve Beraud 12 Nov '24

12 Nov '24

Hey Stackers, We propose to schedule a one hour OFTC IRC meeting (#opentack-eventlet-removal) each first and third week of each month. Let's find together the right UTC time slot for our one hour OFTC IRC meeting. Once the vote is done, meetings will be scheduled each first and third <selected time slot in the poll> of each month. Please, vote here: https://framadate.org/rUghO3j6ZkRhKD7m Votes will be considered as closed by Nov. 8th First meeting will be forced to 4pm UTC on Tuesday Nov 5th and then moved to the selected time slot for all the following meetings. See the proposed time slots in the poll link above. Join us on the #opentack-eventlet-removal OFTC IRC channel. Thanks in advance for your votes -- Hervé Beraud Senior Software Engineer at Red Hat irc: hberaud https://github.com/4383/

1 2

OpenStack Cinder - Volume-ID in cinder DB
by Harsh Ailani 12 Nov '24

12 Nov '24

Hello contributors, I am working on clone operation for a volumegroup and in that the volume ID auto-generated by OpenStack is not the one with which the clone volumes are being created on my IBM storage backend. This is as per design that IBM Storage backend will take the source-volume ID in the source-volumegroup and add a suffix to it while cloning. Now since there is a mismatch in the IDs between cinder and storage backend, because of which I am unable to perform any volume operations on this cloned volume. So, I was wondering if I can change the volume ID in cinder DB? I know that it’s a unique key and there is a limit of 36 characters for this field in cinder DB, but since I cannot change the volume in storage backend, I wanted to check if there is a way I can alter the volume ID in cinder DB? Kindly share your thoughts/suggestions on this matter. Thanks and Regards, Harsh Ailani

2 4

[Cinder][Ceph] Volumetria errada no cluster ceph
by lfsilva＠binario.cloud 07 Nov '24

07 Nov '24

Hello, I'm experiencing an issue with the volume usage of the Ceph cluster, which is currently used in OpenStack for volumes. I am working with Ceph version Octopus (15.2.17). When I run the ceph df command, I get this output: --- RAW STORAGE --- CLASS SIZE AVAIL USED RAW USED %RAW USED hdd 120 TiB 33 TiB 87 TiB 87 TiB 72.72 TOTAL 120 TiB 33 TiB 87 TiB 87 TiB 72.72 --- POOLS --- POOL ID PGS STORED OBJECTS USED %USED MAX AVAIL device_health_metrics 1 32 69 MiB 101 137 MiB 0 11 TiB images 2 256 1.3 TiB 166.90k 2.5 TiB 10.72 11 TiB vms 3 64 574 KiB 21 2.5 MiB 0 11 TiB volumes 4 2048 41 TiB 3.94M 82 TiB 79.42 11 TiB backups 5 1024 407 GiB 111.85k 818 GiB 3.63 11 TiB And when I run the rbd du -p volumes command, I get this total: NAME PROVISIONED USED <TOTAL> 14 TiB 8.3 TiB This "volumes" pool is currently set to replica 2, and mirroring is not enabled. I have checked for any locked or deleted snapshots but found none. I also ran a ceph osd pool deep-scrub volumes, but it didn't resolve the issue. Has anyone encountered this problem before? Could someone provide assistance?

3 21

Re: GPU PCI passthrough woes.
by Andy Speagle 04 Nov '24

04 Nov '24

Aye, that seems to have done it. Strange that this wasn't necessary for my old 1080 cards. Thanks! On Wed, 2024-10-30 at 08:12 +0000, Alexander Dibbo - STFC UKRI wrote: > ⚠️CAUTION⚠️ > This email originated outside of Toyon's network. Do not download > attachments or click links unless you know the content is safe. > > Hi Andy, > > We've seen many issues with pci passthrough and I suspect that the > l40s will have a seen similar issues with the a100s > > Here is an alias string that works for our a100s: > `alias = {"vendor_id": "10de", "product_id": "20f1", "device_type": > "type-PF", "name": "nvidia-tesla-a100-pcie-vga"}` > > I believe it is the `device_type` that will be your issue > > Regards > > Alexander Dibbo – Cloud Architect / Cloud Operations Group Leader > For STFC Cloud Documentation visit > https://urldefense.us/v3/__https://stfc.atlassian.net/wiki/spaces/CLOUDKB/o… > To raise a support ticket with the cloud team please email > cloud-support(a)stfc.ac.uk > To receive notifications about the service please subscribe to our > mailing list at: > https://urldefense.us/v3/__https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=S… > To receive fast notifications or to discuss usage of the cloud please > join our Slack: > https://urldefense.us/v3/__https://stfc-cloud.slack.com/__;!!A4KLn-w15EM4BR… > > -----Original Message----- > From: Andy Speagle <aspeagle(a)toyon.com> > Sent: 30 October 2024 02:15 > To: openstack-discuss(a)lists.openstack.org > Subject: GPU PCI passthrough woes. > > Hey Folks, > > I could use a little assistance getting GPU passthrough working. I > had > this working already for one flavor of nvidia gpu... and I've added > some hosts with a much newer gpu... > > I've updated the pci_alias and pci_passthrough variables and those > seem > to be getting set properly in nova.conf > > passthrough_whitelist = [{"vendor_id":"10de", > "product_id":"1b06"},{"vendor_id":"10de", "product_id":"26b9"}] > > alias = {"name": "gpu", "product_id": "1b06", "vendor_id": "10de"} > alias = {"name": "gpu-l40s", "product_id": "26b9", "vendor_id": > "10de"} > > I believe I have all of the iommu stuff configured and have the pci- > stub module entries... dmesg output shows that the GPUs are being > claimed by the stub module. > > $ openstack flavor show t1.small_gpu_l40s > +----------------------------+--------------------------------------+ > > Field | Value | > +----------------------------+--------------------------------------+ > > OS-FLV-DISABLED:disabled | False | > > OS-FLV-EXT-DATA:ephemeral | 0 | > > access_project_ids | None | > > description | None | > > disk | 0 | > > id | af70c94e-0026-4a39-bc1e-dfb93b286a54 | > > name | t1.small_gpu_l40s | > > os-flavor-access:is_public | True | > > properties | pci_passthrough:alias='gpu-l40s:1' | > > ram | 2048 | > > rxtx_factor | 1.0 | > > swap | 0 | > > vcpus | 1 | > +----------------------------+--------------------------------------+ > > Yet... I can't seem to get an instance to run using that new > flavor... > keeps complaining about there not being enough hosts available. > > Fault: > code: 500 > created: 2024-10-30T02:04:33Z > message: "No valid host was found. There are not enough hosts > available." > details: | > Traceback (most recent call last): > File "/usr/lib/python3/dist- > packages/nova/conductor/manager.py", > line 1580, in schedule_and_build_instances > host_lists = self._schedule_instances(context, > request_specs[0], > File "/usr/lib/python3/dist- > packages/nova/conductor/manager.py", > line 940, in _schedule_instances > host_lists = self.query_client.select_destinations( > File "/usr/lib/python3/dist- > packages/nova/scheduler/client/query.py", line 41, in > select_destinations > return self.scheduler_rpcapi.select_destinations(context, > spec_obj, > File "/usr/lib/python3/dist-packages/nova/scheduler/rpcapi.py", > line 160, in select_destinations > return cctxt.call(ctxt, 'select_destinations', **msg_args) > File "/usr/lib/python3/dist- > packages/oslo_messaging/rpc/client.py", line 189, in call > result = self.transport._send( > File "/usr/lib/python3/dist- > packages/oslo_messaging/transport.py", line 123, in _send > return self._driver.send(target, ctxt, message, > File "/usr/lib/python3/dist- > packages/oslo_messaging/_drivers/amqpdriver.py", line 689, in send > return self._send(target, ctxt, message, wait_for_reply, > timeout, > File "/usr/lib/python3/dist- > packages/oslo_messaging/_drivers/amqpdriver.py", line 681, in _send > raise result > nova.exception_Remote.NoValidHost_Remote: No valid host was > found. > There are not enough hosts available. > > Traceback (most recent call last): > File "/usr/lib/python3/dist- > packages/oslo_messaging/rpc/server.py", line 241, in inner > return func(*args, **kwargs) > > File "/usr/lib/python3/dist- > packages/nova/scheduler/manager.py", > line 223, in select_destinations > selections = self._select_destinations( > > File "/usr/lib/python3/dist- > packages/nova/scheduler/manager.py", > line 250, in _select_destinations > selections = self._schedule( > > File "/usr/lib/python3/dist- > packages/nova/scheduler/manager.py", > line 416, in _schedule > self._ensure_sufficient_hosts( > > File "/usr/lib/python3/dist- > packages/nova/scheduler/manager.py", > line 455, in _ensure_sufficient_hosts > raise exception.NoValidHost(reason=reason) > > nova.exception.NoValidHost: No valid host was found. There are > not > enough hosts available. > > Any clues on how/where to dig into this more to see what might be > missing? Thanks. > > -- > Andy Speagle > -- Andy Speagle Sr. Site Reliability Engineer Toyon Research Corporation 316.617.2431

3 2

Is HPET recommended for x86 instance?
by Jiatong Shen 04 Nov '24

04 Nov '24

Hello Community Experts, I found that to enable the HPET timer for x86 guests, the operator should set hw_time_hpet on the given image. By default, HPET is disabled. My question is what are the pros and cons for using HPET? Thanks. -- Best Regards, Jiatong Shen

2 3

[neutron][ovn-octavia-provider][octavia][ovn] Octavia OVN DBs sync cmd
by Rico Lin 04 Nov '24

04 Nov '24

Hi all I would like to raise attention to the patch set that introduces Octavia OVN DBs sync cmd. patches: https://review.opendev.org/q/topic:%22add-ovn-db-sync-utils%22 With command `octavia-ovn-db-sync-util {--debug}` It should allows us to do the OVN DB sync similar to what `neutron-ovn-db-sync-util` does. And thanks to Fernando Royo, the patches are small and earlier to review in focused functions. Would be great if we could have some more reviews for it. *Rico Lin*

2 1

[devstack][nova][cinder][ironic][glance][swift][neutron][all] Deprecating/removing non-uWSGI deployment mechanisms?
by Stephen Finucane 01 Nov '24

01 Nov '24

o/ tl;dr: I'd like to remove the ability to deploy services under mod_wsgi from DevStack. I'd also like to remove the ability to deploy services under eventlet, where this feature is not currently being tested, and deprecate it where it is. I took a foray through DevStack last week and realised that, depending on the service in question, there are currently 3 possible ways an API can be deployed by DevStack: * Eventlet server * Apache + uWSGI * Apache + mod_wsgi In addition, there are a number of overlapping config options that could be set to modify which of these was in use: * ENABLE_HTTPD_MOD_WSGI_SERVICES * WSGI_MODE * Service-specific options like NOVA_USE_MOD_WSGI, CINDER_USE_MOD_WSGI etc. Potential confusion aside, there are two issues I see with this. Firstly, and most obviously, we are working on eventlet removal, which will necessarily involve eventlet server removal [1]. In addition, there is a goal [2][3] to remove use of '.wsgi' scripts in favour of Python module paths. Quite a few services have migrated to this already, and those that haven't are going to need to do so by time pip 25.0 comes out (since that removes the deprecated legacy editable install flow [4] meaning these files will not be generated by pbr and DevStack deployments will start failing...you have been warned). However, mod_wsgi need .wsgi scripts. I didn't realise we cared about mod_wsgi in DevStack still before last week, so to complete the wsgi script -> module migration for services that can be deployed with mod_wsgi (e.g. Keystone), we're going to have to either (a) package a .wsgi script in DevStack or (b) remove the ability to deploy with mod_wsgi. My immediate thought on this was to simply remove all non-uWSGI deployment paths immediately and deprecate the config options for removal to force DevStack plugins to adapt. However, some services still want to test eventlet server for their API while they stabilise support for standard WSGI-based deployments (neutron), while others explicitly rely on evenlet to generate a single standalone binary (ironic, glance, ...). That means we can't just go remove everything now and must instead take a phased approach. I would therefore like to propose we do the following: * Remove the ability to deploy services under mod_wsgi from DevStack and deprecate/remove the WSGI_MODE flag. To be clear, other deployment tools can continue to use mod_wsgi but DevStack won't. * Entirely remove the ability to deploy under eventlet from services that are not currently testing this. This includes Nova and Cinder. * Invert the default for services that are testing eventlet, so that they test with uWSGI by default now and jobs will need to opt-in to eventlet rather than opt-in to uWSGI. I think this just affects Neutron. * Optionally, and likely only one Neutron is no longer testing eventlet (next cycle?), deprecate ENABLE_HTTPD_MOD_WSGI_SERVICES and {SERVICE}_USE_MOD_WSGI config options entirely and replace it with {SERVICE}_STANDALONE options like Glance has. This will affect Ironic and maybe more. I have proposed a series that will do much of this. The WIP patches indicate the things that I am less clear about, such as Swift's readiness for deploying under uWSGI. I would be grateful if folks from the individual service teams could take a look and let me know what they think: https://review.opendev.org/c/openstack/devstack/+/932193/1 Cheers, Stephen [1] https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack… [2] https://review.opendev.org/c/openstack/governance/+/902807 [3] https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack… [4] https://ichard26.github.io/blog/2024/08/whats-new-in-pip-24.2/

6 10

[openstack helm] [Cinder] [Ceph] keyring not getting decode when doing cinder-storage-init
by daniel890723＠gmail.com 31 Oct '24

31 Oct '24

Hello today i was testing to deploy openstack with helm in my kubernetes cluster and i encounter an error when trying to deploy cinder Error: UPGRADE FAILED: post-upgrade hooks failed: 1 error occurred: * job cinder-storage-init failed: BackoffLimitExceeded when i print the log of the cinder-storage-init pod it shows 2024-10-30T03:22:14.303+0000 7efc1353d640 -1 auth: failed to load /etc/ceph/ceph.client.admin.keyring: (5) Input/output error 2024-10-30T03:22:14.303+0000 7efc1353d640 -1 auth: error parsing file /etc/ceph/ceph.client.admin.keyring: error setting modifier for [client.admin] type=key val=: Malformed input [buffer:3] 2024-10-30T03:22:14.303+0000 7efc1353d640 -1 auth: failed to load /etc/ceph/ceph.client.admin.keyring: (5) Input/output error 2024-10-30T03:22:14.303+0000 7efc1353d640 -1 monclient: keyring not found [errno 5] RADOS I/O error (error connecting to the cluster) and the keyring secret is use looks like this: apiVersion: v1 data: key: AQDtmCFnh3STCBAAGKQIyIGuZjnmX3L4mPCQXw== kind: Secret metadata: name: pvc-ceph-client-key namespace: openstack after checking secret and the rook-ceph cluster works fine i change the template and makes the pod not executing the command so i can look inside the pod and it's keyrings the keyring looks like this: [client.admin] key = �!g�tȁ�f9�_r��_ so apparently the keyring secert or when it mount into the pod got some problem and change the key into the normal one fix the problem is there a reason why it cause this ? because the secret and ceph-cluster are all works fine

2 4

[manila] 2025.1 Epoxy PTG summary
by Carlos Silva 31 Oct '24

31 Oct '24

Hello everyone! Thank you for the great participation at the PTG last week. We've had great discussions and a good turnout. The recordings for the sessions are available on YouTube [0]. If you would like to check on the notes, please take a look at the PTG etherpad [1]. *2024.2 Dalmatian Retrospective* ========================== - New core reviewers in the manila group were impactful in reviews, we should continue actively working on maintaining/growing the core reviewer team. - We had the mid-cycle and managed to combine it with our well-known collaborative review sessions, around feature proposal freeze. This had a good impact on raising awareness on the changes being proposed, as well as prioritizing the reviews. - Great contributions ranging from new third party drivers to successful internships on manila-ui, bandit and the ongoing OpenAPI internships. *Action items:* - Carlos (carloss) will work with the manila team to help people gain context on the bug czar role and work with the team to rotate it. - Vida Haririan (vhari) will jot down the details of the Bug Czar role - Follow the discussions on teams joining the VMT and get Manila included too. - Spread the word on the removal of the manila client and switch to OpenStackClient *Share backup enhancements* ========================= - Out of place restore isn't supported currently. We have agreed that this is a good use case and that a design specification should be proposed to document this. - DataManager / BackupDriver - forcing the backup process to go through the DataManager service is supported through a config option, but Manila is currently not honoring it. We agreed that this is an issue in the code, and we will review the proposed change [2] to make the data manager honor this config. - DataManager to allow for a backup driver to provide reports on API call progress: Currently, the data manager fetches the progress of a backup using a generic get progress call, but it is failing with the generic backup driver. We suggested that this should be fixed in the base driver. - Context for Backup API calls: currently, only objects representing a Share and Backup are passed to the backup driver. The request context should also be forwarded in these calls. The backup driver interface can be changed for this, but we should be mindful of out of tree drivers that could break. *Action items:* - Zach Goggins (zachgoggins) will look into: - Proposing a spec for the share backup out of place restore. - Updating the backup driver interface and adding context to the methods that need it. - Updating the backup driver interface and adding the abstract methods/capabilities that will help with the `get_restore_progress` and `get_backup_progress` methods. - The manila team will provide feedback on [2] *All things CephFS* =============== *Updates from previous cycles* ------------------------------------------ *State of the Standalone NFS Ganesha protocol helper:* - We added a deprecation warning at the end of the previous SLURP release, and we are planning to complete the removal during the 2025.1/Epoxy release. There were no objections to this so far at the PTG. When this is removed, CephFS-via-NFS will only work with cephadm deployed ceph-nfs clusters. *Testing and stabilization:* - devstack-plugin-ceph has been refactored to deploy a standalone NFS-Ganesha service with a ceph orch deployed cluster. We also dropped support for package-based and container-based installation of ceph. cephadm is used to deploy/orchestrate ceph. - Bumped Ceph version to Reef in Antelope, Bobcat, Caracal, Dalmatian, as well as started testing with Squid. - There are some failures on stable branches jobs which are being triaged and fixed. *Manage/unmanage:* - Implementation completed in Dalmatian and the documentation has been updated. We are currently working to enable the tests on CI permanently, as well as doing some small refactors to the CI jobs. *Ensure shares:* - Merged in Dalmatian but testing is still challenging, as running the tests mean that the service would temporarily have a different status and shares within the backend would have their status changed, which is harmful for test concurrency. *Preferred export locations and export location metadata:* - The core feature merged, but we are still working to get the newly implemented tests passing and merged. *Plans for 2025.1/Epoxy* -------------------------------- - NFSv3 + testing: we are looking into enabling NFSv3 support as soon as the patch is merged in Ceph. We agreed that we should enable the tests within manila-tempest-plugin and make any necessary changes to the tests structure, so we can ensure that we are testing some scenarios with both NFSv3 and NFSv4. - We will start to investigate support for SMB/CIFS shares and look at the necessary changes for setting up devstack and testing. *Action items:* - Carlos (carloss) will write an email to the openstack-discuss mailing list announcing the removal of the deprecated ganesha helper - Carlos pursue the manage/unmanage testing patches to have tests enabled in the CephFS jobs during Epoxy. - Carlos will look into approaches to test ensure shares APIs. - Ashley (ashrod98) will continue working on the export location metadata tempest changes and drive them to completion. - The manila team will look into updating manila-tempest-plugin tests and enabling NFSv3 tests in the Ceph NFS jobs - Goutham (gouthamr) will be submitting a prototype of the SMB/CIFS integration *Tech Debt* ======== *Eventlet removal* ---------------------- Our main concerns: - Performance should not be degraded with the default configuration when we switch. - Synchronous calls do not take a big hit and become asynchronous. - Impact to the SSH Pool (used by many drivers) should be minimal. *Action items for 2025.1 Epoxy:* - Tackle the low-hanging-fruit changes. - Participating in the pop-up team discussions. - Removing the affected console scripts in Manila. - Working on performance tests to understand what will be the impact on the SSH pool that is used by some drivers. - Look into enhancing our rally/browbeat test coverage. *CI and testing images* ------------------------------- We started working on the migration of the CI to Ubuntu 24.04 in all of the manila repositories (manila-image-elements, python-manilaclient, manila-ui, manila, manila-specs). Currently, the Ceph job is broken [3]. *Action items:* - We should clean up our CI job variants, as they have a lot of workarounds and we can start moving away from them. *Stable branches* ---------------------- We currently have 5 "unmaintained" branches, so we should be looking at sunsetting them. *Action items:* - Carlos (carloss) will start the conversation for the transition of some of these branches in the openstack-discuss mailing list. *Allowing force delete of a share network subnet* ======================================= We currently can add subnets (which translates to adding new network interfaces) to a share server but we can't remove them. This is a proposal to add this removal feature and being able to detach a network interface of a share server. We agreed that: - This is a good use case and something that can be enhanced. - The enhancement should add a force-delete API. - We should not allow the last subnet to be deleted, otherwise the shares won't have an export path. - A bug should be filed for a tangential issue that the NetApp driver is using "neutron_net_id" (and possibly "neutron_subnet_id" to name resources on the backend: ipspaces, broadcast domains, and possibly concurrency control / locks. *Action Items:* - sylvanld will look into proposing a spec to document this behavior *NetApp Driver: Enforcing lifs limit per HA pair* ===================================== - The NetApp ONTAP storage has a limit of network interfaces per node in a HA pair. In case the sum of allocated network interfaces in the two nodes of the HA pair is bigger than the limit of the single node, then the failover operation is compromised and will fail. - NetApp maintainers would like to fix this issue, and we agreed that: - The fix should be as dynamic as possible, not relying on users/admin input or configuration. - The ONTAP driver must look up all of the interfaces already created and allow/deny the request in case it would compromise the failover. - The NetApp ONTAP driver should keep an updated capability with the max network interfaces support number, and possible the number of allocated network interfaces at the moment. *NetApp Driver: Implement Certificate based authentication* ================================================ - The NetApp ONTAP driver currently handles only user/password authentication, but in an environment that password should change quarterly, this means updating the local.conf at least every three months. This enhancement proposes also adding the possibility of adding certificate based authentication. - We agreed that this is something that is going to be important for operators and will allow them to add their certificates with a longer expiration date, avoiding the disruptions caused by needing to update the user/password. *Manage Share affinity relationships by annotation/label* ============================================= Currently the manila scheduler uses affinity/anti-affinity hints and we base ourselves on share IDs. The idea now would be to have the affinite hints to be based in an affinity policy, as possible with Nova. We considered the proposed approaches, and agreed that: - If we are adding new policies, they should end up becoming a new resource/entity within the manila database - If there is a way to reuse the share groups mechanism, we should prioritize it *Action items:* - Chuan (chuanm) will propose a design spec to document this new behavior. *Share encryption* ============== This feature is currently waiting for more reviews and testing on gerrit. In the Dalmatian release mid-cycle we talked about the importance of testing this feature against a first party driver, to ensure that the APIs and integration with Barbican and Castellan work. We agreed that: - We should do some research on how to do this testing with the generic driver (which uses Cinder and Nova) - The testing will focus on the APIs and behavior of this feature, not the encryption of the shares. *Action items:* - gouthamr will help with some research on how to test this with the generic driver - The manila team will discuss this again in the upcoming manila weekly meetings. [0] https://www.youtube.com/watch?v=8UxrjEr6yik&list=PLnpzT0InFrqDHGfSDPhiGtSeX… [1] https://etherpad.opendev.org/p/epoxy-ptg-manila [2] https://review.opendev.org/c/openstack/manila/+/907983 [3] https://www.spinics.net/lists/ceph-users/msg83201.html

1 0