March 2024 - openstack-discuss - lists.openstack.org

[swift] Multi-policy container
by m.victorello＠luizalabs.com 13 Mar '24

13 Mar '24

Hello, Is there any plan to support multi-policy containers? Why the container's database have a 'storage_policy_index' column for objects? As far as I know the policy index comes from the container (broker.get_info). I don't know that much about changing the policy of a container, i think those are related.

2 1

[magnum] ClusterAPI driver out of tree
by Jake Yip 13 Mar '24

13 Mar '24

Hi all, After the previous discussions[1] on the ML, Magnum and StackHPC have decided to create and maintain separate repos[2] for StackHPC's contributed cluster api driver. The repos are currently 1. magnum-capi-helm - driver implementation 2. magnum-capi-helm-charts - helm charts used by driver This will allow the driver to be out of tree, yet still be in governance. StackHPC and Magnum Core team will work together to test the driver in CI. I hope this will please everyone and be a way forward. Regards, Jake [1] https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack… [2] https://review.opendev.org/c/openstack/governance/+/910240 [3] https://review.opendev.org/c/openstack/project-config/+/910239 -- Jake Yip Technical Lead, ARDC Nectar Research Cloud

1 0

opensearch dashboards example
by Satish Patel 13 Mar '24

13 Mar '24

Folks, Recently I have deployed opensearch logging for ELK replacement and trying to create some visualization dashboard but its new tool and learning curve. Will appreciate If anyone has any example dashboard or visualizer for logging etc. ~S

1 0

Re: [keystone] application creds issue with permission
by Satish Patel 13 Mar '24

13 Mar '24

Hi Artem, You are 100% correct here, I have noticed as soon as I give the reader the role it works. Thank you! On Tue, Mar 12, 2024 at 4:02 PM Artem Goncharov <artem.goncharov(a)gmail.com> wrote: > Can you please check which roles are bound to the appcred? The reason I am > asking is https://review.opendev.org/c/openstack/keystone/+/910337 (wrong > treatment of implied roles when you create appcred explicitly specifying > role that implies another role (basically if you create appcred with member > role - reader role is not given) > > ---- > typed from mobile, auto-correct typos assumed > ---- > > On Tue, Mar 12, 2024, 20:57 Satish Patel <satish.txt(a)gmail.com> wrote: > >> Folks, >> >> I am running the 2023.1 release of openstack using kolla-ansible. I have >> integrated with LDAP and everything works fine. But today when I created >> application creds and used that to access openstack api I got all kinds of >> strange auth errors. >> >> #!/usr/bin/env bash >> export OS_AUTH_TYPE=v3applicationcredential >> export OS_AUTH_URL=https://openstack-eng.example.com:5000 >> export OS_IDENTITY_API_VERSION=3 >> export OS_REGION_NAME="eng" >> export OS_INTERFACE=public >> export OS_APPLICATION_CREDENTIAL_ID=ee17300916b1401f912f6140ce9cd642 >> export OS_APPLICATION_CREDENTIAL_SECRET=XXXXXXXXX >> >> # openstack server list >> ForbiddenException: 403: Client Error for url: >> https://openstack-eng.example.com:8774/v2.1/servers/detail?deleted=False, >> Policy doesn't allow os_compute_api:servers:detail to be performed. >> >> >> # openstack image list >> ForbiddenException: 403: Client Error for url: >> https://openstack-eng.example.com:9292/v2/images, You are not authorized >> to complete get_images action.<br /><br /> >> >> What is wrong here because I have one more setup of openstack which is >> not using LDAP but in that cloud application creds working fine. >> >> # cat /etc/keystone/keystone.conf >> >> [DEFAULT] >> debug = False >> transport_url = hiding.... >> log_file = /var/log/kolla/keystone/keystone.log >> use_stderr = True >> >> [oslo_middleware] >> enable_proxy_headers_parsing = True >> >> [database] >> connection = mysql+pymysql:// >> keystone:hiding...@openstack-eng.example.com:3306/keystone >> connection_recycle_time = 10 >> max_pool_size = 1 >> max_retries = -1 >> >> [identity] >> domain_specific_drivers_enabled = true >> domain_config_dir = /etc/keystone/domains >> >> [token] >> revoke_by_id = False >> provider = fernet >> expiration = 86400 >> allow_expired_window = 172800 >> >> [fernet_tokens] >> max_active_keys = 3 >> >> [cache] >> backend = oslo_cache.memcache_pool >> enabled = True >> memcache_servers = 10.0.25.201:11211,10.0.25.202:11211,10.0.25.203:11211 >> >> [oslo_messaging_notifications] >> transport_url = rabbit://openstack:hiding.... >> driver = noop >> >> [oslo_messaging_rabbit] >> heartbeat_in_pthread = True >> amqp_durable_queues = true >> rabbit_quorum_queue = true >> kombu_reconnect_delay = 0.5 >> rabbit_transient_queues_ttl = 60 >> >> >> # cat /etc/keystone/domains/keystone.eng.conf >> >> # Ansible managed >> >> [identity] >> driver = ldap >> >> [ldap] >> debug_level = 4095 >> group_allow_create = False >> group_allow_delete = False >> group_allow_update = False >> group_id_attribute = cn >> hiding.... >> hiding.... >> >>

1 0

[keystone] application creds issue with permission
by Satish Patel 12 Mar '24

12 Mar '24

Folks, I am running the 2023.1 release of openstack using kolla-ansible. I have integrated with LDAP and everything works fine. But today when I created application creds and used that to access openstack api I got all kinds of strange auth errors. #!/usr/bin/env bash export OS_AUTH_TYPE=v3applicationcredential export OS_AUTH_URL=https://openstack-eng.example.com:5000 export OS_IDENTITY_API_VERSION=3 export OS_REGION_NAME="eng" export OS_INTERFACE=public export OS_APPLICATION_CREDENTIAL_ID=ee17300916b1401f912f6140ce9cd642 export OS_APPLICATION_CREDENTIAL_SECRET=XXXXXXXXX # openstack server list ForbiddenException: 403: Client Error for url: https://openstack-eng.example.com:8774/v2.1/servers/detail?deleted=False, Policy doesn't allow os_compute_api:servers:detail to be performed. # openstack image list ForbiddenException: 403: Client Error for url: https://openstack-eng.example.com:9292/v2/images, You are not authorized to complete get_images action.<br /><br /> What is wrong here because I have one more setup of openstack which is not using LDAP but in that cloud application creds working fine. # cat /etc/keystone/keystone.conf [DEFAULT] debug = False transport_url = hiding.... log_file = /var/log/kolla/keystone/keystone.log use_stderr = True [oslo_middleware] enable_proxy_headers_parsing = True [database] connection = mysql+pymysql:// keystone:hiding...@openstack-eng.example.com:3306/keystone connection_recycle_time = 10 max_pool_size = 1 max_retries = -1 [identity] domain_specific_drivers_enabled = true domain_config_dir = /etc/keystone/domains [token] revoke_by_id = False provider = fernet expiration = 86400 allow_expired_window = 172800 [fernet_tokens] max_active_keys = 3 [cache] backend = oslo_cache.memcache_pool enabled = True memcache_servers = 10.0.25.201:11211,10.0.25.202:11211,10.0.25.203:11211 [oslo_messaging_notifications] transport_url = rabbit://openstack:hiding.... driver = noop [oslo_messaging_rabbit] heartbeat_in_pthread = True amqp_durable_queues = true rabbit_quorum_queue = true kombu_reconnect_delay = 0.5 rabbit_transient_queues_ttl = 60 # cat /etc/keystone/domains/keystone.eng.conf # Ansible managed [identity] driver = ldap [ldap] debug_level = 4095 group_allow_create = False group_allow_delete = False group_allow_update = False group_id_attribute = cn hiding.... hiding....

1 0

[ansible][aio][glance] Cannot create image, TypeError: Response is undefined
by Tang Chai Keong 12 Mar '24

12 Mar '24

I deployed openstack-ansible AIO and is trying to create a new image but is encountering "TypeError: Response is undefined" and "Error: Unable to create the image". The image creation task is just stuck in the "queued" mode forever after that. I also do not see any of the service logs in the lxc containers under /var/logs/ folder. Appreciate any advice on how I can troubleshoot this. Thank you. Regards, Ben

2 2

[tc] Next Meeting Tomorrow, March 12, 1800 UTC
by Jay Faulkner 12 Mar '24

12 Mar '24

Hi all, the next weekly meeting of the OpenStack Technical Committee is tomorrow, March 12, at 1800 UTC. It will be held on #openstack-tc in OFTC. *US-based contributors please note that the meeting time does not change with DST.* The proposed agenda is: - Roll call - Follow up on tracked action items - Gate health check - Implementation of Unmaintained branch statuses - Testing runtime for 2024.2 release (gmann) - https://review.opendev.org/c/openstack/governance/+/908862 - TC vPTG 2024.2 / TC Tracker 2024.1 - https://etherpad.opendev.org/p/tc-2024.1-tracker - https://etherpad.opendev.org/p/apr2024-ptg-os-tc - Open Discussion and Reviews Thanks, Jay Faulkner OpenStack TC Chair

1 0

[swift][s3api] put-object-acl in SLO objects
by Thiago De Moraes Teixeira 12 Mar '24

12 Mar '24

Hi, I've got this error when trying to put acl in objects that are a multipart upload with s3api client. *aws s3api put-object-acl --bucket ramac --key darwin --acl public-readAn error occurred (BadDigest) when calling the PutObjectAcl operation (reached max retries: 4): The Content-MD5 you specified did not match what we received.* And for objects greater than 5Gb the error is: *An error occurred (EntityTooLarge) when calling the PutObjectAcl operation: Your proposed upload exceeds the maximum allowed object size.* I noticed this is happening because s3 middleware did not recognize this object as an SLO, so when the request reaches the ServerSideCopyMiddleware, the request did not have the query param "multipart-manifest". I tried this swift patch, to solve this problem, identifying if the object is an SLO, on head object in acl_handlers.py class, and retrieving this information before request transformation to swift_request, adding the query param. While this patch resolves the immediate issue, I believe there's room for improvement. I welcome any suggestions or feedback on enhancing this patch further. Your insights would be greatly appreciated. https://paste.openstack.org/show/bd2nAHIHPtUtDNbWLorD/ Best regards -- _‘Esta mensagem é direcionada apenas para os endereços constantes no cabeçalho inicial. Se você não está listado nos endereços constantes no cabeçalho, pedimos-lhe que desconsidere completamente o conteúdo dessa mensagem e cuja cópia, encaminhamento e/ou execução das ações citadas estão imediatamente anuladas e proibidas’._ * **‘Apesar do Magazine Luiza tomar todas as precauções razoáveis para assegurar que nenhum vírus esteja presente nesse e-mail, a empresa não poderá aceitar a responsabilidade por quaisquer perdas ou danos causados por esse e-mail ou por seus anexos’.*

1 0

[swift][middleware][s3] Content-type of s3api versioning response
by m.victorello＠luizalabs.com 11 Mar '24

11 Mar '24

Hello, We are having some compatibility problems caused by the lack of content-type in this part: https://github.com/openstack/swift/blob/5e5fde12400fc76b9df43b3fcd13eb8a226… def GET(self, req): """ Handles GET Bucket versioning. """ sysmeta = req.get_container_info(self.app).get('sysmeta', {}) elem = Element('VersioningConfiguration') if sysmeta.get('versions-enabled'): SubElement(elem, 'Status').text = ( 'Enabled' if config_true_value(sysmeta['versions-enabled']) else 'Suspended') body = tostring(elem) return HTTPOk(body=body, content_type=None) Is there a reason for it?

1 0

[swift][s3][multipart] S3 multipart re-upload leaves orphan segments
by m.victorello＠luizalabs.com 11 Mar '24

11 Mar '24

Hello, I understand that this is a known issue for some time, but I dont know if it has been resolved. As far as I know stable/2023.2 still has this problem, is there any related work? We think there are two fronts: - Segments of incomplete multipart, forgotten by the user - Completed re-uploads We plan to fix those: - Using the object-expirer in a new middleware - Deploying workers to review all files, checking if they are still in use What are your thoughts?

1 0