)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":27615,"name":"Rajat Dhasmana","email":"rajatdhasmana@gmail.com","username":"whoami-rajat"},"change_message_id":"11d5d748bfdfb8e899a2e53174f08909bf4207c2","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"a232dc4e_6025c06a","updated":"2023-05-23 11:23:00.000000000","message":"CI is passing. Although A/A is not tested in the CI, it is responsibility of the vendor team to manually test this feature. I am assuming LightBits team has successfully tested their driver in A/A env hence the change so LGTM.","commit_id":"0198e3ecfab9cdd0e4c8d2ec00ea66de95244d66"},{"author":{"_account_id":27615,"name":"Rajat Dhasmana","email":"rajatdhasmana@gmail.com","username":"whoami-rajat"},"change_message_id":"caaeabf643a2d5ef555f646d284679fdc3508481","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"2e6fdfa9_82d1f75d","updated":"2023-05-23 07:18:54.000000000","message":"run-Lightbits CI","commit_id":"0198e3ecfab9cdd0e4c8d2ec00ea66de95244d66"},{"author":{"_account_id":33612,"name":"yuval","email":"yuval@lightbitslabs.com","username":"yuval"},"change_message_id":"14e6b014071f707f2a31f999994b9001c92c7043","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"c38f1c91_893e4b72","in_reply_to":"a232dc4e_6025c06a","updated":"2023-05-24 07:46:52.000000000","message":"hey rajat, Thank you for the review.\nmy testing include 2 cinder machines with cluster mode enabled (in the cinder.conf I add cluster \u003d lightos_cluster)\nthan I ran some sanity tests like create, attach, delete etc...\n\nI have a question regarding the backend name. in both cinder machines, the backend name was identical - is that a problem? so for example \ncinder1 with backendname \u003d lightos, default backend is also lightos\ncinder2 with backendname \u003dlightos, default backend is also lightos\n\nbut cinder.conf are using the cluster mode.\nI see the requests reaching both machines, I wonder if is it because the backendname is identical or because of the cluster mode enabled?\n\nalso are there any suggested tests I should consider?\n\nand again thanks for the review","commit_id":"0198e3ecfab9cdd0e4c8d2ec00ea66de95244d66"},{"author":{"_account_id":9535,"name":"Gorka Eguileor","email":"geguileo@redhat.com","username":"Gorka"},"change_message_id":"346ebc481cd4b5a32aaa8832604346d4ed69f1ea","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"7b4615d2_7193a4d7","in_reply_to":"c38f1c91_893e4b72","updated":"2023-05-24 12:22:36.000000000","message":"Everything in the configuration of the backend cinder.conf section has to be the same, with the exception of the backend_host that CANNOT be the same.\nSo for A/A the host and backend_host configuration options should never be the same on the different cinder-volume services.","commit_id":"0198e3ecfab9cdd0e4c8d2ec00ea66de95244d66"}],"cinder/volume/drivers/lightos.py":[{"author":{"_account_id":9535,"name":"Gorka Eguileor","email":"geguileo@redhat.com","username":"Gorka"},"change_message_id":"346ebc481cd4b5a32aaa8832604346d4ed69f1ea","unresolved":false,"context_lines":[{"line_number":365,"context_line":"    VERSION \u003d \u00272.3.12\u0027"},{"line_number":366,"context_line":"    # ThirdPartySystems wiki page"},{"line_number":367,"context_line":"    CI_WIKI_NAME \u003d \"LightbitsLabs_CI\""},{"line_number":368,"context_line":"    SUPPORTS_ACTIVE_ACTIVE \u003d True"},{"line_number":369,"context_line":""},{"line_number":370,"context_line":"    def __init__(self, *args, **kwargs):"},{"line_number":371,"context_line":"        super(LightOSVolumeDriver, self).__init__(*args, **kwargs)"}],"source_content_type":"text/x-python","patch_set":1,"id":"9cf6df42_9c6ca75b","line":368,"updated":"2023-05-24 12:22:36.000000000","message":"I know that enabling the Active-Active feature in a driver is a vendor responsibility and it\u0027s turned on under their responsibility, but I feel somewhat concern with how the ACLs on a volume are handled, because we retrieve the ACLs from the storage system, then modify it in Cinder, and set the updated value.\n\nWe do this without any locks, so potentially there could be a race condition between 2 requests that modify the ACLs on the same volume.  It\u0027s true that Cinder should prevent it and there should be no issue, but my lack of understanding of their driver and storage system makes me concerned.\n\nAs I said, this is a vendor\u0027s responsibility, so if they say it\u0027s ok, then I\u0027ll trust them.","commit_id":"0198e3ecfab9cdd0e4c8d2ec00ea66de95244d66"}]}