Example Analysis of ceph scrub processing flow
Editor to share with you the example analysis of ceph scrub processing process, I believe that most people do not know much about it, so share this article for your reference, I hope you can learn a lot after reading this article, let's go to know it!
PG::sched_scrub () / PG::_finish_recovery ()
| | _ PG::queue_scrub () |
| | _ state_set (PG_STATE_SCRUBBING) |
| | _ PG::requeue_scrub () |
| | _ OSD::queue_for_scrub () |
| | _ OSD::op_wq.queue () |
PGQueueable::RunVis::operator () (PGScrub)
| | _ PG::scrub () |
| | _ PG::chunky_scrub () |
| | _ _ scrubber.state = initial state of INACTIVE |
| | _ scrubber.epoch_start = info.history.same_interval_since |
| | _ scrubber.state = NEW_CHUNK |
| | _ _ initialize primary_scrubmap and received_maps data structures (used to store scrubmap information of primary and replicas, respectively) |
| | _ _ PGBackend::objects_list_partial () reads the specified number of objects |
| | _ _ traverse the actingbackfill queue and call PG::_request_scrub_map () |
| | _ _ create MOSDRepScrub message object |
| | _ _ OSD::send_message_osd_cluster () sends the MOSDRepScrub message object to the OSDs in the actingbackfill queue |
| | _ scrubber.state = WAIT_PUSHES |
| | _ _ wait for all replicas to send scrub information back to primary, and then set scrubber.state = WAIT_LAST_UPDATE |
| | _ _ for last_update_applied > = scrubber.subset_last_update, set scrubber.state = BUILD_MAP |
| | _ PG::build_scrub_map_chunk () |
| | _ PGBackend::objects_list_range () |
| | _ PGBackend::be_scan_list () |
| | _ _ read the object size information |
| | _ _ read the attribute information of the object |
| | _ _ for those requiring deep scrub, read the object itself and the object property itself, and then calculate the corresponding digest value |
| | _ PG::_scan_rollback_obs () |
| | _ PG::_scan_snaps () |
| | _ _ set scrubber.state = WAIT_REPLICAS |
| | _ _ wait for all replicas to return scrub information, then set scrubber.state = COMPARE_MAPS |
| | _ PG::scrub_compare_maps () |
| | _ PGBackend::be_compare_scrubmaps () |
| | _ PG::_scrub () |
| | _ _ set scrubber.state = WAIT_DIGEST_UPDATE |
| | _ _ after completing the scrub operation for all PG to be scrub, set scrubber.state = FINISH |
| | _ PG::scrub_finish () |
| | _ ReplicatedPG::_scrub_finish () |
The processing flow after Replicas receives the MOSDRepScrub message
ReplicatePG::do_request ()
| | _ PG::replicas_scrub () |
| | _ PG::build_scrub_map_chunk () |
| | _ _ create a MOSDSubOp message object and set the message op = CEPH_OSD_OP_SCRUB_MAP |
| | _ OSD::send_message_osd_cluster () |
The processing flow after Primary receives the MOSDSubOp message
ReplicatedPG::do_sub_op ()
| | _ PG::sub_op_scrub_map () |
| | _ _ write the scrubmap of the received replicas to the scrubber.received_maps |
| | _ _ if scrubmap is returned for all replicas, call the PG::requeue_scrub () function to re-enter the scrub operation |
The above is all the contents of the article "sample Analysis of ceph scrub processing flow". Thank you for reading! I believe we all have a certain understanding, hope to share the content to help you, if you want to learn more knowledge, welcome to follow the industry information channel!
