Bug #37525: unprime_split_children may discard query - RADOS - Ceph

Actions

Copy link

Bug #37525

closed

unprime_split_children may discard query

Added by Sage Weil over 5 years ago. Updated over 5 years ago.

Status:

Resolved

Priority:

Urgent

Assignee:

Sage Weil

Category:

Target version:

% Done:

Source:

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Component(RADOS):

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

2018-12-04 15:38:00.577 7fcb86729700  1 -- 172.21.15.60:6813/40472 <== osd.0 172.21.15.60:6805/13037 886 ==== pg_query(8.18 epoch 423) v4 ==== 162+0+0 (1807327245 0 0) 0x106ab400 con 0xd242d00
2018-12-04 15:38:00.577 7fcb86729700  7 osd.3 420 handle_fast_pg_query pg_query(8.18 epoch 423) v4 from osd.0
2018-12-04 15:38:00.577 7fcb86729700 15 osd.3 420 enqueue_peering_evt 8.18 epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)
2018-12-04 15:38:00.577 7fcb86729700 20 osd.3 op_wq(0) _enqueue OpQueueItem(8.18 PGPeeringEvent(epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)) prio 255 cost 10 e423)
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq(0) _process 8.18 (new) to_process <> waiting <> waiting_peering {}
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq(0) _process OpQueueItem(8.18 PGPeeringEvent(epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)) prio 255 cost 10 e423) queued
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq(0) _process 8.18 to_process <OpQueueItem(8.18 PGPeeringEvent(epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)) prio 255 cost 10 e423)> waiting <> waiting_peering {}
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq(0) _process OpQueueItem(8.18 PGPeeringEvent(epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)) prio 255 cost 10 e423) pg 0
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq(0) _process 8.18 map 423 > 420
2018-12-04 15:38:00.577 7fcb60da8700 20 osd.3 op_wq _add_slot_waiter 8.18 peering, item epoch is 423, will wait on OpQueueItem(8.18 PGPeeringEvent(epoch_sent: 423 epoch_requested: 423 MQuery 8.18 from 0 query_epoch 423 query: query(info 0'0 epoch_sent 423)) prio 255 cost 10 e423)
...
2018-12-04 15:38:03.205 7fcb7429b700 20 osd.3:0.identify_splits_and_merges identify_splits_and_merges slot 8.18 has no pg and waiting_for_split 
...
2018-12-04 15:38:03.205 7fcb7429b700 20 osd.3:0.consume_map consume_map 8.18
2018-12-04 15:38:03.205 7fcb7429b700 20 osd.3:0.consume_map consume_map  8.18 waiting for split 421
...
2018-12-04 15:38:03.205 7fcb60da8700 20 osd.3 421 try_finish_pg_delete 8.8 0x103a5400
2018-12-04 15:38:03.205 7fcb60da8700 10 osd.3:0._detach_pg 8.8 0x103a5400
2018-12-04 15:38:03.205 7fcb60da8700 10 osd.3:0.unprime_split_children unprime_split_children parent 8.8 clearing 8.18

/a/sage-2018-12-04_12:53:56-rados-wip-sage2-testing-2018-12-03-2211-distro-basic-smithi/3306155