[MDEV-33134] GCF-1060 test hangs when SR transaction deadlocks trying to complete rollback Created: 2023-12-28  Updated: 2023-12-29

Status: Open
Project: MariaDB Server
Component/s: Galera
Affects Version/s: 10.4.32
Fix Version/s: 10.4

Type: Bug Priority: Major
Reporter: Denis Protivensky Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: HTML File backtrace_all     File mysqld.1.err     File mysqld.2.err    
Issue Links:
Relates
relates to MDEV-32160 GCF-1060 test failure due to wsrep MD... Stalled

 Description   

Local SR transaction gets aborted:

2023-12-28 14:56:16 10 [Note] WSREP: MDL conflict·
schema:  test
request: (10    seqno 57    wsrep (toi, exec, aborted) cmd 3 8  TRUNCATE TABLE t1)
granted: (11    seqno -1    wsrep (local, exec, certifying) cmd 3 5     INSERT INTO t1 VALUE (1,'x'),(2,'x'),(3,'x'))
2023-12-28 14:56:16 10 [Note] WSREP: MDL ticket: type: MDL_SHARED_WRITE space: TABLE db: test name: t1 (Waiting for table metadata lock)
2023-12-28 14:56:16 10 [Note] WSREP: MDL conflict-> BF abort
schema:  test
request: (10    seqno 57    wsrep (toi, exec, aborted) cmd 3 8  TRUNCATE TABLE t1)
granted: (11    seqno -1    wsrep (local, exec, certifying) cmd 3 5     INSERT INTO t1 VALUE (1,'x'),(2,'x'),(3,'x'))
2023-12-28 14:56:16 10 [Note] WSREP: MDL ticket: type: MDL_SHARED_WRITE space: TABLE db: test name: t1 (Waiting for table metadata lock)
2023-12-28 14:56:16 10 [Note] WSREP: Seqno -1 successfully BF aborted 60 victim_seqno -1
2023-12-28 14:56:16 10 [Note] WSREP: client: 11 txc: 60 state: certifying -> must_abort

but later it cannot complete rollback:

2023-12-28 14:56:16 11 [Note] WSREP: after_statement: enter(11,exec,local,deadlock_error,3,toi: -1,nbo: -1)
2023-12-28 14:56:16 11 [Note] WSREP: after_statement_enter
    server: 17006100-a578-11ee-975f-1683cf31f43b, client: 11, state: exec, mode: local
    trx_id: 60, seqno: -1, flags: 1
    state: aborted, bfa_state: certifying, error: deadlock_error, status: 3
    is_sr: 0, frags: 0, frags size: 0, unit: 0, size: 1, counter: 0, log_pos: 0, sr_rb: 0
    own: 1 thread_id: 7febc0145700

hanging here:

#3  0x00007febc9093ee2 in gu::Lock::wait (this=0x7febc0143810, cond=...) at /home/denis/dev/galera-bugs/galerautils/src/gu_lock.hpp:46
#4  0x00007febc909dcdf in galera::Monitor<galera::ReplicatorSMM::CommitOrder>::enter (this=0x5618223f0fa0, obj=...) at /home/denis/dev/galera-bugs/galera/src/monitor.hpp:163
#5  0x00007febc9086a5b in galera::ReplicatorSMM::release_rollback (this=0x5618223efdf0, trx=...) at /home/denis/dev/galera-bugs/galera/src/replicator_smm.cpp:1509
#6  0x00007febc905d278 in galera_release (gh=0x5618223dbb50, ws_handle=0x7febc0143e88) at /home/denis/dev/galera-bugs/galera/src/wsrep_provider.cpp:905
#7  0x00005618208f55a5 in wsrep::wsrep_provider_v26::release (this=<optimized out>, ws_handle=...) at /home/denis/dev/mariadb-server-features/wsrep-lib/src/wsrep_provider_v26.cpp:976
#8  0x00005618208f1d27 in wsrep::transaction::after_statement (this=this@entry=0x7feb6c008990, lock=...) at /home/denis/dev/mariadb-server-features/wsrep-lib/src/transaction.cpp:913
#9  0x00005618208cd947 in wsrep::client_state::after_statement (this=this@entry=0x7feb6c008928) at /home/denis/dev/mariadb-server-features/wsrep-lib/src/client_state.cpp:265


Generated at Thu Feb 08 10:36:38 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.