Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Done
-
None
Description
mleich produced an rr replay trace where multiple purge coordinator tasks are running simulataneously.
ssh pluto
|
rr replay /data/results/1654860797/001037/1/rr/latest-trace
|
|
bb-10.10-MDEV-11026 2acc2601910d5292f1ea46e06e5affeb56dc8010 |
(rr) c
|
Continuing.
|
mysqld: /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1186: ulint trx_purge_attach_undo_recs(ulint): Assertion `node->undo_recs.empty()' failed.
|
|
|
Thread 3 received signal SIGABRT, Aborted.
|
[Switching to Thread 2420074.2420795]
|
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
|
50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
|
(rr) backtrace
|
#0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
|
#1 0x000073e3778e7859 in __GI_abort () at abort.c:79
|
#2 0x000073e3778e7729 in __assert_fail_base (
|
fmt=0x73e377a7d588 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n",
|
assertion=0x55a3b83be3e0 "node->undo_recs.empty()",
|
file=0x55a3b83bc860 "/data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc", line=1186, function=<optimized out>) at assert.c:92
|
#3 0x000073e3778f8f36 in __GI___assert_fail (
|
assertion=0x55a3b83be3e0 "node->undo_recs.empty()",
|
file=0x55a3b83bc860 "/data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc", line=1186,
|
function=0x55a3b83be380 "ulint trx_purge_attach_undo_recs(ulint)")
|
at assert.c:101
|
#4 0x000055a3b71a4f75 in trx_purge_attach_undo_recs (n_purge_threads=1)
|
at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1186
|
#5 0x000055a3b71a5e04 in trx_purge (n_tasks=1, truncate=false)
|
at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1346
|
#6 0x000055a3b717b874 in purge_coordinator_state::do_purge (
|
this=0x55a3ba001500 <purge_state>)
|
at /data/Server/bb-10.10-MDEV-11026/storage/innobase/srv/srv0srv.cc:1611
|
#7 0x000055a3b71794e5 in purge_coordinator_callback ()
|
at /data/Server/bb-10.10-MDEV-11026/storage/innobase/srv/srv0srv.cc:1765
|
#8 0x000055a3b752e6b8 in tpool::task_group::execute (
|
…
|
(rr) break trx0purge.cc:1254
|
(rr) reverse-continue
|
Continuing.
|
|
|
Thread 3 received signal SIGABRT, Aborted.
|
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
|
50 in ../sysdeps/unix/sysv/linux/raise.c
|
(rr) reverse-continue
|
Continuing.
|
[Switching to Thread 2420074.2420757]
|
|
|
Thread 9 hit Breakpoint 1, trx_purge_attach_undo_recs (n_purge_threads=2) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1254
|
1254 node->undo_recs.push(purge_rec);
|
(rr) continue
|
Continuing.
|
mysqld: /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1186: ulint trx_purge_attach_undo_recs(ulint): Assertion `node->undo_recs.empty()' failed.
|
|
|
Thread 3 received signal SIGABRT, Aborted.
|
[Switching to Thread 2420074.2420795]
|
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
|
50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
|
(rr) thread apply 9 backtrace
|
|
|
Thread 9 (Thread 2420074.2420757):
|
#0 0x0000000067df63b7 in sched_yield () at /home/roc/rr/rr/src/preload/overrides.c:215
|
#1 0x000055a3b70352ee in __gthread_yield () at /usr/include/x86_64-linux-gnu/c++/9/bits/gthr-default.h:693
|
#2 0x000055a3b7035389 in std::this_thread::yield () at /usr/include/c++/9/thread:356
|
#3 0x000055a3b71a3f93 in trx_purge_choose_next_log () at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1018
|
#4 0x000055a3b71a466d in trx_purge_get_next_rec (n_pages_handled=0x29b97acab720, heap=0x621000050100) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1083
|
#5 0x000055a3b71a4ba8 in trx_purge_fetch_next_rec (roll_ptr=0x29b97acab768, n_pages_handled=0x29b97acab720, heap=0x621000050100) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1154
|
#6 0x000055a3b71a5511 in trx_purge_attach_undo_recs (n_purge_threads=2) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1225
|
#7 0x000055a3b71a5e04 in trx_purge (n_tasks=2, truncate=false) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/trx/trx0purge.cc:1346
|
#8 0x000055a3b717b874 in purge_coordinator_state::do_purge (this=0x55a3ba001500 <purge_state>) at /data/Server/bb-10.10-MDEV-11026/storage/innobase/srv/srv0srv.cc:1611
|
#9 0x000055a3b71794e5 in purge_coordinator_callback () at /data/Server/bb-10.10-MDEV-11026/storage/innobase/srv/srv0srv.cc:1765
|
#10 0x000055a3b752e6b8 in tpool::task_group::execute (this=0x55a3ba001980 <purge_coordinator_task_group>, t=0x55a3ba001a40 <purge_coordinator_task>) at /data/Server/bb-10.10-MDEV-11026/tpool/task_group.cc:68
|
The design constraint is that there is only one purge coordinator task running at a time. Before MDEV-16264 this was trivially satisfied by the fact that there was a dedicated purge coordinator thread.
Attachments
Issue Links
- is caused by
-
MDEV-11026 Make InnoDB number of IO write/read threads dynamic
-
- Closed
-