[MDEV-19261] InnoDB: Failing assertion: dfield->type.mtype == 0 || dfield->len == len upon crash upgrade from 10.0.14 Created: 2019-04-16  Updated: 2019-11-21

Status: Open
Project: MariaDB Server
Component/s: Storage Engine - InnoDB
Affects Version/s: 10.1
Fix Version/s: 10.1

Type: Bug Priority: Minor
Reporter: Elena Stepanova Assignee: Marko Mäkelä
Resolution: Unresolved Votes: 0
Labels: None


 Description   

10.1 409dddf6

2019-04-16 18:29:23 7fdc447fc700  InnoDB: Assertion failure in thread 140584018757376 in file trx0rec.cc line 1133
InnoDB: Failing assertion: dfield->type.mtype == 0 || dfield->len == len
InnoDB: We intentionally generate a memory trap.
 
#5  0x00007fdc7078c3fa in abort () from /lib/x86_64-linux-gnu/libc.so.6
#6  0x000056549f963cab in trx_undo_rec_get_partial_row (ptr=0x7fdc42c208c7 "\a\003\200าณ\b\004", index=0x7fdc42c59378, update=0x7fdc42d13b68, row=0x7fdc6fe1eaf0, ignore_prefix=0, heap=0x7fdc6fc4b980) at /data/src/10.1/storage/xtradb/trx/trx0rec.cc:1132
#7  0x000056549f90dc24 in row_purge_parse_undo_rec (node=0x7fdc6fe1ea78, undo_rec=0x7fdc42c20278 "\tW\214", updated_extern=0x7fdc447fbd0e, thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/row/row0purge.cc:828
#8  0x000056549f90df49 in row_purge (node=0x7fdc6fe1ea78, undo_rec=0x7fdc42c20278 "\tW\214", thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/row/row0purge.cc:913
#9  0x000056549f90e17f in row_purge_step (thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/row/row0purge.cc:994
#10 0x000056549f8abe3c in que_thr_step (thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/que/que0que.cc:1089
#11 0x000056549f8ac040 in que_run_threads_low (thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/que/que0que.cc:1151
#12 0x000056549f8ac1cc in que_run_threads (thr=0x7fdc5df6eec0) at /data/src/10.1/storage/xtradb/que/que0que.cc:1192
#13 0x000056549f95c015 in trx_purge (n_purge_threads=1, batch_size=300, truncate=false) at /data/src/10.1/storage/xtradb/trx/trx0purge.cc:1230
#14 0x000056549f940513 in srv_do_purge (n_threads=1, n_total_purged=0x7fdc447fbea0) at /data/src/10.1/storage/xtradb/srv/srv0srv.cc:3448
#15 0x000056549f940ab6 in srv_purge_coordinator_thread (arg=0x0) at /data/src/10.1/storage/xtradb/srv/srv0srv.cc:3597
#16 0x00007fdc728a5494 in start_thread (arg=0x7fdc447fc700) at pthread_create.c:333
#17 0x00007fdc7084093f in clone () from /lib/x86_64-linux-gnu/libc.so.6

To reproduce,

Non-debug build says

2019-04-16 18:38:09 139796648805952 [Note] InnoDB: Restoring possible half-written data pages from the doublewrite buffer...
InnoDB: 1 transaction(s) which must be rolled back or cleaned up
InnoDB: in total 4 row operations to undo
InnoDB: Trx id counter is 59392

and doesn't crash.

I'm setting it to Minor for now, because crash-upgrade from 10.0 is only supported in 10.1, and it's not the high priority version for corner cases (which it probably is).



 Comments   
Comment by Elena Stepanova [ 2019-11-21 ]

Same assertion was also observed on MySQL 5.6 => 10.1 crash upgrade.

Generated at Thu Feb 08 08:50:16 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.