MariaDB keepts crashing on start

Yesterday, we had a sudden crash on MariaDB 10.0.38

Afterwards it crashed on each restart.

Afterwards, I was only able to start it with InnoDB Recovery Mode 3, so I needed to dump the database and restore it to an empty data directory (which solved the situation).

Is there any way, to find out the root cause?

We only had the same issue some weeks ago with another server and MariaDB 10.3.17 (not sure this has been the same root cause).

First Crash on 10.0.38

2019-11-10 13:30:26 7fd52b3fa700 InnoDB: Assertion failure in thread 140553530353408 in file fsp0fsp.cc line 1521
InnoDB: Failing assertion: frag_n_used > 0
InnoDB: We intentionally generate a memory trap.
InnoDB: Submit a detailed bug report to https://jira.mariadb.org/
InnoDB: If you get repeated assertion failures or crashes, even
InnoDB: immediately after the mysqld startup, there may be
InnoDB: corruption in the InnoDB tablespace. Please refer to
InnoDB: http://dev.mysql.com/doc/refman/5.6/en/forcing-innodb-recovery.html
InnoDB: about forcing recovery.
191110 13:30:26 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.

To report this bug, see https://mariadb.com/kb/en/reporting-bugs

We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.

Server version: 10.0.38-MariaDB-0+deb8u1
key_buffer_size=33554432
read_buffer_size=131072
max_used_connections=110
max_threads=752
thread_count=5
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 1684034 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x0
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0x0 thread_stack 0x48000
/usr/sbin/mysqld(my_print_stacktrace+0x2e)[0xc0885e]
/usr/sbin/mysqld(handle_fatal_signal+0x3af)[0x738a2f]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xf890)[0x7fdad02e0890]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0x37)[0x7fdacee89067]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x148)[0x7fdacee8a448]
/usr/sbin/mysqld[0xb0fa3c]
/usr/sbin/mysqld[0xb17ea9]
/usr/sbin/mysqld[0xa74b97]
/usr/sbin/mysqld[0xa7998d]
/usr/sbin/mysqld[0xa70ed7]
/usr/sbin/mysqld[0xa71b99]
/usr/sbin/mysqld[0xa721dd]
/usr/sbin/mysqld[0xb3b289]
/usr/sbin/mysqld[0xb2696d]
/usr/sbin/mysqld[0xb333b2]
/usr/sbin/mysqld[0xb335ae]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x8064)[0x7fdad02d9064]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d)[0x7fdacef3c62d]
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.

Following Crashes

191110 13:30:35 mysqld_safe Number of processes running now: 0
191110 13:30:35 mysqld_safe mysqld restarted
191110 13:30:35 [Warning] 'THREAD_CONCURRENCY' is deprecated and will be removed in a future release.
191110 13:30:35 [Warning] options --log-slow-admin-statements, --log-queries-not-using-indexes and --log-slow-slave-statements have no effect if --log_slow_queries is not set
191110 13:30:35 [Note] /usr/sbin/mysqld (mysqld 10.0.38-MariaDB-0+deb8u1) starting as process 4365 ...
2019-11-10 13:30:35 7f1f0f6c5780 InnoDB: Warning: Using innodb_additional_mem_pool_size is DEPRECATED. This option may be removed in future releases, together with the option innodb_use_sys_malloc and with the I
nnoDB's internal memory allocator.
191110 13:30:35 [Note] InnoDB: Using mutexes to ref count buffer pool pages
191110 13:30:35 [Note] InnoDB: The InnoDB memory heap is disabled
191110 13:30:35 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
191110 13:30:35 [Note] InnoDB: GCC builtin __atomic_thread_fence() is used for memory barrier
191110 13:30:35 [Note] InnoDB: Compressed tables use zlib 1.2.8
191110 13:30:35 [Note] InnoDB: Using Linux native AIO
191110 13:30:35 [Note] InnoDB: Using CPU crc32 instructions
191110 13:30:35 [Note] InnoDB: Initializing buffer pool, size = 20.0G
191110 13:30:36 [Note] InnoDB: Completed initialization of buffer pool
191110 13:30:36 [Note] InnoDB: Highest supported file format is Barracuda.
191110 13:30:36 [Note] InnoDB: Starting crash recovery from checkpoint LSN=12502111848897
191110 13:30:38 [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 0 row operations to undo
InnoDB: Trx id counter is 951470848
191110 13:30:38 [Note] InnoDB: Starting final batch to recover 6970 pages from redo log
InnoDB: Last MySQL binlog file position 0 417264732, file name /var/lib/mysql/share/log_bin.001456
InnoDB: Cleaning up trx with id 951470008
2019-11-10 13:30:41 7f1f0f6c5780 InnoDB: Assertion failure in thread 139771379472256 in file fsp0fsp.cc line 1521
InnoDB: Failing assertion: frag_n_used > 0
InnoDB: We intentionally generate a memory trap.
InnoDB: Submit a detailed bug report to https://jira.mariadb.org/
InnoDB: If you get repeated assertion failures or crashes, even
InnoDB: immediately after the mysqld startup, there may be
InnoDB: corruption in the InnoDB tablespace. Please refer to
InnoDB: http://dev.mysql.com/doc/refman/5.6/en/forcing-innodb-recovery.html
InnoDB: about forcing recovery.
191110 13:30:41 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.

To report this bug, see https://mariadb.com/kb/en/reporting-bugs

We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.

Server version: 10.0.38-MariaDB-0+deb8u1
key_buffer_size=33554432
read_buffer_size=131072
max_used_connections=0
max_threads=752
thread_count=0
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 1684034 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x0
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0x0 thread_stack 0x48000
/usr/sbin/mysqld(my_print_stacktrace+0x2e)[0xc0885e]
/usr/sbin/mysqld(handle_fatal_signal+0x3af)[0x738a2f]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xf890)[0x7f1f0f2a4890]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0x37)[0x7f1f0de4d067]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x148)[0x7f1f0de4e448]
/usr/sbin/mysqld[0xb0fa3c]
/usr/sbin/mysqld[0xb17ea9]
/usr/sbin/mysqld[0xa74b97]
/usr/sbin/mysqld[0xa7998d]
/usr/sbin/mysqld[0xa71bcf]
/usr/sbin/mysqld[0xa6a737]
/usr/sbin/mysqld[0x9dbe4e]
/usr/sbin/mysqld[0xa577ea]
/usr/sbin/mysqld[0x989665]
/usr/sbin/mysqld(_Z24ha_initialize_handlertonP13st_plugin_int+0x5e)[0x73aa4e]
/usr/sbin/mysqld[0x5d1a05]
/usr/sbin/mysqld(_Z11plugin_initPiPPci+0x4b0)[0x5d2be0]
/usr/sbin/mysqld[0x52ac45]
/usr/sbin/mysqld(_Z11mysqld_mainiPPc+0x57b)[0x5307db]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f1f0de39b45]
/usr/sbin/mysqld[0x526221]
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.

First Crash on 10.3.17

191022 11:59:06 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.

To report this bug, see https://mariadb.com/kb/en/reporting-bugs

We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.

Server version: 10.3.18-MariaDB-1:10.3.18+maria~bionic
key_buffer_size=33554432
read_buffer_size=131072
max_used_connections=58
max_threads=102
thread_count=22
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 256994 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x7fa2cc000c08
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0x7fa3038a5c48 thread_stack 0x49000
/usr/sbin/mysqld(my_print_stacktrace+0x2e)[0x55a62e7cba7e]
/usr/sbin/mysqld(handle_fatal_signal+0x515)[0x55a62e262285]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x12890)[0x7fa42c468890]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0xc7)[0x7fa42b97ce97]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x141)[0x7fa42b97e801]
/usr/sbin/mysqld(+0xa2f0a1)[0x55a62e5370a1]
/usr/sbin/mysqld(+0x95976f)[0x55a62e46176f]
/usr/sbin/mysqld(+0x95c156)[0x55a62e464156]
/usr/sbin/mysqld(+0xa3db88)[0x55a62e545b88]
/usr/sbin/mysqld(+0xa4befd)[0x55a62e553efd]
/usr/sbin/mysqld(+0xa4e608)[0x55a62e556608]
/usr/sbin/mysqld(+0x9b555f)[0x55a62e4bd55f]
/usr/sbin/mysqld(+0x9b7629)[0x55a62e4bf629]
/usr/sbin/mysqld(+0x9b849b)[0x55a62e4c049b]
/usr/sbin/mysqld(+0x9729d0)[0x55a62e47a9d0]
/usr/sbin/mysqld(+0xa028e7)[0x55a62e50a8e7]
/usr/sbin/mysqld(+0x9e4540)[0x55a62e4ec540]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x76db)[0x7fa42c45d6db]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x3f)[0x7fa42ba5f88f]
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort.
Query (0x0):
Connection ID (thread ID): 1
Status: NOT_KILLED

Optimizer switch: index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_mer
ge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort
_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=off,exists_to_in=on,ord
erby_uses_equalities=on,condition_pushdown_for_derived=on,split_materialized=on

The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
information that should help you find out what is causing the crash.
Writing a core file...
Working directory at /var/lib/mysql/share
Resource Limits:
Limit Soft Limit Hard Limit Units
Max cpu time unlimited unlimited seconds
Max file size unlimited unlimited bytes
Max data size unlimited unlimited bytes
Max stack size 8388608 unlimited bytes
Max core file size 0 unlimited bytes
Max resident set unlimited unlimited bytes
Max processes 62987 62987 processes
Max open files 16364 16364 files
Max locked memory 16777216 16777216 bytes
Max address space unlimited unlimited bytes
Max file locks unlimited unlimited locks
Max pending signals 768028 768028 signals
Max msgqueue size 819200 819200 bytes
Max nice priority 0 0
Max realtime priority 0 0
Max realtime timeout unlimited unlimited us
Core pattern: core

Following Crash

2019-10-22 11:59:24 0 [Note] InnoDB: Using Linux native AIO
2019-10-22 11:59:24 0 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
2019-10-22 11:59:24 0 [Note] InnoDB: Uses event mutexes
2019-10-22 11:59:24 0 [Note] InnoDB: Compressed tables use zlib 1.2.11
2019-10-22 11:59:24 0 [Note] InnoDB: Number of pools: 1
2019-10-22 11:59:24 0 [Note] InnoDB: Using SSE2 crc32 instructions
2019-10-22 11:59:24 0 [Note] InnoDB: Initializing buffer pool, total size = 4G, instances = 8, chunk size = 128M
2019-10-22 11:59:24 0 [Note] InnoDB: Completed initialization of buffer pool
2019-10-22 11:59:24 0 [Note] InnoDB: If the mysqld execution user is authorized, page cleaner thread priority can be changed. See the man page of setpriority().
2019-10-22 11:59:24 0 [Note] InnoDB: Starting crash recovery from checkpoint LSN=478136761706
2019-10-22 11:59:24 0 [Note] InnoDB: Starting final batch to recover 832 pages from redo log.
2019-10-22 11:59:33 0 [Note] InnoDB: 128 out of 128 rollback segments are active.
2019-10-22 11:59:33 0 [Note] InnoDB: Removed temporary tablespace data file: "ibtmp1"
2019-10-22 11:59:33 0 [Note] InnoDB: Creating shared tablespace for temporary tables
2019-10-22 11:59:33 0 [Note] InnoDB: Setting file './ibtmp1' size to 12 MB. Physically writing the file full; Please wait ...
2019-10-22 11:59:33 0 [Note] InnoDB: File './ibtmp1' size is now 12 MB.
2019-10-22 11:59:33 0 [Note] InnoDB: Waiting for purge to start
2019-10-22 11:59:33 0 [Note] InnoDB: 10.3.18 started; log sequence number 478137116730; transaction id 151762556
2019-10-22 11:59:33 0 [Note] Plugin 'FEEDBACK' is disabled.
2019-10-22 11:59:33 0 [Note] InnoDB: Loading buffer pool(s) from /var/lib/mysql/share/ib_buffer_pool
2019-10-22 11:59:33 0 [Note] Recovering after a crash using tc.log
2019-10-22 11:59:33 0 [Note] Starting crash recovery...
2019-10-22 11:59:33 0 [Note] Crash recovery finished.
2019-10-22 11:59:33 0 [Note] Server socket created on IP: '::'.
2019-10-22 11:59:33 1 [ERROR] [FATAL] InnoDB: Rec offset 99, cur1 offset 15289, cur2 offset 16131
191022 11:59:33 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.

To report this bug, see https://mariadb.com/kb/en/reporting-bugs

We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.

Server version: 10.3.18-MariaDB-1:10.3.18+maria~bionic
key_buffer_size=33554432
read_buffer_size=131072
max_used_connections=0
max_threads=102
thread_count=6
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 256994 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x7f1714000c08
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0x7f1755118c48 thread_stack 0x49000

• • • buffer overflow detected ***: /usr/sbin/mysqld terminated