Bug #96599 Test innodb_undo.i_s_files_4k hangs on PPC64LE in RELEASE mode
Submitted: 21 Aug 2019 8:35 Modified: 22 Oct 2019 13:09
Reporter: Tony Reix Email Updates:
Status: Not a Bug Impact on me:
None 
Category:MySQL Server Severity:S3 (Non-critical)
Version:8.0.17 OS:Fedora (fc30)
Assigned to: CPU Architecture:POWER (PPC64LE)

[21 Aug 2019 8:35] Tony Reix
Description:
Building MySQL v8.0.16 on Fedora30 on PPC64LE in DEBUG mode, I've faced about 5 hangs with tests: innodb.blob-crash-4k (available in debug mode only), innodb_undo.i_s_files_4k and opt_trace.bugs_no_prot_all after several tries, first with all tests then with only these 3 tests.
It was quite easy to reproduce the hang with innodb.blob-crash-4k . With this test, the hang is random but appears about 1 times out of 2-3 tries.

I was using the official community-mysql-8.0.16-2.fc31.src.rpm file provided by the Fedora project for building MySQL, simply adding: --trace -DCMAKE_BUILD_TYPE=DEBUG  to the cmake command line in the community-mysql.spec file.
I used a modified version of the %check command in order to run the tests. See "How to repeat" for details.

Moving back to the original community-mysql.spec file, thus compiling in RELEASE mode, it took about 9 hours of running alternatively and infinitely the tests innodb_undo.i_s_files_4k and opt_trace.bugs_no_prot_all before the hang happened again, apparently in the same context than in DEBUG mode and with test innodb_undo.i_s_files_4k.

$ ps -edf | grep mysqld
mysql    42486 42485 20 04:37 pts/3    01:11:46 /var/lib/mysql/rpmbuild/BUILD/mysql-8.0.16/build/runtime_output_directory/mysqld --no-defaults --initialize-insecure --loose-skip-ndbcluster --tmpdir=/tmp/mysql-notroot/tmp/ --core-file --datadir=/tmp/mysql-notroot/mysqld.1/data --secure-file-priv=/tmp/mysql-notroot --innodb_buffer_pool_size=24M --innodb-log-file-size=5M --innodb_autoextend_increment=8 --character-sets-dir=/var/lib/mysql/rpmbuild/BUILD/mysql-8.0.16/share/charsets --loose-auto_generate_certs=OFF --loose-sha256_password_auto_generate_rsa_keys=OFF --loose-caching_sha2_password_auto_generate_rsa_keys=OFF --innodb-page-size=4k --init-file=/tmp/mysql-notroot/tmp/bootstrap.sql

My PPC64LE Fedora VM has 32 POWER8 logical processors at 3425MHz.
4 virtual processors  X  SMT8 .

HMC says:
Processing Units:
  Minimum:      0.10
  Sharing mode:	Uncapped
  Assigned:	2.00
  Weight:	128
  Maximum:	4.00
	
Virtual Processors:
  Minimum:	1.0
  Assigned:	4.0
  Maximum:	8.0

$ ppc64_cpu --smt
SMT=8

$ ppc64_cpu --info
Core   0:    0*    1*    2*    3*    4*    5*    6*    7* 
Core   1:    8*    9*   10*   11*   12*   13*   14*   15* 
Core   2:   16*   17*   18*   19*   20*   21*   22*   23* 
Core   3:   24*   25*   26*   27*   28*   29*   30*   31* 

$ ppc64_cpu --cores-present
Number of cores present = 4

On Fedora30 x86_64, either in DEBUG or RELEASE mode, I was unable to reproduce the hang after about 14 hours of testing the same 2 tests as on PPC64LE.

I'll provide stack trace from gdb then.

How to repeat:
su - mysql

Build:
  rpmbuild -ba community-mysql.spec

Test:
while true
do
  sh community-mysql-8.0.16-2.fc31.spec.TEST innodb_undo.i_s_files_4k ; sh community-mysql-8.0.16-2.fc31.spec.TEST opt_trace.bugs_no_prot_all
done

with (adapt to your own PATH): community-mysql-8.0.16-2.fc31.spec.TEST :
cd /var/lib/mysql/rpmbuild//BUILD/mysql-8.0.16
cd build
cd mysql-test
export MTR_BUILD_THREAD=64
./mtr \
  --verbose \
  --parallel=1 \
  --force \
  --retry=0 \
  --mysqld=--binlog-format=mixed \
  --suite-timeout=2000 \
  --testcase-timeout=60 \
  --report-unstable-tests \
  --max-save-core=100 \
  --max-test-fail=0 \
  --vardir=/tmp/mysql-notroot \
  $@
[21 Aug 2019 8:40] Tony Reix
Details about traces & logs, plus gdb information

Attachment: 96599 (application/octet-stream, text), 81.18 KiB.

[21 Aug 2019 8:54] Tony Reix
xz-compressed file providing the trace of building MySQL in RELEASE mode

Attachment: mysql-bug-data-96599-community-mysql-8.0.16-2.fc31.spec.res.6-RELEASE-NotRoot.xz (application/x-xz, text), 373.59 KiB.

[22 Aug 2019 6:53] Tony Reix
The machine where the hang occurred is still up and the hang is waiting for more commands to be run under gdb if needed. I'll probably have to release this hang soon, probably next week.
So, do not hesitate to asking me for more information and data by running gdb commands.
About version 8.0.17 of MySQL, I'm waiting for the Fedora project to provide a .src.rpm file.
[16 Sep 2019 15:05] Tony Reix
I've been able to reproduce the hang with version 8.0.17 in RELEASE mode with test innodb_undo.i_s_files_4k .

I've used a "while true" infinite loop around a call of mtr which leads to:
 /root/rpmbuild/BUILD/mysql-8.0.17/mysql-test/mysql-test-run.pl  --warnings --verbose --parallel=1 --force --retry=0 --suite-timeout=2000 --testcase-timeout=60 --report-unstable-tests --max-save-core=100 --max-test-fail=0 --tmpdir=/My --user=root --mark-progress opt_trace.bugs_no_prot_all innodb_undo.i_s_files_4k

I see that this test innodb_undo.i_s_files_4k has been "pass" several times before I got a hang. The code is blocked for more than 4 hours.

# ps -edf | grep mysqld
root      4786  4681  0 11:44 pts/1    00:00:00 /root/rpmbuild/BUILD/mysql-8.0.17/build/runtime_output_directory/mysqltest_safe_process --verbose -- /root/rpmbuild/BUILD/mysql-8.0.17/build/runtime_output_directory/mysqld --no-defaults --initialize-insecure --loose-skip-ndbcluster --tmpdir=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/tmp/ --core-file --datadir=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/mysqld.1/data --secure-file-priv=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var --innodb_buffer_pool_size=24M --innodb-log-file-size=5M --innodb_autoextend_increment=8 --character-sets-dir=/root/rpmbuild/BUILD/mysql-8.0.17/share/charsets --loose-auto_generate_certs=OFF --loose-sha256_password_auto_generate_rsa_keys=OFF --loose-caching_sha2_password_auto_generate_rsa_keys=OFF --innodb-page-size=4k --init-file=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/tmp/bootstrap.sql

root      4787  4786  0 11:44 pts/1    00:00:10 /root/rpmbuild/BUILD/mysql-8.0.17/build/runtime_output_directory/mysqld --no-defaults --initialize-insecure --loose-skip-ndbcluster --tmpdir=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/tmp/ --core-file --datadir=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/mysqld.1/data --secure-file-priv=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var --innodb_buffer_pool_size=24M --innodb-log-file-size=5M --innodb_autoextend_increment=8 --character-sets-dir=/root/rpmbuild/BUILD/mysql-8.0.17/share/charsets --loose-auto_generate_certs=OFF --loose-sha256_password_auto_generate_rsa_keys=OFF --loose-caching_sha2_password_auto_generate_rsa_keys=OFF --innodb-page-size=4k --init-file=/root/rpmbuild/BUILD/mysql-8.0.17/build/mysql-test/var/tmp/bootstrap.sql

# date
lun. sept. 16 15:32:13 CEST 2019

So, it was launched at 11:44 and is still blocked at 15:32 (and now at 16:52).

However, though mysqld was built with "-O2 -g" and is not stripped, I have no information by means of gdb about what the thread is doing:

(gdb) where
#0  0x000020000008a114 in ?? ()
#1  0x000020000008a0f4 in ?? ()
#2  0x0000200000089e30 in ?? ()
#3  0x000000012c1a6964 in ?? ()
#4  0x000000012ae7b7fc in ?? ()
#5  0x000000012be40fb4 in ?? ()
#6  0x000000012be3b8d0 in ?? ()
#7  0x000000012aa475f4 in ?? ()
#8  0x000000012aa4df74 in ?? ()
#9  0x000000012a9faba0 in ?? ()
#10 0x00002000011a4d28 in ?? ()
#11 0x00002000011a4f24 in ?? ()
#12 0x0000000000000000 in ?? ()

(gdb) i thread
  Id   Target Id         Frame 
* 1    LWP 4787 "mysqld" 0x000020000008a114 in ?? ()
  2    LWP 4788 "mysqld" 0x000020000009144c in ?? ()
  3    LWP 4790 "mysqld" 0x00002000012a7c74 in ?? ()
  4    LWP 4791 "mysqld" 0x00002000012a7c74 in ?? ()
  5    LWP 4792 "mysqld" 0x00002000012a7c74 in ?? ()
  6    LWP 4793 "mysqld" 0x00002000012a7c74 in ?? ()
  7    LWP 4794 "mysqld" 0x00002000012a7c74 in ?? ()
  8    LWP 4795 "mysqld" 0x00002000012a7c74 in ?? ()
  9    LWP 4796 "mysqld" 0x00002000012a7c74 in ?? ()
  10   LWP 4797 "mysqld" 0x00002000012a7c74 in ?? ()
  11   LWP 4798 "mysqld" 0x00002000012a7c74 in ?? ()
  12   LWP 4799 "mysqld" 0x00002000012a7c74 in ?? ()
  13   LWP 4800 "mysqld" 0x000020000009144c in ?? ()
  14   LWP 4807 "mysqld" 0x0000200000091950 in ?? ()
  15   LWP 4808 "mysqld" 0x0000200000091950 in ?? ()
  16   LWP 4809 "mysqld" 0x0000200000091950 in ?? ()
  17   LWP 4810 "mysqld" 0x0000200000091950 in ?? ()
  18   LWP 4811 "mysqld" 0x0000200000091950 in ?? ()
  19   LWP 4812 "mysqld" 0x0000200000091950 in ?? ()

# file ./build/runtime_output_directory/mysqld
./build/runtime_output_directory/mysqld: ELF 64-bit LSB shared object, 64-bit PowerPC or cisco 7500, version 1 (SYSV), dynamically linked, interpreter /lib64/ld64.so.2, for GNU/Linux 3.10.0, BuildID[sha1]=8009252cb72ada8755314af03f2fbed9a8f491a0, with debug_info, not stripped, too many notes (256)
[16 Sep 2019 15:05] Tony Reix
Using strace, I see (more details in future attached file):

# strace -s 99 -ffp 4787
strace: Process 4787 attached with 19 threads
[pid  4812] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4811] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4810] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4809] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4808] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4807] restart_syscall(<... resuming interrupted futex ...> <unfinished ...>
[pid  4800] futex(0x200004046e30, FUTEX_WAIT_PRIVATE, 0, NULL <unfinished ...>
[pid  4799] io_getevents(0x200003fa0000, 1, 256,  <unfinished ...>
[pid  4798] io_getevents(0x200003f90000, 1, 256,  <unfinished ...>
[pid  4797] io_getevents(0x200003f80000, 1, 256,  <unfinished ...>
[pid  4796] io_getevents(0x200003f70000, 1, 256,  <unfinished ...>
[pid  4795] io_getevents(0x200003ef0000, 1, 256,  <unfinished ...>
[pid  4794] io_getevents(0x200003ee0000, 1, 256,  <unfinished ...>
[pid  4793] io_getevents(0x200003ed0000, 1, 256,  <unfinished ...>
[pid  4792] io_getevents(0x200003ec0000, 1, 256,  <unfinished ...>
[pid  4791] io_getevents(0x200003f10000, 1, 256,  <unfinished ...>
[pid  4790] io_getevents(0x200003f00000, 1, 256,  <unfinished ...>
[pid  4788] futex(0x20000400b954, FUTEX_WAIT_PRIVATE, 0, NULL <unfinished ...>
[pid  4787] futex(0x200003a5f130, FUTEX_WAIT, 4788, NULL <unfinished ...>
[pid  4807] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4807] futex(0x200004212720, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4807] futex(0x200004212774, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=9999762} <unfinished ...>
[pid  4808] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4808] futex(0x2000042127c0, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4808] futex(0x200004212810, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=99999740} <unfinished ...>
[pid  4807] <... futex resumed>)        = -1 ETIMEDOUT (Connection timed out)
[pid  4807] futex(0x200004212720, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4807] futex(0x200004212774, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=9999723} <unfinished ...>
[pid  4809] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4809] futex(0x200004212900, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4809] futex(0x200004212950, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=99999797} <unfinished ...>
[pid  4812] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4810] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4812] futex(0x200004216710, FUTEX_WAKE_PRIVATE, 1 <unfinished ...>
[pid  4810] futex(0x2000042167b0, FUTEX_WAKE_PRIVATE, 1 <unfinished ...>
[pid  4812] <... futex resumed>)        = 0
[pid  4810] <... futex resumed>)        = 0
[pid  4812] futex(0x200004216764, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=99999739} <unfinished ...>
[pid  4810] futex(0x200004216804, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=99999758} <unfinished ...>
[pid  4811] <... restart_syscall resumed>) = -1 ETIMEDOUT (Connection timed out)
[pid  4811] futex(0x200004212860, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4811] futex(0x2000042128b4, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=99999723} <unfinished ...>
[pid  4807] <... futex resumed>)        = -1 ETIMEDOUT (Connection timed out)
[pid  4807] futex(0x200004212720, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4807] futex(0x200004212774, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=9999785}) = -1 ETIMEDOUT (Connection timed out)
[pid  4807] futex(0x200004212720, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4807] futex(0x200004212774, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=9999736}) = -1 ETIMEDOUT (Connection timed out)
[pid  4807] futex(0x200004212720, FUTEX_WAKE_PRIVATE, 1) = 0
[pid  4807] futex(0x200004212774, FUTEX_WAIT_PRIVATE, 0, {tv_sec=0, tv_nsec=9999735}) = -1 ETIMEDOUT (Connection timed out)

Using :
 for i in /proc/4787/task/*/stack; do echo $i >> /tmp/Stacks; cat $i >> /tmp/Stacks ; done
I have:

/proc/4787/task/4787/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4788/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4790/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4791/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4792/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4793/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4794/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4795/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4796/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4797/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4798/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4799/stack
[<0>] __switch_to+0x340/0x520
[<0>] read_events+0x144/0x1d0
[<0>] do_io_getevents+0xa4/0x190
[<0>] sys_io_getevents+0x6c/0xd0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4800/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4807/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4808/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4809/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4810/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4811/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
/proc/4787/task/4812/stack
[<0>] __switch_to+0x340/0x520
[<0>] futex_wait_queue_me+0x108/0x1e0
[<0>] futex_wait+0x180/0x320
[<0>] do_futex+0x1d4/0x1230
[<0>] sys_futex+0x118/0x1e0
[<0>] system_call+0x5c/0x70
[16 Sep 2019 15:06] Tony Reix
output of strace

Attachment: 96599.strace.4787 (application/octet-stream, text), 155.03 KiB.

[16 Sep 2019 15:47] Tony Reix
My mistake: I wa
[16 Sep 2019 15:52] Tony Reix
My mistake: I was using the DEBUG version of mysqld 8.0.17, not the RELEASE version. Thus, using the correct mysqld executable, gdb does work and shows:

(gdb) where
#0  0x000020000008a114 in __GI___pthread_timedjoin_ex (threadid=35184433295456, thread_return=0x0, abstime=<optimized out>, block=<optimized out>) at pthread_join_common.c:84
#1  0x0000200000089e30 in __pthread_join (threadid=<optimized out>, thread_return=<optimized out>) at pthread_join.c:24
#2  0x000000012c1a6964 in my_thread_join (thread=<optimized out>, value_ptr=<optimized out>) at /root/rpmbuild/BUILD/mysql-8.0.17/mysys/my_thread.cc:100
#3  0x000000012ae7b7fc in bootstrap::run_bootstrap_thread (file=0x0, boot_handler=<optimized out>, thread_type=<optimized out>)
    at /root/rpmbuild/BUILD/mysql-8.0.17/sql/bootstrap.cc:436
#4  0x000000012be40fb4 in dd::Dictionary_impl::init (dd_init=<optimized out>) at /root/rpmbuild/BUILD/mysql-8.0.17/sql/dd/impl/dictionary_impl.cc:146
#5  0x000000012be3b8d0 in dd::init (dd_init=<optimized out>) at /root/rpmbuild/BUILD/mysql-8.0.17/sql/dd/impl/dd.cc:65
#6  0x000000012aa475f4 in init_server_components () at /root/rpmbuild/BUILD/mysql-8.0.17/sql/mysqld.cc:5484
#7  0x000000012aa4df74 in mysqld_main (argc=<optimized out>, argv=<optimized out>) at /root/rpmbuild/BUILD/mysql-8.0.17/sql/mysqld.cc:6482
#8  0x000000012a9faba0 in main (argc=<optimized out>, argv=<optimized out>) at /root/rpmbuild/BUILD/mysql-8.0.17/sql/main.cc:25

(gdb) i thread
  Id   Target Id                                 Frame 
* 1    Thread 0x200000049d50 (LWP 4787) "mysqld" 0x000020000008a114 in __GI___pthread_timedjoin_ex (threadid=35184433295456, thread_return=0x0, abstime=<optimized out>, 
    block=<optimized out>) at pthread_join_common.c:84
  2    Thread 0x200003a5f060 (LWP 4788) "mysqld" 0x000020000009144c in futex_wait_cancelable (private=0, expected=0, futex_word=0x20000400b954)
    at ../sysdeps/unix/sysv/linux/futex-internal.h:88
  3    Thread 0x20000840f060 (LWP 4790) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  4    Thread 0x20000a75f060 (LWP 4791) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  5    Thread 0x20000ab6f060 (LWP 4792) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  6    Thread 0x20000af7f060 (LWP 4793) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  7    Thread 0x20000b38f060 (LWP 4794) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  8    Thread 0x20000b79f060 (LWP 4795) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  9    Thread 0x20000bbaf060 (LWP 4796) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  10   Thread 0x20000bfbf060 (LWP 4797) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  11   Thread 0x20002c40f060 (LWP 4798) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  12   Thread 0x20002c81f060 (LWP 4799) "mysqld" syscall () at ../sysdeps/unix/sysv/linux/powerpc/syscall.S:29
  13   Thread 0x20002cc2f060 (LWP 4800) "mysqld" 0x000020000009144c in futex_wait_cancelable (private=0, expected=0, futex_word=0x200004046e30)
    at ../sysdeps/unix/sysv/linux/futex-internal.h:88
  14   Thread 0x20002f86f060 (LWP 4807) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20002f86e0c8, expected=0, 
    futex_word=0x200004212774) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
  15   Thread 0x20003c40f060 (LWP 4808) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20003c40e138, expected=0, 
    futex_word=0x200004212810) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
  16   Thread 0x20003cc2f060 (LWP 4809) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20003cc2e0c8, expected=0, 
    futex_word=0x200004212950) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
  17   Thread 0x20002fc7f060 (LWP 4810) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20002fc7e0e8, expected=0, 
    futex_word=0x200004216804) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
  18   Thread 0x20003c81f060 (LWP 4811) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20003c81e0c8, expected=0, 
    futex_word=0x2000042128b4) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
  19   Thread 0x20003d03f060 (LWP 4812) "mysqld" 0x0000200000091950 in futex_reltimed_wait_cancelable (private=<optimized out>, reltime=0x20003d03de68, expected=0, 
    futex_word=0x200004216764) at ../sysdeps/unix/sysv/linux/futex-internal.h:142
[16 Sep 2019 15:53] Tony Reix
gdb data about several threads

Attachment: 96599.gdb.4787 (application/octet-stream, text), 16.04 KiB.

[22 Oct 2019 13:09] MySQL Verification Team
Hello Mr. Reix,

Thank you for your bug report.

However, it is not a bug.

Supported CPUs are listed here:

https://www.mysql.com/support/supportedplatforms/database.html