Bug #95993 | after kill -9 of mysqld, than cannot to start, crash | ||
---|---|---|---|
Submitted: | 26 Jun 2019 7:43 | Modified: | 30 Aug 2019 10:05 |
Reporter: | wang huai | Email Updates: | |
Status: | Duplicate | Impact on me: | |
Category: | MySQL Server | Severity: | S2 (Serious) |
Version: | 8.0.16 | OS: | Any |
Assigned to: | CPU Architecture: | Any | |
Tags: | crash |
[26 Jun 2019 7:43]
wang huai
[26 Jun 2019 12:46]
MySQL Verification Team
Hi Mr. Huai, Thank you for your bug report. This bug turned out to be reported internally already and was already fixed. It is yet unknown whether it is fixed in 8.0.17 or 8.0.18. We shall let you know when we find out. Also, we do not know when any of those two releases will be published. Thank you for your contribution.
[27 Jun 2019 12:48]
MySQL Verification Team
Hi Mr. Huai, We have made enquiries and discovered that a bug is fixed in 8.0.17. We do not know when will 8.0.17 be fixed, but when that release is published, all our pages on dev.mysql.com will be updated.
[27 Aug 2019 14:13]
Bernardo Perez
Hello, Could it be possible that you confirm if this bug is related to this one https://bugs.mysql.com/bug.php?id=93768 ? Best regards, We have a signature now that seems to be a bit of a mix of Bug #95993 and Bug #93768 stack: 2019-08-27T05:32:24.378216Z 26139 [ERROR] [MY-013183] [InnoDB] Assertion failure: row0umod.cc:149:!dummy_big_rec thread 47668352853760 InnoDB: We intentionally generate a memory trap. InnoDB: Submit a detailed bug report to http://bugs.mysql.com. 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/8.0/en/forcing-innodb-recovery.html InnoDB: about forcing recovery. 05:32:24 UTC - 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. Attempting to collect some information that could help diagnose the problem. As this is a crash and something is definitely wrong, the information collection process might fail. key_buffer_size=16777216 read_buffer_size=262144 max_used_connections=24 max_threads=312 thread_count=8 connection_count=7 It is possible that mysqld could use up to key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 179579 K bytes of memory Hope that's ok; if not, decrease some variables in the equation. Thread pointer: 0x2b5b195c3000 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 = 2b5aa73d1d00 thread_stack 0x40000 /rdsdbbin/mysql/bin/mysqld(my_print_stacktrace(unsigned char*, unsigned long)+0x2e) [0x1c891ee] /rdsdbbin/mysql/bin/mysqld(handle_fatal_signal+0x413) [0xecd013] /lib64/libpthread.so.0(+0xf100) [0x2b5a78826100] /lib64/libc.so.6(gsignal+0x37) [0x2b5a795a25f7] /lib64/libc.so.6(abort+0x148) [0x2b5a795a3ce8] /rdsdbbin/mysql/bin/mysqld(ut_dbg_assertion_failed(char const*, char const*, unsigned long)+0x2c1) [0x1fa5f61] /rdsdbbin/mysql/bin/mysqld() [0x2134f44] /rdsdbbin/mysql/bin/mysqld() [0x2135526] /rdsdbbin/mysql/bin/mysqld(row_undo_mod(undo_node_t*, que_thr_t*)+0xbdf) [0x2138aaf] /rdsdbbin/mysql/bin/mysqld(row_undo_step(que_thr_t*)+0x4f) [0x1f26f3f] /rdsdbbin/mysql/bin/mysqld(que_run_threads(que_thr_t*)+0x94f) [0x1ec6b1f] /rdsdbbin/mysql/bin/mysqld() [0x1f83edb] /rdsdbbin/mysql/bin/mysqld(trx_rollback_to_savepoint(trx_t*, trx_savept_t*)+0x22) [0x1f84162] /rdsdbbin/mysql/bin/mysqld(trx_rollback_last_sql_stat_for_mysql(trx_t*)+0x31) [0x1f841a1] /rdsdbbin/mysql/bin/mysqld() [0x1df2a00] /rdsdbbin/mysql/bin/mysqld(ha_rollback_low(THD*, bool)+0xe7) [0xfbbf97] /rdsdbbin/mysql/bin/mysqld(MYSQL_BIN_LOG::rollback(THD*, bool)+0xe4) [0x19db3e4] /rdsdbbin/mysql/bin/mysqld(ha_rollback_trans(THD*, bool)+0x11c) [0xfbc1cc] /rdsdbbin/mysql/bin/mysqld(trans_rollback_stmt(THD*)+0x29) [0xe9cf19] /rdsdbbin/mysql/bin/mysqld(mysql_execute_command(THD*, bool)+0x2c7d) [0xdc1a5d] /rdsdbbin/mysql/bin/mysqld(mysql_parse(THD*, Parser_state*, bool)+0x33f) [0xdc370f] /rdsdbbin/mysql/bin/mysqld(dispatch_command(THD*, COM_DATA const*, enum_server_command)+0x2c1b) [0xdc67ab] /rdsdbbin/mysql/bin/mysqld(do_command(THD*)+0x179) [0xdc7289] /rdsdbbin/mysql/bin/mysqld() [0xec00c8] /rdsdbbin/mysql/bin/mysqld() [0x21617f7] /lib64/libpthread.so.0(+0x7dc5) [0x2b5a7881edc5] /lib64/libc.so.6(clone+0x6d) [0x2b5a79663c9d]
[27 Aug 2019 15:52]
MySQL Verification Team
Mr. huai, Those truly look identical !!!!
[29 Aug 2019 1:49]
wang huai
I think these two bugs are not same, because Bug #95993 was happend in mysqld main thread, and current_thd of main thread is null, so it is crash, but another one is not the main thread, it is the user thread. I have show the method of fixed(where thd main thread's current_thd is null), so I think they are not same.
[29 Aug 2019 12:25]
MySQL Verification Team
Hi Mr. huai, Let us know if you can repeat the crash with 8.0.17 .... Then, we can take a look at it ....
[30 Aug 2019 10:05]
wang huai
when I fixed it, the mysqld working well, so the crash not happend in the future, but as I known, we can infer and repeat it from the crash running stack, and it is confirmed that, in the start phrase, the current_thd is null, so if used it(binlog_recover), it will certain crash.
[30 Aug 2019 12:40]
MySQL Verification Team
Hi Mr. huai, First of all, I did not understand much of what you wrote. Second, this is a user thread, not main thread. Main thread only spawns user threads and does not execute any statements. Third, when `current_thd` is NULL it is mostly the case when it is optimised away, id est, it is found in one of the registers. Last, but not least, analysis of the original bug clearly shows that this one is a duplicate.
[30 Aug 2019 12:42]
MySQL Verification Team
One more comment ..... A developer who fixed the original bug in 8.0.17 is very adamant that this one is a duplicate. Hence, try 8.0.17.
[30 Aug 2019 13:39]
Bernardo Perez
Hello Sinisa, Bernardo Perez here. I made the comment of the previous stack and the other bug that I thought It could be related. Thanks for confirming that they are duplicates and that my signature is the same as this bug. Regards,
[30 Aug 2019 13:42]
MySQL Verification Team
Mr. Perez, You are welcome .......