[MDEV-13872] innodb.innodb-64k-crash failed in buildbot with error on startup and SIGSEGV Created: 2017-09-22  Updated: 2023-11-29  Resolved: 2023-11-29

Status: Closed
Project: MariaDB Server
Component/s: Storage Engine - InnoDB, Tests
Affects Version/s: 10.1
Fix Version/s: 10.2.2

Type: Bug Priority: Major
Reporter: Elena Stepanova Assignee: Marko Mäkelä
Resolution: Cannot Reproduce Votes: 0
Labels: None

Issue Links:
Relates
relates to MDEV-12699 Improve crash recovery of corrupted d... Closed

 Description   

http://buildbot.askmonty.org/buildbot/builders/kvm-bintar-quantal-x86/builds/8526/steps/test/logs/stdio

innodb.innodb-64k-crash 'innodb_plugin'  w2 [ fail ]
        Test ended at 2017-09-06 16:37:11
 
CURRENT_TEST: innodb.innodb-64k-crash
 
 
Server [mysqld.1 - pid: 15550, winpid: 15550, exit: 256] failed during test run
Server log from this test:
----------SERVER LOG START-----------
2017-09-06 16:36:57 3058517760 [Note] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld (mysqld 10.1.27-MariaDB) starting as process 15551 ...
2017-09-06 16:36:57 3058517760 [Warning] Could not increase number of max_open_files to more than 1024 (request: 4162)
2017-09-06 16:36:57 3058517760 [Note] Plugin 'partition' is disabled.
2017-09-06 16:36:57 3058517760 [Note] Plugin 'SEQUENCE' is disabled.
2017-09-06 16:36:57 3058517760 [Note] InnoDB:  innodb-page-size has been changed from the default value 16384 to 65536.
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Using mutexes to ref count buffer pool pages
2017-09-06 16:36:57 3058517760 [Note] InnoDB: The InnoDB memory heap is disabled
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
2017-09-06 16:36:57 3058517760 [Note] InnoDB: GCC builtin __atomic_thread_fence() is used for memory barrier
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Compressed tables use zlib 1.2.3
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Using Linux native AIO
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Using generic crc32 instructions
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Initializing buffer pool, size = 32.0M
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Completed initialization of buffer pool
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Highest supported file format is Barracuda.
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Starting crash recovery from checkpoint LSN=12640612
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Restoring possible half-written data pages from the doublewrite buffer...
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Starting a batch to recover 6 pages from redo log
2017-09-06 16:36:57 3058517760 [Note] InnoDB: Starting a batch to recover 1 pages from redo log
2017-09-06 16:36:58 3058517760 [Note] InnoDB: Starting a batch to recover 2 pages from redo log
2017-09-06 16:36:58 3058517760 [Note] InnoDB: Starting a batch to recover 16 pages from redo log
2017-09-06 16:36:59 3058517760 [Note] InnoDB: Starting a batch to recover 223 pages from redo log
2017-09-06 16:36:59 3058517760 [Note] InnoDB: Starting a batch to recover 166 pages from redo log
2017-09-06 16:37:00 3058517760 [Note] InnoDB: Starting a batch to recover 174 pages from redo log
2017-09-06 16:37:01 3058517760 [Note] InnoDB: Starting a batch to recover 178 pages from redo log
2017-09-06 16:37:01 3058517760 [Note] InnoDB: Starting a batch to recover 242 pages from redo log
2017-09-06 16:37:02 3058517760 [Note] InnoDB: Starting a batch to recover 115 pages from redo log
2017-09-06 16:37:02 3058517760 [Note] InnoDB: Starting a batch to recover 196 pages from redo log
2017-09-06 16:37:03 3058517760 [Note] InnoDB: Starting a batch to recover 66 pages from redo log
2017-09-06 16:37:03 3058517760 [Note] InnoDB: Starting a batch to recover 40 pages from redo log
2017-09-06 16:37:04 3058517760 [Note] InnoDB: Starting a batch to recover 41 pages from redo log
2017-09-06 16:37:04 3058517760 [Note] InnoDB: Starting a batch to recover 42 pages from redo log
2017-09-06 16:37:05 3058517760 [Note] InnoDB: Starting a batch to recover 41 pages from redo log
2017-09-06 16:37:05 3058517760 [Note] InnoDB: Starting a batch to recover 56 pages from redo log
2017-09-06 16:37:06 3058517760 [Note] InnoDB: Starting a batch to recover 60 pages from redo log
2017-09-06 16:37:06 3058517760 [Note] InnoDB: Starting a batch to recover 64 pages from redo log
2017-09-06 16:37:07 3058517760 [Note] InnoDB: Starting a batch to recover 55 pages from redo log
2017-09-06 16:37:08 3058517760 [Note] InnoDB: Starting a batch to recover 47 pages from redo log
2017-09-06 16:37:08 3058517760 [Note] InnoDB: Starting a batch to recover 152 pages from redo log
2017-09-06 16:37:09 3058517760 [Note] InnoDB: Starting a batch to recover 155 pages from redo log
InnoDB: Set innodb_force_recovery to ignore this error.
2017-09-06 16:37:09 3058517760 [ERROR] Plugin 'InnoDB' init function returned error.
2017-09-06 16:37:09 3058517760 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_DATAFILES' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_TABLESTATS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_CMP' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'XTRADB_READ_VIEW' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'XTRADB_RSEG' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_DELETED' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_LOCK_WAITS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'CHANGED_PAGE_BITMAPS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_TABLESPACES_ENCRYPTION' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_FIELDS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'XTRADB_INTERNAL_HASH_TABLES' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_TABLESPACES_SCRUBBING' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_CMPMEM_RESET' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_INDEX_TABLE' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_BEING_DELETED' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_MUTEXES' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_CONFIG' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'FEEDBACK' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_INDEX_CACHE' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_FT_DEFAULT_STOPWORD' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_CMP_RESET' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_COLUMNS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_CMP_PER_INDEX_RESET' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_TABLESPACES' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_SEMAPHORE_WAITS' is disabled.
2017-09-06 16:37:09 3058517760 [Note] Plugin 'INNODB_SYS_INDEXES' is disabled.
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown variable 'loose-feedback-debug-startup-interval=20'
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown variable 'loose-feedback-debug-first-interval=60'
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown variable 'loose-feedback-debug-interval=60'
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown option '--loose-pam-debug'
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown option '--loose-skip-plugin-innodb-changed-pages'
2017-09-06 16:37:09 3058517760 [Warning] /usr/local/mariadb-10.1.27-linux-i686/bin/mysqld: unknown variable 'loose-debug-sync-timeout=300'
2017-09-06 16:37:09 3058517760 [ERROR] Unknown/unsupported storage engine: InnoDB
2017-09-06 16:37:09 3058517760 [ERROR] Aborting
 
170906 16:37:09 [ERROR] mysqld got signal 11 ;
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: 10.1.27-MariaDB
read_buffer_size=131072
max_used_connections=0
thread_count=0
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...
----------SERVER LOG END-------------
mysqltest failed but provided no output
The result from queries just before the failure was:
< snip >
UPDATE t2 SET ec=@i,fc=@i,gc=@i,hc=@i,ic=@i,jc=@i;
UPDATE t2 SET kc=@i,lc=@i,mc=@i,nc=@i,oc=@i,pc=@i;
UPDATE t2 SET qc=@i,rc=@i,sc=@i,tc=@i,uc=@i;
UPDATE t2 SET vc=@i,wc=@i,xc=@i,yc=@i,zc=@i;
COMMIT;
BEGIN;
INSERT INTO t1 SELECT * from t2;
UPDATE t1 SET a=@c,b=@c,c=@c,d=@c,e=@c,f=@c,g=@c,h=@c,i=@c,j=@c,
k=@c,l=@c,m=@c,n=@c,o=@c,p=@c,q=@c,r=@c,s=@c,t=@c,u=@c,
v=@c,w=@c,x=@c,y=@c,z=@c,
aa=@c,ba=@c,ca=@c,da=@c,ea=@c,fa=@c,ga=@c,ha=@c,ia=@c,ja=@c,
ka=@c,la=@c,ma=@c,na=@c,oa=@c,pa=@c,qa=@c,ra=@c,sa=@c,ta=@c,ua=@c,
va=@c,wa=@c,xa=@c,ya=@c,za=@c,
ab=@c,bb=@c,cb=@c,db=@c,eb=@c,fb=@c,gb=@c,hb=@c,ib=@c,jb=@c,
kb=@c,lb=@c,mb=@c,nb=@c,ob=@c,pb=@c,qb=@c,rb=@c,sb=@c,tb=@c,ub=@c,
vb=@c,wb=@c,xb=@c,yb=@c,zb=@c,
ac=@c,bc=@c,cc=@c,dc=@c,ec=@c,fc=@c,gc=@c,hc=@c,ic=@c,jc=@c,
kc=@c,lc=@c,mc=@c,nc=@c,oc=@c,pc=@c,qc=@c,rc=@c,sc=@c,tc=@c,uc=@c,
vc=@c,wc=@c,xc=@c,yc=@c,zc=@c;
# Kill and restart



 Comments   
Comment by Marko Mäkelä [ 2023-11-29 ]

The following is the latest signal 11 in a main branch for this test:

10.1 ecd3a7e00de28d279ef180e1d4defa97

innodb.innodb-64k-crash 'innodb_plugin'  w3 [ fail ]
        Test ended at 2018-12-07 14:27:37
2018-12-07 14:27:36 70366838784000 [ERROR] Unknown/unsupported storage engine: InnoDB
2018-12-07 14:27:36 70366838784000 [ERROR] Aborting
 
181207 14:27:37 [ERROR] mysqld got signal 11 ;
/usr/sbin/mysqld(my_print_stacktrace+0x64)[0x2824ef44]
mysys/stacktrace.c:267(my_print_stacktrace)[0x27c39410]
addr2line: '': No such file
[0x3fff8f750478]
[0x3fff86e5f3d8]
[0x3fff86e3f0ac]
pthread_create.c:0(start_thread)[0x3fff8f708728]
/lib64/libc.so.6(clone+0x98)[0x3fff8ee9d210]

I remember seeing this type of shutdown crashes after failed InnoDB startup, most of the time in the test innodb.doublewrite in MariaDB Server 10.1 and possibly earlier versions. In MariaDB Server 10.2, I refactored the InnoDB startup code so that all errors would be properly propagated to the plugin initialization function.

Generated at Thu Feb 08 08:09:01 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.