Description:
Oh this will be fun to repeat :)
Backtrace of mysqld:
(gdb) where
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7f0da98 in pthread_kill () from /lib/tls/i686/cmov/libpthread.so.0
#2 0x083a3015 in write_core (sig=6) at stacktrace.c:220
#3 0x0823c945 in handle_segfault (sig=6) at mysqld.cc:2134
#4 <signal handler called>
#5 0xffffe410 in __kernel_vsyscall ()
#6 0xb7cba9a1 in raise () from /lib/tls/i686/cmov/libc.so.6
#7 0xb7cbc2b9 in abort () from /lib/tls/i686/cmov/libc.so.6
#8 0xb7cb3f51 in __assert_fail () from /lib/tls/i686/cmov/libc.so.6
#9 0x085ee02b in NdbEventOperationImpl::readBlobParts (this=0x8dd1a78,
buf=0x8dd5650 'X' <repeats 200 times>..., blob=0x8dd5418, part=0, count=1)
at NdbEventOperationImpl.cpp:489
#10 0x0861b4e2 in NdbBlob::readEventParts (this=0x8dd5418, buf=0x8dd5650 'X' <repeats 200 times>...,
part=0, count=1) at NdbBlob.cpp:1116
#11 0x0861b634 in NdbBlob::readParts (this=0x8dd5418, buf=0x8dd5650 'X' <repeats 200 times>..., part=0,
count=1) at NdbBlob.cpp:1085
#12 0x0861d2a9 in NdbBlob::readDataPrivate (this=0x8dd5418,
buf=0x8d55658 '\ufffd' <repeats 120 times>, "h4z\025I", bytes=@0xb7bdc0a4) at NdbBlob.cpp:926
#13 0x0861d4a5 in NdbBlob::readData (this=0x8dd5418, data=0x8d55558, bytes=@0xb7bdc0a4) at NdbBlob.cpp:857
#14 0x083fd69f in get_ndb_blobs_value (table=0x8dca970, value_array=0x8dca3a0, buffer=@0xb7bdc154,
buffer_size=@0xb7bdc14c, ptrdiff=0) at ha_ndbcluster.cc:874
#15 0x0842589a in ndb_binlog_thread_handle_data_event (ndb=0x8a499d0, pOp=0x8dbfca0, row=@0xb7bdc2bc,
trans=@0xb7bdc35c) at ha_ndbcluster_binlog.cc:2844
#16 0x084285b1 in ndb_binlog_thread_func (arg=0x0) at ha_ndbcluster_binlog.cc:3438
#17 0xb7f0a341 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#18 0xb7d5b4ee in clone () from /lib/tls/i686/cmov/libc.so.6
nparts was 0 and count equaled 1
How to repeat:
Was ndb_restore a DBT2 database into a fresh cluster.
started getting "increase MaxNoConcurrentTransactions", so did a rolling upgrade (with restore running) upping MaxNoOfConcurrentTransactions and Operations.
I have the backup I was restoring.
of course, stupid me (by accident) deleted the fs for this crash.
Suggested fix:
don't crash. perhaps change the assert to a return an error?
Description: Oh this will be fun to repeat :) Backtrace of mysqld: (gdb) where #0 0xffffe410 in __kernel_vsyscall () #1 0xb7f0da98 in pthread_kill () from /lib/tls/i686/cmov/libpthread.so.0 #2 0x083a3015 in write_core (sig=6) at stacktrace.c:220 #3 0x0823c945 in handle_segfault (sig=6) at mysqld.cc:2134 #4 <signal handler called> #5 0xffffe410 in __kernel_vsyscall () #6 0xb7cba9a1 in raise () from /lib/tls/i686/cmov/libc.so.6 #7 0xb7cbc2b9 in abort () from /lib/tls/i686/cmov/libc.so.6 #8 0xb7cb3f51 in __assert_fail () from /lib/tls/i686/cmov/libc.so.6 #9 0x085ee02b in NdbEventOperationImpl::readBlobParts (this=0x8dd1a78, buf=0x8dd5650 'X' <repeats 200 times>..., blob=0x8dd5418, part=0, count=1) at NdbEventOperationImpl.cpp:489 #10 0x0861b4e2 in NdbBlob::readEventParts (this=0x8dd5418, buf=0x8dd5650 'X' <repeats 200 times>..., part=0, count=1) at NdbBlob.cpp:1116 #11 0x0861b634 in NdbBlob::readParts (this=0x8dd5418, buf=0x8dd5650 'X' <repeats 200 times>..., part=0, count=1) at NdbBlob.cpp:1085 #12 0x0861d2a9 in NdbBlob::readDataPrivate (this=0x8dd5418, buf=0x8d55658 '\ufffd' <repeats 120 times>, "h4z\025I", bytes=@0xb7bdc0a4) at NdbBlob.cpp:926 #13 0x0861d4a5 in NdbBlob::readData (this=0x8dd5418, data=0x8d55558, bytes=@0xb7bdc0a4) at NdbBlob.cpp:857 #14 0x083fd69f in get_ndb_blobs_value (table=0x8dca970, value_array=0x8dca3a0, buffer=@0xb7bdc154, buffer_size=@0xb7bdc14c, ptrdiff=0) at ha_ndbcluster.cc:874 #15 0x0842589a in ndb_binlog_thread_handle_data_event (ndb=0x8a499d0, pOp=0x8dbfca0, row=@0xb7bdc2bc, trans=@0xb7bdc35c) at ha_ndbcluster_binlog.cc:2844 #16 0x084285b1 in ndb_binlog_thread_func (arg=0x0) at ha_ndbcluster_binlog.cc:3438 #17 0xb7f0a341 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0 #18 0xb7d5b4ee in clone () from /lib/tls/i686/cmov/libc.so.6 nparts was 0 and count equaled 1 How to repeat: Was ndb_restore a DBT2 database into a fresh cluster. started getting "increase MaxNoConcurrentTransactions", so did a rolling upgrade (with restore running) upping MaxNoOfConcurrentTransactions and Operations. I have the backup I was restoring. of course, stupid me (by accident) deleted the fs for this crash. Suggested fix: don't crash. perhaps change the assert to a return an error?