You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
STD::CONDITION_VARIABLE::WAIT_FOR IN SOLARIS/X86
When a MySQL Server built with ndbcluster engine is run in a
Solaris/x86 platform, it crashes everytime when
NDB_SCHEMA_OBJECT::client_wait_completed() method is called. The
segfault happens when this method calls the
std::condition_variable::wait_for() method. This particular crash
happens only in the release builds. The reason seems to be an issue with
the compiler used in Solaris. When the optimization level -xO2 is used,
it fills the wait_for() call with a bogus code that triggers this
segfault during runtime.
Fixed the issue by making the release builds use optimization
level -xO1 for ndbcluster instead.
Change-Id: I8471970169cd971958ac2f5aa6dcb0ccd8d083a3
0 commit comments