Possibly crasher in HostDB serialization (disk sync) #1322
Description
Hi Thomas,
we're seeing this odd looking crash, fairly often:
#0 0x00002aaaad74337d in __libc_waitpid (pid=<optimized out>, stat_loc=<optimized out>, options=<optimized out>)
at ../sysdeps/unix/sysv/linux/waitpid.c:41
#1 0x00002aaaaab83297 in crash_logger_invoke (signo=11, info=0x2aaaf20007f0, ctx=0x2aaaf20006c0) at Crash.cc:164
#2 <signal handler called>
#3 0x0000000000000000 in ?? ()
#4 0x00002aaaaaf06270 in Continuation::handleEvent (data=0x2ae608f1b140, event=1, this=<optimized out>) at I_Continuation.h:153
#5 EThread::process_event (this=0x2aaaf1c0da00, e=0x2ae608f1b140, calling_code=1) at UnixEThread.cc:143
#6 0x00002aaaaaf0727f in EThread::execute (this=0x2aaaf1c0da00) at UnixEThread.cc:197
#7 0x00002aaaaaf05ba6 in spawn_thread_internal (a=0x2aaac51fccc0) at Thread.cc:84
#8 0x00002aaaad73baa1 in start_thread (arg=0x2aaaf2001700) at pthread_create.c:301
#9 0x00002aaaae5f293d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115
Not particularly helpful, I know, but in frame #4, Randal found that the data seems related to the serialization continuation:
(gdb) frame 4
#4 0x00002aaaaaf06270 in Continuation::handleEvent (data=0x2ae608f1b140, event=1, this=<optimized out>) at I_Continuation.h:153
153 I_Continuation.h: No such file or directory.
(gdb) print *((Event*)data).continuation
$1 = {<force_VFPT_to_top> = {_vptr.force_VFPT_to_top = 0x2aaaab1d9670 <vtable for RefCountCacheSerializer<RefCountCache<HostDBInfo> >+16>},
handler = (int (Continuation::*)(Continuation * const, int,
void *)) 0x2aaaaadcf7a0 <RefCountCacheSerializer<RefCountCache<HostDBInfo> >::initialize_storage(int, Event*)>, mutex = {
m_ptr = 0x2aaaf0e46e40}, link = {<SLink<Continuation>> = {next = 0x0}, prev = 0x0}, control_flags = {raw_flags = 0}}
We're thinking of turning off HostDB sync again to see if that helps, but let me know if you have any brilliant ideas as to why this would fail here (unfortunately the above is all we got :-/).