[KUDU-2325] tserver crashes in kudu::StackTraceCollector::RevokeSigData

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 1.7.0
• Fix Version/s: 1.7.0
• Component/s: tserver, util
• Code Review: http://gerrit.cloudera.org:8080/9430
• Assignee: Todd Lipcon
• Reporter: Alexey Serbin
• Created: 23/Feb/18 05:01
• Updated: 28/Feb/18 07:39
• Resolved: 28/Feb/18 07:39

Description

While running raft_consensus_stress-itest, I noticed tserver sometimes crashes with the stack trace like below:

    @     0x7f188e6c92fd  google::LogMessage::Fail() at ??:0
    @     0x7f188e6cb1bd  google::LogMessage::SendToLog() at ??:0
    @     0x7f188e6c8e39  google::LogMessage::Flush() at ??:0
    @     0x7f188e6cbc5f  google::LogMessageFatal::~LogMessageFatal() at ??:0
    @     0x7f189028e5af  kudu::StackTraceCollector::RevokeSigData() at ??:0
    @     0x7f189028ecba  kudu::StackTraceCollector::AwaitCollection() at ??:0
    @     0x7f189028ee86  kudu::GetThreadStack() at ??:0
    @     0x7f189028efb7  kudu::DumpThreadStack() at ??:0
    @     0x7f189031cc49  kudu::KernelStackWatchdog::RunThread() at ??:0
    @     0x7f1890320af9  boost::_mfi::mf0<>::operator()() at ??:0
    @     0x7f1890320a60  boost::_bi::list1<>::operator()<>() at ??:0
    @     0x7f1890320a0a  boost::_bi::bind_t<>::operator()() at ??:0
    @     0x7f18903207ed  boost::detail::function::void_function_obj_invoker0<>::invoke() at ??:0
    @     0x7f1891eb2568  boost::function0<>::operator()() at ??:0
    @     0x7f18903aeaad  kudu::Thread::SuperviseThread() at ??:0
    @     0x7f189418d184  start_thread at ??:0
    @     0x7f188c6b7ffd  clone at ??:0
    @              (nil)  (unknown)