Skip to main content

Summary This article clarifies the investigation of a crash of the Artix container (it_container).
Article Number 32665
Environment Artix 5.5 Solaris 10
Question/Problem Description Artix container (it_container) crashing and generating core dump.



Clarifying Information
Error Message pstack output of the crashing thread: ff1cd1e4 _lwp_kill (6, 0, ff235090, ff1ac928, ffffffff, 6) 8 ff141bac abort (15dd08, 1, f9b35480, ee930, ff2334d8, 0) 110 f9b2b0fc void os::abort(bool) (1, f9e15fc4, 1, f9dfc000, 19fc4, 19c00) 5c f9c4d6f0 void VMError::report_and_die() (f9e3a814, f9e2d7b8, 1, f9e40508, f9e15f9c, 0) cd8 f95c0df4 JVM_handle_solaris_signal (b, fb87fb78, fb87f8c0, b1400, 0, a074) a8c ff1c961c __sighndlr (b, fb87fb78, fb87f8c0, f95c0330, 0, 1) c ff1bdce8 call_user_handler (b, ffbffeff, c, 0, fbbc1200, fb87f8c0) 3b8 ff1bded0 sigacthandler (b, fb87fb78, fb87f8c0, 0, 0, 0) 60 --- called from signal handler with signal 11 (SIGSEGV) --- febd098c memcpy (0, 23299ea0, 0, 4000, ff2303d8, 0) 584 fad4d52c void IT_ATLI2_HTTP::HTTPBufferReader::make_body_contigious(unsigned long,IT_ATLI2_HTTP::HTTPInputStream*) (20b79280, 0, 230dc558, 114f0ff, 0, 0) 230 fad2ee00 bool IT_ATLI2_HTTP::HTTPInputStream::has_enough_data() (230dc558, fe0ca464, 20b79280, e6, 114f019, fad893c4) 2f8 fad32590 bool IT_ATLI2_HTTP::HTTPInputStream::process_input(const TimeBase::UtcT&) (230dc558, fb87fea0, 495409c2, 0, fad893c4, 230dc5e0) 558 fad2e31c IT_WorkQueue::WorkItemStatus IT_ATLI2_HTTP::HTTPInputStream::execute() (230dc558, fe11eba4, fad8cb94, 3990c, 0, ffffffff) 24 fc8d48c8 IT_WorkQueue::WorkItemStatus IT_ATLI2_IP::IPPoolImpl::execute() (2848b20, 1, fb87ff2c, 1, fe11eba4, 0) 3b4 fdda5ae0 void*IT_Work_WorkerThread::run() (1383e0, 2, c1710, c1710, c1710, fe11eba4) 100 ff1c94f0 _lwp_start (0, 0, 0, 0, 0, 0)
Defect/Enhancement Number
Cause pmap output indicates total 4063568K of memory allocated in the JVM with:

00030000 589696K rwx-- [ heap ] // that's the native heap
2E400000 1228800K rw--- // these three are the Java heap
79400000 1916928K rw--- //
EE400000 131072K rw--- //


pstack output indicates that the crash is caused by
febd098c memcpy (0, 23299ea0, 0, 4000, ff2303d8, 0) 584

The crash occurs due to exhaustion of the _native_ heap, most likely due to the Java heap that's too large.
Resolution Ensure the application is being profiled, the memory consumption analyzed, and the corresponding code is investigated for any memory leaks.
Workaround
Notes Should the Artix container crash, ensure you collect pstack/pmap analysis files and provide these to Micro Focus Technical Support.
Attachment

Created date: 19 June 2012
Last Modified: 13 February 2013
Last Published: 19 June 2012
First Published date: 19 June 2012

#Orbix
#KnowledgeDocs