Crash on startup with Couchbase Server 2.1.1

Hi,

In a Couchbase 2.1.1 (community edition) deployment, the beam.smp process cores periodically on startup. The backtrace from the core file is:

(gdb) bt
#0 0x0000003888e32925 in raise () from /lib64/libc.so.6
#1 0x0000003888e34105 in abort () from /lib64/libc.so.6
#2 0x000000000044350a in erl_exit (n=1, fmt=0x58b788 “%s: Cannot %s %lu bytes of memory (of type “%s”).\n”) at beam/erl_init.c:1613
#3 0x0000000000428e2c in erts_alc_fatal_error (error=1, func=0, n=130) at beam/erl_alloc.c:1642
#4 0x0000000000428ee7 in erts_alloc_n_enomem (n=6, size=18446744073709551615) at beam/erl_alloc.c:1667
#5 0x00000000004ec28b in erts_alloc (p=0x7f971000e478, new_sz=75025, objv=0x7f977f28acc0, nobj=1) at beam/erl_alloc.h:205
#6 do_minor (p=0x7f971000e478, new_sz=75025, objv=0x7f977f28acc0, nobj=1) at beam/erl_gc.c:897
#7 0x00000000004f0115 in minor_collection (p=0x7f971000e478, objv=0x7f977f28acc0, nobj=1, need=0) at beam/erl_gc.c:769
#8 erts_garbage_collect (p=0x7f971000e478, objv=0x7f977f28acc0, nobj=1, need=0) at beam/erl_gc.c:376
#9 0x00000000004f0619 in erts_gc_after_bif_call (p=0x7f971000e478, result=140288306407762, regs=, arity=)
at beam/erl_gc.c:326
#10 0x00000000005223e8 in process_main () at beam/beam_emu.c:2545
#11 0x000000000048db06 in sched_thread_func (vesdp=0x7f97806f4040) at beam/erl_process.c:3782
#12 0x0000000000586369 in thr_wrapper (vtwd=0x7fff1b16ff30) at pthread/ethread.c:106
#13 0x00000038892079d1 in start_thread () from /lib64/libpthread.so.0
#14 0x0000003888ee8b6d in clone () from /lib64/libc.so.6

The backtrace looks like the backtrace mentioned in MB-7562 (https://issues.couchbase.com/browse/MB-7562) which mentions that this issue has been fixed in version 2.0.1. So, my questions are:

  • What is the cause of the issue?
  • Has this been fixed in the 2.1.1 Community Edition version?
  • Is it fixed in a later community edition version?

Thanks

Not sure what fix are on which branch for the community edition, maybe @cihangirb can help you.