I've back ported to-3140 to 5.2.x, which seems to fix this.
> On Mar 24, 2015, at 3:05 PM, Bryan Call <bc...@apache.org> wrote: > > I opened: https://issues.apache.org/jira/browse/TS-3464 > <https://issues.apache.org/jira/browse/TS-3464> > > Alan, > Good find! > > > -Bryan > > >> On Mar 24, 2015, at 1:26 PM, Leif Hedstrom <zw...@apache.org> wrote: >> >> >>> On Mar 24, 2015, at 2:14 PM, Bryan Call <bc...@apache.org >>> <mailto:bc...@apache.org>> wrote: >>> >>> I am able to reproduce the issue Alan is seeing... >>> >>> (gdb) bt full >>> #0 0x00007ffff4ef28c7 in raise () from /lib64/libc.so.6 >>> No symbol table info available. >>> #1 0x00007ffff4ef452a in abort () from /lib64/libc.so.6 >>> No symbol table info available. >>> #2 0x00007ffff7bb79fb in ink_die_die_die (retval=1) at ink_error.cc:43 >>> No locals. >>> #3 0x00007ffff7bb7aca in ink_fatal_va(int, const char *, typedef >>> __va_list_tag __va_list_tag *) (return_code=1, >>> fmt=0x7ffff7bc8398 "%s:%d: failed assert `%s`", ap=0x7ffff1f67148) >>> at ink_error.cc:67 >>> msg = "FATAL: HttpSM.cc <http://httpsm.cc/>:2499: failed assert >>> `server_entry == NULL`", '\000' <repeats 964 times> >>> len = 7 >>> #4 0x00007ffff7bb7b6d in ink_fatal (return_code=1, >>> message_format=0x7ffff7bc8398 "%s:%d: failed assert `%s`") >>> at ink_error.cc:75 >>> ap = {{gp_offset = 40, fp_offset = 48, >>> overflow_arg_area = 0x7ffff1f67220, >>> reg_save_area = 0x7ffff1f67160}} >>> #5 0x00007ffff7bb5590 in _ink_assert ( >>> expression=0x7a719b "server_entry == NULL", file=0x7a6edc "HttpSM.cc >>> <http://httpsm.cc/>", >>> line=2499) at ink_assert.cc:37 >> >> >> Ok, sounds like we have a regression. Did anyone track down the offending >> commit / Jira? Or open a new Jira? >> >> — Leif >
