Replicator Crash
I have replicator working well enough…to crash:
Jan 28 11:26:12 la dovecot[1187]: replicator: Panic: data stack: Out of memory when allocating 268435496 bytes Jan 28 11:26:12 la dovecot[1187]: replicator: Error: Raw backtrace: #0 t_askpass[0x7fa59db0c0] -> #1 backtrace_append[0x7fa59db374] -> #2 backtra
ce_get[0x7fa59db510] -> #3 execvp_const[0x7fa59e7ba4] -> #4 i_syslog_fatal_handler[0x7fa59e8510] -> #5 i_panic[0x7fa5936808] -> #6 t_pop_pass_str
[0x7fa59e0eb4] -> #7 connection_deinit[0x7fa59e6430] -> #8 pool_datastack_create[0x7fa5a0e7b0] -> #9 array_bsearch_insert_pos_i[0x7fa59d46d0] ->
#10 t_base64_scheme_decode[0x7fa59dcb24] -> #11 buffer_append[0x7fa59dd3d0] -> #12 replicator_queue_push[0x556e8f4f40] -> #13 replicator_brain_in
it[0x556e8f53e4] -> #14 _start[0x556e8f3d70] -> #15 _start[0x556e8f3f04] -> #16 io_loop_call_io[0x7fa5a02f20] -> #17 io_loop_handler_run_internal
[0x7fa5a05450] -> #18 io_loop_handler_run[0x7fa5a05680] -> #19 io_loop_run[0x7fa5a05944] -> #20 master_service_run[0x7fa59684b0] -> #21 main[0x55
6e8f3930] -> #22 __libc_init_first[0x7fa57676d0] -> #23 __libc_start_main[0x7fa5767780] -> #24 _start[0x556e8f3b80] Jan 28 11:26:12 la dovecot[1187]: replicator: Fatal: master: service(replicator): child 1185 killed with signal 6 (core dumps disabled - https://
dovecot.org/bugreport.html#coredumps)
This is on Debian 12.
# dovecot --version 2.3.19.1 (9b53102964)
There are gigabytes of free RAM, though it says it is trying to allocate 268,435,496 bytes--on the stack--and that does seem wrong in any case. The Maildir being synced is over 430MB, is that too much for dsync??
Suggestions?
Thanks,
-kb
On 28/01/2025 21:38 EET Kent Borg via dovecot <dovecot@dovecot.org>
wrote:
I have replicator working well enough…to crash:
Jan 28 11:26:12 la dovecot[1187]: replicator: Panic: data
stack: Out
of memory when allocating 268435496 bytes
Jan 28 11:26:12 la dovecot[1187]: replicator: Error: Raw
backtrace: #0
t_askpass[0x7fa59db0c0] -> #1 backtrace_append
[0x7fa59db374] -> #2
backtra
ce_get[0x7fa59db510] -> #3 execvp_const[0x7fa59e7ba4] -> #4
i_syslog_fatal_handler[0x7fa59e8510] -> #5 i_panic
[0x7fa5936808] -> #6
t_pop_pass_str
[0x7fa59e0eb4] -> #7 connection_deinit[0x7fa59e6430] -> #8
pool_datastack_create[0x7fa5a0e7b0] -> #9
array_bsearch_insert_pos_i[0x7fa59d46d0] ->
#10 t_base64_scheme_decode[0x7fa59dcb24] -> #11
buffer_append[0x7fa59dd3d0] -> #12 replicator_queue_push
[0x556e8f4f40]
-> #13 replicator_brain_in
it[0x556e8f53e4] -> #14 _start[0x556e8f3d70] -> #15
_start[0x556e8f3f04] -> #16 io_loop_call_io[0x7fa5a02f20] -
> #17
io_loop_handler_run_internal
[0x7fa5a05450] -> #18 io_loop_handler_run[0x7fa5a05680] -
> #19
io_loop_run[0x7fa5a05944] -> #20 master_service_run
[0x7fa59684b0] ->
#21 main[0x55
6e8f3930] -> #22 __libc_init_first[0x7fa57676d0] -> #23
__libc_start_main[0x7fa5767780] -> #24 _start[0x556e8f3b80]
Jan 28 11:26:12 la dovecot[1187]: replicator: Fatal:
master:
service(replicator): child 1185 killed with signal 6 (core
dumps
disabled - https://
dovecot.org/bugreport.html#coredumps)
This is on Debian 12.
# dovecot --version
2.3.19.1 (9b53102964)
There are gigabytes of free RAM, though it says it is trying to
allocate
268,435,496 bytes--on the stack--and that does seem wrong in any
case.
The Maildir being synced is over 430MB, is that too much for dsync??
Suggestions?
Thanks,
-kb
_______________________________________________
dovecot mailing list -- dovecot@dovecot.org
To unsubscribe send an email to dovecot-leave@dovecot.org
It's maybe https://github.com/dovecot/core/commit/
d3a097075afb7b5d8955978ed4a2ae99add11f88.patch
Aki
28.01.2025 23:53, Aki Tuomi via dovecot пишет:
On 28/01/2025 21:38 EET Kent Borg via dovecot <dovecot@dovecot.org> wrote: I have replicator working well enough…to crash: Jan 28 11:26:12 la dovecot[1187]: replicator: Panic: data stack: Out of memory when allocating 268435496 bytes Jan 28 11:26:12 la dovecot[1187]: replicator: Error: Raw backtrace: #0 t_askpass[0x7fa59db0c0] -> #1 backtrace_append [0x7fa59db374] -> #2 backtra\ ce_get[0x7fa59db510] -> #3 execvp_const[0x7fa59e7ba4] -> #4 i_syslog_fatal_handler[0x7fa59e8510] -> #5 i_panic [0x7fa5936808] -> #6 t_pop_pass_str\ [0x7fa59e0eb4] -> #7 connection_deinit[0x7fa59e6430] -> #8 pool_datastack_create[0x7fa5a0e7b0] -> #9 array_bsearch_insert_pos_i[0x7fa59d46d0] -> \ #10 t_base64_scheme_decode[0x7fa59dcb24] -> #11 buffer_append[0x7fa59dd3d0] -> #12 replicator_queue_push [0x556e8f4f40] -> #13 replicator_brain_in\ it[0x556e8f53e4] -> #14 _start[0x556e8f3d70] -> #15 _start[0x556e8f3f04] -> #16 io_loop_call_io[0x7fa5a02f20] - > #17 io_loop_handler_run_internal\ [0x7fa5a05450] -> #18 io_loop_handler_run[0x7fa5a05680] - > #19 io_loop_run[0x7fa5a05944] -> #20 master_service_run [0x7fa59684b0] -> #21 main[0x55\ 6e8f3930] -> #22 __libc_init_first[0x7fa57676d0] -> #23 __libc_start_main[0x7fa5767780] -> #24 _start[0x556e8f3b80] Jan 28 11:26:12 la dovecot[1187]: replicator: Fatal: master: service(replicator): child 1185 killed with signal 6 (core dumps disabled - https://\ dovecot.org/bugreport.html#coredumps) This is on Debian 12. # dovecot --version 2.3.19.1 (9b53102964) There are gigabytes of free RAM, though it says it is trying to allocate 268,435,496 bytes--on the stack--and that does seem wrong in any case. The Maildir being synced is over 430MB, is that too much for dsync?? Suggestions? Thanks, -kb _______________________________________________ dovecot mailing list -- dovecot@dovecot.org To unsubscribe send an email to dovecot-leave@dovecot.orgIt's maybe https://github.com/dovecot/core/commit/ d3a097075afb7b5d8955978ed4a2ae99add11f88.patch
Aki
Thank you!
This patch works.
On 1/28/25 11:53 AM, Aki Tuomi wrote:
It's maybe https://github.com/dovecot/core/commit/d3a097075afb7b5d8955978ed4a2ae99add11...
Oh, fun. Time to figure out how to build a .deb!
And…it's not that bad. The correct way to clear the cruft to build subsequent times wasn't immediately obvious, but I managed.
But I got some test failures:
array_free ........................................................... : ok test-backtrace.c:14: Assert failed: strstr(str_c(bt), "test_backtrace") != NULL backtrace_append ..................................................... : FAILED test-backtrace.c:38: Assert failed: strstr(bt, "test_backtrace") != NULL backtrace_get ........................................................ : FAILED base32_encode() with padding ......................................... : ok
Hmmm. Went back to original sources and did a build and they get the same errors.
So I put the decrement back in, commented out the unhappy asserts, and did another build.
I get over 30 .deb-files for output, I'm going to verify I know what I need to install, have a revert-plan ready, and try it tomorrow, when I am fresh.
But I thought I'd mention the test failures…
-kb, the Kent who is unnerved by failing tests, even when they look innocuous.
participants (3)
-
Aki Tuomi
-
Dmitry Melekhov
-
Kent Borg