ShummGen commented on issue #63609:
URL: https://github.com/apache/doris/issues/63609#issuecomment-4533264115
Adding more context about the crash dump.
We have 5 BE core dump files from repeated crashes:
```text
/var/lib/systemd/coredump/core.doris_be.82134.lz4
/var/lib/systemd/coredump/core.doris_be.83921.lz4
/var/lib/systemd/coredump/core.doris_be.84765.lz4
/var/lib/systemd/coredump/core.doris_be.85632.lz4
/var/lib/systemd/coredump/core.doris_be.86955.lz4
The latest investigated crash is:
process: doris_be
pid: 86955
time: around 2026-05-25 09:14
At this moment, I have not yet extracted a richer gdb bt from the compressed
core dump, but the crash signature from be.out is stable across repeated
crashes:
*** Signal 11 (SIGSEGV) received by PID 86955 ***
PC: @ 0x7fdd1f000000 (unknown)
*** SIGSEGV address not mapped to object (@0x7fdd1f000000) received by PID
86955 ***
Stack trace:
#0 __memmove_avx_unaligned_erms ()
#1 memcpy ()
#2 doris::vectorized::ColumnStr<unsigned int>::serialize_impl(...)
#3 doris::vectorized::ColumnStr<unsigned int>::serialize_vec(...)
#4 doris::vectorized::DistinctStreamingAgg(...)
So the currently confirmed crash path is:
DistinctStreamingAgg
-> ColumnStr::serialize_vec
-> ColumnStr::serialize_impl
-> memcpy
-> SIGSEGV
This is the same failure pattern we observed repeatedly when corrupted
segment files were read.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
