segmentation fault with address sanitizer

[Shbinging]

Description

When I add --enable-address-sanitizer to configure and compile, there will be a probability of segmentation fault when running zebra or mgmtd even vtysh. The specific error message is as follows

AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
AddressSanitizer:DEADLYSIGNAL
Segmentation fault (core dumped)

In the absence of seg fault, when exiting zebra, there will be the following output

==128==LeakSanitizer has encountered a fatal error.
==128==HINT: For debugging, try setting environment variable LSAN_OPTIONS=verbosity=1:log_threads=1
==128==HINT: LeakSanitizer does not work under ptrace (strace, gdb, etc)

Version

FRRouting 10.2-dev-my-manual-build (c1f2b43ca294) on Linux(6.5.0-35-generic)

Ubuntu 22.04

How to reproduce

add --enable-address-sanitizer to configure and compile, then run zebra

I use the dockerfile provided by the dev doc(https://docs.frrouting.org/projects/dev-guide/en/latest/building-docker.html)

Expected behavior

normal operation

Actual behavior

segmentation fault

Additional context

No response

Checklist

• I have searched the open issues for this bug.
• I have not included sensitive information in this report.

[rzalamena]

I ran into a similar issue sometime ago. In my case I was running out of memory due to bad defaults.

To check if it is the same issue: What's the output of ulimit -n? If FRR maximum file descriptors is not configured it will use that value.

Possible solutions:

1. Set a lower ulimit -n or configure /etc/security/limits.conf (example: * hard nofile 4096, replace * as appropriated).
2. Configure MAX_FDS in /etc/frr/daemons (if using watchfrr)
3. Use the command line: --limit-fds 4096 in all daemons (documentation)

[Shbinging]

I followed your instructions, but even after launching 3 to 4 times, I still encounter this issue once.