Hi,
I have a 5 node rabbitmq cluster with classic queues with no
mirroring. Recently 4 nodes have crashed (on random days, not together)
giving a segmentation fault in beam.smp (seen in dmesg)
[Tue Jul 23 09:37:29 2024] 3_scheduler[1046232]: segfault at 7f4c7dfe3f68 ip 000055878990cd4e sp 00007f4f96db49a0 error 4 in beam.smp[5587897da000+2b1000]
[Tue Jul 23 09:37:29 2024] Code: 0f 84 ce 00 00 00 48 8b 7b 08 48 c7 c0 ff ff ff ff f0 48 0f c1 47 10 48 83 f8 01 0f 84 03 01 00 00 48 8b 5b 10 48 85 db 74 70 <48> 8b 03 83 e0 3c 48 83 f8 14 74 6e 48 83 f8 20 75 c0 4c 8b 63 18
Some other symptoms are server is on high load around the crash (Machine has 40 cores and load is around 150+)
Queues are in flow state as well.
RMQ version is 3.12.10
Erlang version is :25.3.2.9
OS is : Ubuntu 20.04.6 LTS
What can be the reason for this, is there anything we are missing out here ?