Have many Ubuntu 20.04 servers running netdata. Version 1.44.3 had/has no issues. Some systems (all physical servers) are having issues with 1.45.3. The netdata service will try to start, fail, and try to restart, then continue in this loop with the below log for kernel segfault libpthread.
kernel: [4448124.523030] netdata[3777917]: segfault at 18 ip 00007fa35a072fc4 sp 00007ffd4d864ae8 error 4 in libpthread-2.31.so[7fa35a06e000+11000]
I tried a fresh install of 1.45.3 via ubuntu repo and running the netdata kickstart, but no luck. I had to roll the Ubuntu 20.04 systems back to 1.44.3 and all nodes are running fine without issues on the older netdata version.
Nothing stood out in the 1.45 changelog where it would cause it to act like this.
netdata.conf:
[global]
error log = syslog
access log = none
[web]
# all defaults
[plugins]
# all defaults
[health]
enabled = yes
run at least every seconds = 15
[registry]
enabled = no
# all defaults
[statsd]
enabled = yes
max private charts allowed = 5000
max private charts hard limit = 5000
private charts history = 60
update every (flushInterval) = 15
If possible, could you check also if there is anything the logs that could point to a possible issue?
Note: Sometime back our logging changed from file into systemd journal by default, in case you need some details - Netdata Logging | Learn Netdata