Closed hl3w22bupt closed 2 years ago
After a cursory investigation, it appears this may be due to limited disk space while flushing mmap. My guess is this is the same issue. Is it possible your node is low on disk space when this error occurs?
After a cursory investigation, it appears this may be due to limited disk space while flushing mmap. My guess is this is the same issue. Is it possible your node is low on disk space when this error occurs?
emm, not sure but possible. I will notice the disk space if it reoccur
Closing this as I believe it is due to disk space. Please re-open if this is not the case.
I launched liftbridge server several days ago, and writes some data through nats-pub client continuously, unfortunately, liftbridge server paniced and exited because of
fatal error: fault SIGBUS
from the callstack dump. I have encountered panic problem before, such asfatal error: runtime: out of memory
, It is intuitive and easy to understand, I could avoidout of memory
by controlling the write qps. but in this time, theSIGBUS panic
error may implies some unsafe access. So I doubt there maybe been some bug in unsafe mmap when process commitlog, Is there anyone else encountered this?By the way, the liftbridge cluster only contains 1 server which bootstraps as raft seed with command option
--raft-bootstrap-seed
in my caseAgain, my liftbridge server is v1.5.0, golang version is 1.15.5