cassandra连不上,报Stop listening for CQL clients, Failed managing commit log segments
less /var/log/cassandra/system.log
这个时候其实是cassandra检测到commit log不一致,如果conf里commit_failure_policy: stop,就会停下来
如果忽略这个错误,改为commit_failure_policy: ignore 强制启动,会报错:
这就是说,system.peers这个系统表的sstable,即内存表已经不一致了,就是rdms里常说的坏表
这就需要repair掉corrupt table:
./sstablescrub system peers
fix后,就可以正常启动了
sort of, I have some experience in the domain of database(MySQL/mongo), java, python, front-end, etc. I'll willing to give and accept bits of help from others.
now base in Singapore.