obd部署了 3节点的集群，想测试下高可用，手动kill了 一个observer进程。目前用如下命令拉不起来，报错没看懂。

[admin@localhost lib]$ /home/admin/myoceanbase/oceanbase/bin/observer -r 10.186.57.94:2882:2881;10.186.57.7:2882:2881;10.186.57.5:2882:2881 -p 2881 -P 2882 -z zone1 -n myoceanbase -c 1695279065 -d /home/admin/myoceanbase/oceanbase/store -o __min_full_resource_pool_memory=2147483648,enable_syslog_recycle=True,enable_syslog_wf=False,max_syslog_file_count=4,memory_limit=9G,datafile_size=60G,system_memory=3G,log_disk_size=22G,cpu_count=16
/home/admin/myoceanbase/oceanbase/bin/observer -r 10.186.57.94:2882:2881
rs list: 10.186.57.94:2882:2881
create pid dir fail: ./run/
============= [AFTER_DESTROY] begin to show unstopped thread =============
[AFTER_DESTROY] detect unstopped thread, tid: 2993, name: observer
============= [AFTER_DESTROY] finish to show unstopped thread =============
-bash: 10.186.57.7:2882:2881: command not found
-bash: 10.186.57.5:2882:2881: command not found