port lost due to mem fulled after running for a while

Please answer these questions before submitting your issue. Thanks!

(Please mention that if the issue you filed is solved, you may wish to close it by yourself. Thanks again.)

(PS, you can remove 3 lines above, including this one, before post your issue.)

What version of shadowsocks-libev are you using?

shadowsocks-libev 3.3.5

What operating system are you using?

centos 7 el7.elrepo.x86_64 with 512Mem

What did you do?

nohup ss-manager -c /etc/shadowsocks-libev/config.json -u &

What did you expect to see?

the server run its config stable and last

What did you see instead?

some port will lost after half a day or a day later, and the memory will be occupied.

What is your config in detail (with all sensitive info masked)?

{ "server":"0.0.0.0", "local_port":1080, "port_password":{ "21135" : "xxxxxxx", "21887" : "xxxxx", "41266" : "xxx", "51266" : "xxxxx" }, "timeout":60, "method":"chacha20-ietf-poly1305", "mode":"tcp_and_udp", "ipv6_first":false, "dns_ipv6":false, "fast_open":true, "plugin":"obfs-server", "plugin_opts":"obfs=http;failover=127.0.0.1:10080;fast-open" } I run nohup ss-manager -c /etc/shadowsocks-libev/config.json -u & command to offer service for myself and friends but after a day or a while.some port like port 21887 will be close. I check using netstat -an | grep 0.0.0.0.and i found the the mem is fulled with command free -m