icymoon
icymoon
IP分片支持+1 sessio级别的日志+1
Updated to the latest version. New problem, many connections don't release after long large pressure. Debug is enabled: CFLAGS += -D CONFIG_RECORD_BIG_LOOP CFLAGS += -D CONFIG_DPVS_SAPOOL_DEBUG CFLAGS += -D CONFIG_DPVS_IPVS_DEBUG...
flood ping, curl for http, dig for dns look up And many kinds of tcp/udp/icmp packets built and sent by scapy script: 1）icmp echo-request & echo reply 2) TCP syn,...
This issue should be closed now I think. I tested it over a weekend and no core file generated. Thanks.
No session leak find last night, thank you.
It easy to reproduce with workload. Monitors call dpip, ipvsadm command with different parameters.
每段开始出问题的时间，都是以这样的消息开头的 MSGMGR: multicast_msg_send: mc_msg(type:11, cid:0->slaves) timeout(2000 us), drop... SERVICE: dp_vs_copy_stats: send message fail. 或者 msg_send: uc_msg(type:8, cid:0->13, flags=2) timeout(2000 us), drop... NETIF: [netif_sockopt_get] msg dropped for netif sockmsg opt 502)...
reproduced with this work load: \# ./wrk -t26 -c5000 -d15m http://xxxx/ Running 15m test @ http://xxx/ 26 threads and 5000 connections Thread Stats Avg Stdev Max +/- Stdev Latency 4.01ms...
谢谢@ywc689 或者我只尝试读per core的计数然后在用户态加起来，可以规避这个问题？
谢谢@ywc689 统计数据抖一下没关系，最多是图不好看，我最怕的是处理卡顿意味着有稳定性或者性能隐患。。。