When the business volume of our search service is large, the application may fail to get the redis connection from time to time. All threads of the entire program are stuck in the following position, resulting in the inaccessibility of new requests from the front end, the search service is suspended and the entire program does not respond.
Thread 4 (thread 0x7ff97222d700 (lwp 222201 )):
#0 0x000000339f2e15e3 in select () from/lib64/libc. so.6
#1 0x00000000009628fc in rediscontextwaitready ()
#2 0x0000000000962e08 in rediscontextconnecttcp ()
#3 0x0000000000960293 in redisconnectwithtimeout ()
#4 0x000000000076ab07 in redisoperator: getredisconnection ()()
This problem is very strange. Sometimes it does not occur once in a few months. Sometimes it occurs multiple times in a few days, which makes debugging very difficult.
Debugging and Analysis of redis connection problems cannot be obtained under high concurrency