tengine version is 2.2.2

also see some memory error in tengine log:
2019/01/09 20:04:02 [info] 42#0: *259 [dyups] post upstream name: elasticsearch, client: 127.0.0.1, server: , request:
"POST /upstream/elasticsearch HTTP/1.1", host: "localhost:8081"
2019/01/09 20:04:02 [info] 42#0: [dyups] create upstream dyups_upstream_sandbox
2019/01/09 20:04:02 [crit] 42#0: ngx_slab_alloc() failed: no memory
2019/01/09 20:04:02 [notice] 39#0: signal 17 (SIGCHLD) received
2019/01/09 20:04:02 [alert] 39#0: worker process 42 exited on signal 11 (core dumped)
2019/01/09 20:04:02 [alert] 39#0: shared memory zone "ngx_http_upstream_check#1" was locked by 42
2019/01/09 20:04:02 [alert] 39#0: shared memory zone "ngx_http_dyups_module#1" was locked by 42

Dyups works ok in 2.2.0, but not in 2.2.1 and 2.2.2 for me

Seem to be related to no of servers when calling http://localhost:8081/upstream
If I call with 6 servers it works ok, but when using 7 it fails:

server x.x.x.1:8080;server x.x.x.2:8080;server x.x.x.3:8080;server x.x.x.4:8080;server x.x.x.5:8080;server x.x.x.6:8080
this works ok

server x.x.x.1:8080;server x.x.x.2:8080;server x.x.x.3:8080;server x.x.x.4:8080;server x.x.x.5:8080;server x.x.x.6:8080; server x.x.x.7:8080

fails with this after posting to /upstream:
'Connection aborted.', RemoteDisconnected('Remote end closed connection without response',))

Error happens after about 4 calls to /upstream and then starts to what seems core dump in nginx, when calling:
curl -v --data @test2.txt localhost:8081/upstream/test

• Trying 127.0.0.1...

• Connected to localhost (127.0.0.1) port 8081 (#0)
  POST /upstream/test HTTP/1.1
  Host: localhost:8081
  User-Agent: curl/7.47.0
  Accept: /
  Content-Length: 366
  Content-Type: application/x-www-form-urlencoded

• upload completely sent off: 366 out of 366 bytes

• Empty reply from server

• Connection #0 to host localhost left intact
  curl: (52) Empty reply from server

same call have worked initial before, so seems like a memory issue as per below debug lines in nginx error-log for above curl call:

019/01/24 21:04:23 [debug] 95#0: *1112 [dyups] interface read post body
2019/01/24 21:04:23 [info] 95#0: *1112 [dyups] post upstream name: test, client: 127.0.0.1, server: , request: "POST /upstream/test HTTP/1.1", host: "localhost:8081"
2019/01/24 21:04:23 [debug] 95#0: shmtx lock
2019/01/24 21:04:23 [debug] 95#0: [dyups] read msg 95
2019/01/24 21:04:23 [debug] 95#0: [dyups] process 95 update time 1548363863209
2019/01/24 21:04:23 [debug] 95#0: [dyups] find dynamic upstream
2019/01/24 21:04:23 [info] 95#0: [dyups] create upstream dyups_upstream_sandbox
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 00000000016D5080:512 @16
2019/01/24 21:04:23 [debug] 95#0: malloc: 00000000016DD9E0:512
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 000000000162D4E0:512 @16
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 000000000162D6F0:512 @16
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 000000000162D900:512 @16
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 000000000162DB10:512 @16
2019/01/24 21:04:23 [debug] 95#0: malloc: 00000000016E0C50:768
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 00000000016E0F60:512 @16
2019/01/24 21:04:23 [debug] 95#0: malloc: 00000000016E1170:1536
2019/01/24 21:04:23 [debug] 95#0: posix_memalign: 00000000016E1FF0:512 @16
2019/01/24 21:04:23 [debug] 95#0: init least conn
2019/01/24 21:04:23 [debug] 95#0: malloc: 00000000016E2200:1560
2019/01/24 21:04:23 [debug] 95#0: http upstream check add upstream process: 3
2019/01/24 21:04:23 [debug] 95#0: shmtx lock
2019/01/24 21:04:24 [notice] 84#0: signal 17 (SIGCHLD) received
2019/01/24 21:04:24 [alert] 84#0: worker process 95 exited on signal 11 (core dumped)

Have identified the issue starts in commitid 2ece5b2 and works ok in previous 82ebf09

You could try the latest version of Tengine-2.3.0 or Tengine-2.3.1.