failure in shm_ib/test_ucp_perf.envelope/0

[brminich]

Happened with valgrind here: http://hpc-master.lab.mtl.com:8080/job/hpc-ucx-pr/8264/label=hpc-test-node-upstream,worker=0/consoleFull#-853622487b816df86-9afc-46eb-984f-1a9a125d73bb

22:06:31 [ RUN      ] shm_ib/test_ucp_perf.envelope/0
22:06:31 [     INFO ]                tag latency : 0.505 usec (performance not checked)
22:06:32 [     INFO ]            tag iov latency : 1.152 usec (performance not checked)
22:06:32 [     INFO ]                     tag mr : 7.481 Mpps (performance not checked)
22:06:33 [     INFO ]                tag sync mr : 1.906 Mpps (performance not checked)
22:06:34 [     INFO ]                tag wild mr : 6.694 Mpps (performance not checked)
22:06:34 [     INFO ]                     tag bw : 3844.994 MB/sec (performance not checked)
22:06:34 [     INFO ]         tag bw_zcopy_multi : 2567.160 MB/sec (performance not checked)
22:06:34 [     INFO ]                put latency : 0.301 usec (performance not checked)
22:06:35 [     INFO ]                   put rate : 8.512 Mpps (performance not checked)
22:06:35 [     INFO ]                     put bw : 8877.615 MB/sec (performance not checked)
22:06:35 [     INFO ]                get latency : 0.200 usec (performance not checked)
22:06:35 [     INFO ]                     get bw : 7912.120 MB/sec (performance not checked)
22:06:35 [     INFO ]             stream latency : 0.531 usec (performance not checked)
22:06:35 [     INFO ]                  stream bw : 3359.345 MB/sec (performance not checked)
22:06:36 [     INFO ]   stream recv-data latency : 0.547 usec (performance not checked)
22:06:36 [     INFO ]        stream recv-data bw : 9400.656 MB/sec (performance not checked)
22:06:36 [hpc-test-node-upstream:33669:1:56814]      ucp_ep.c:641  UCX Bug: pending request 0x7f01dc6611e8 on ep 0x7f0205b63000 should have been flushed
22:06:36 [hpc-test-node-upstream:33669:0:56813]      ucp_ep.c:641  UCX Bug: pending request 0x7f01ec6463e8 on ep 0x7f0205b23000 should have been flushed
22:06:37 make: *** [test] Segmentation fault (core dumped)
22:06:37 make: Leaving directory `/images/jenkins/workspace/hpc-ucx-pr-2/label/hpc-test-node-upstream/worker/0/build-test/test/gtest'
22:06:37 

Can be related to #2089 or #2354

[yosefe]

@brminich seems it's without valgrind:

22:06:37 make: *** [test] Segmentation fault (core dumped)

in case of valgrind it would have been [test-valgrind]

[brminich]

ok, did not notice that.