This could be a potential bug. Could you file an issue on github about this? Please also provide your reproduction code.
On Thursday, October 4, 2018 at 12:39:44 AM UTC-7, michi....@crowdstrike.com wrote: > > I have a simple C++ grpc client that talks to a grpc echo server written > in Go via a bidir stream. The messages that are sent back and forth are > binary blobs around 200 bytes in size. > > On the C++ client side, I have a reader thread and a writer thread. The > reader sits in a loop and calls a blocking Read() on the stream from the > server and consumes messages; the writer waits for some data to arrive from > an external source and writes it to the stream before waiting for some more > data to arrive. The channel and stream are immutable and created before the > two threads are created. The only concurrent calls into the grpc runtime > are the calls to Read() and Write() on the stream. I'm not calling Read() > concurrently with another Read(), or Write() concurrently with another > Write(). At any one time, there is at most one blocking Read() and one > blocking Write() call in progress, each made from the same single reader or > writer thread each time. > > All this works fine. I can move around 51 k messages per second, and > exchange millions of messages without any issue. The whole process is > basically in steady state. No connection is lost or the like, the channel > and stream are never re-initialized, etc. Everything stable and perfect. > > Then, after 15 or 20 minutes of this, I get the assertion failure below. I > am certain that this isn't something stupid, such as memory corruption or > the like. The code runs clean with address sanitizer and doesn't leak. > All this is with grpc and my own code compiled with gcc-7.3.0 on Ubuntu > 16.04. > > I have two questions: > > 1) Does the stack trace below ring any bells? All other threads are where > I would expect them to be, such as in a poll waiting for data to arrive and > similar. > > 2) I compiled grpc and my own code with thread-sanitizer. When I run the > code, tsan reports a bunch of race conditions. Are these possibly related > to the crash? > > Thanks, > > Michi. > > Stack trace: > https://www.dropbox.com/s/n2494rjnhj0hq56/stack-trace.txt?dl=0 > > Tsan output: https://www.dropbox.com/s/7fkxdu3rmcji6eo/tsan.txt?dl=0 > -- You received this message because you are subscribed to the Google Groups "grpc.io" group. To unsubscribe from this group and stop receiving emails from it, send an email to grpc-io+unsubscr...@googlegroups.com. To post to this group, send email to grpc-io@googlegroups.com. Visit this group at https://groups.google.com/group/grpc-io. To view this discussion on the web visit https://groups.google.com/d/msgid/grpc-io/4237fbb8-57cc-4e88-8ee1-275d0179d522%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
