Impala has two client interfaces with slightly different session behavior: Beewax (default port 21000) - sessions are created when the client connects Hiveserver2 (default port 21050) - sessions are created when OpenSession() is called
In both cases, sessions can be closed either if - the connection ends - someone presses "close" on the /sessions page of the debug webui For hiveserver2, sessions can also be closed if CloseSession() is explicitly called by the client So probably the most likely cause of your issue is that the connections are being dropped somehow, possibly due to the load balancer. If sessions are in fact getting closed due to connections being dropped, you should see some lines of the form "Connection from client <client hostname> closed, closing <num> associated sessions" Fwiw, this behavior was recently changed so that hiveserver2 sessions are not closed immediately when the connection ends but timeout if there hasn't been an associated connection for a configurable timeout: https://issues.apache.org/jira/browse/IMPALA-1653 though we haven't done a release since that work went in On Mon, Aug 5, 2019 at 3:40 PM Antoni Ivanov <aiva...@vmware.com> wrote: > Hi, > > I am investigating the most common errors we see in our Impala Cluster. > The most common is with query status = 'Session Closed' > > I can see from the code ( > https://github.com/apache/impala/blob/72c9370856d7436885adbee3e8da7e7d9336df15/be/src/service/impala-server.cc#L1435 > ) > that it is set when Session is closed and this happens when connection is > closed (ConnectionEnd< > https://github.com/apache/impala/blob/72c9370856d7436885adbee3e8da7e7d9336df15/be/src/service/impala-server.cc#L2094 > >) > and this is called when Thrift transport is closed< > https://github.com/apache/impala/blob/82f753e3044bd2482f35d137fbb28516fc0ef86c/be/src/rpc/TAcceptQueueServer.cpp> > (and query has not completed or failed in some way it would be marked as > Session Closed > > Does this mean that the remote end has simply dropped the connection ? > E.g there has been network interruption or someone killed (SIGKILL) the > remote process ? > We have (TCP) load balancer (HaProxy) and I am wondering if for example > Load Balancer tcp timeout can cause such error. Or can client socket > timeout cause it? > > I'd be grateful for any insides into the semantics of when "Session > Closed" is set. > > > > Thanks, > Antoni >
