One node on production had an issue receiving messages (not responding) but gaining access to the node it was noticed that the recv queue of the rmb connection is filling up but seem the rmb process is not reading that data:
This can be duo to a software bug where the read routing got stuck somehow. Seems that there might be a problem with the detection of connection stalling as well that need to be reviewed
One node on production had an issue receiving messages (not responding) but gaining access to the node it was noticed that the
recv
queue of the rmb connection is filling up but seem the rmb process is not reading that data:This can be duo to a software bug where the
read
routing got stuck somehow. Seems that there might be a problem with the detection of connection stalling as well that need to be reviewed