At irregular intervals our Quickbuild agents will lose all network connection, can't ping in or out, won't contact the QB Coordinator. This last for approximately 60 seconds or less in each case before coming back up as if nothing has happened. We've checked the hosts that the VMs are on, all of the networking in between, there are no issues and for all intents and purposes the servers look absolutely fine. The coordinators will even serve web pages through IIS during the downtime, but the agents simply can't see it.
This happens during a build, and will affect a portion of our build farm but not all of it. It doesn't affect any other servers on the network and is localised entirely within the QB environment. There's nothing in any of the QB logs, nothing in the server event logs that could describe this issue, so we're at a loss.
|
