Update 111206: After discussing this with our customer they agreed to accept the workaround of increasing memory for the impacted machines.
Update 110823: Finally I found the trigger for this STOP error: memory pressure. Further looking into this now to see if we can indeed trigger the problem this way.
Update 110706: We now have an instrumented binary in place to gather additional information.
My colleague Daniel is working on a STOP 0x27 in mrxsmb20!Smb2ReconnectOpen_Finalize, with the following stack:
0: kd> knL
*** Stack trace for last set context - .thread/.cxr resets it
# Child-SP RetAddr Call Site
00 fffff880`0642fba0 fffff880`03203bb8 mrxsmb20!Smb2ReconnectOpen_Finalize+0x74
01 fffff880`0642fc50 fffff880`0302e1b1 mrxsmb!SmbCepFinalizeExchange+0x44
02 fffff880`0642fc80 fffff800`019197c6 rdbss!RxpWorkerThreadDispatcher+0x1a1
03 fffff880`0642fd40 fffff800`01654c26 nt!PspSystemThreadStartup+0x5a
04 fffff880`0642fd80 00000000`00000000 nt!KxStartSystemThread+0x16
We have enabled Redirector tracing and currently we're awaiting a new dump including this tracing. Let me know if you also see this issue!