Kim: Since you can identify the client (because the user is complaining) can you tell me whether this is a client that supports UUIDs or not?
cmdebug <host> 7001 -addr Is this host behind one of the NATs? The message occurs when there is an existing connection associated with a client and the client initiates a new one. The data structures are not being clobbered. The delay is most likely related to the file server seeing a new connection, calling TellMeAboutYourself to determine who the client is, and then matching up the client with any existing host entries and cleaning up current connections. The question to try and answer is why are the clients and the file servers out of sync with regards to their current connection state. Are these clients that have suspended? Lost their network address? or something else? Jeffrey Altman Kim Kimball wrote: > The thing that's interesting is that the messages we've seen are always > the same pattern: > > Sat Jan 21 14:24:56 2006 FindClient: client X(Y) already > had conn 12bbb28 (host 894ea06a), stolen by client X(Y) > > This looks like the same client/SID is stealing from itself?? > > For the 2 hosts almost always identified (host xyz) we're seeing > intermittent data access failures that we're trying to sort out. > > We've seen a couple of hundred of these messages for each of the clients > over a period of < 1 day. > > Any reason for concern? It looks like one client structure is > clobbered/reassigned, and I've been assuming that this is so fast that > the AFS client doesn't notice. > > Kim
smime.p7s
Description: S/MIME Cryptographic Signature
