On Wednesday 19 December 2007 12:48, Linda Walsh wrote: > ... > > ---- > UDP is stateless (no connection needed) -- but it is a reliable > transport for NFS since NFS keeps track of the packet ordering and > the packets which are 'outstanding'.
Not to mention its own checksumming, time-out and retrying. > ... > >> Perhaps things would improve using "async", that's a difference > >> with local filesystems that run "async" by default. But the man > >> page says it can be dangerous. > > > > VERY! In particular, you can get notification of an error (e.g., "disk full") on the reply to a request much later than that of the request which actually encountered the error. Software with complex ordering and error sensitive behavior can be seriously undermined by asynchronous NFS. E.g. one of the most subtle problems I ever debugged was corruption in Berkeley DB files when the aforementioned disk-full condition occurred on a file system being accessed by the BDB code over asynchronous NFS. > ... Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
