Kamesh Jayachandran <kam...@collab.net> writes: > First I had a plan to persist the occurence of preceding 'MKACTIVITY' > on the connection pool to classify subsequent PROPFIND on same > connection as 'commit PROPFIND'.
That does sound like a better solution, but I suppose it depends on how clients use connections. > > Justin was not happy about this as he was saying some bad old clients > can make each requests in its own connection so this detection can be > faulty for those poor clients. Would it be possible for the server to detect the slave error and then try the master? -- Philip
