On Monday 08 March 2004 23:47, you wrote: > What happens if the network or some other part of the infrastructure > goes down such that one of the nodes can't communicate with the master? > Do you loose log messages? Does the node block? > > What happens when the master fills it's log disk, do all the clients > block? > > You're going to need a temporary buffer, i.e. a log file, on each node > and some error recovery strategy. The easiest thing to do would be to > write a scheduled proc which scans the existing log file from the > previous seek point up to the current end of file. You could use > ns_httppost etc. to send back packets of log data. >
Good remarks. You are absolutely correct. But, I did not presented the solution yet, only idea, right? This (exception recovery) should obviously be part of the (whatever) solution. BTW, AS drops log entries on the floor when the filesystem where the logs reside is full. Therefore, the answer to the last question is simple: logs are simply lost. There might be more elaborate mechanism employed for writing to sockets, but this is still to think over. As I said, I have no implementation ready yet, just tinkering with the idea and gathering some ideas/feedback. Now, the idea of scanning the local logile and updating chunks at a time... this could be also a viable solution. Cheers, Zoran -- AOLserver - http://www.aolserver.com/ To Remove yourself from this list, simply send an email to <[EMAIL PROTECTED]> with the body of "SIGNOFF AOLSERVER" in the email message. You can leave the Subject: field of your email blank.
