Sorry, I just have a new question or comment. Will it be possible to use this containers architecture, to enable us to migrate the entire group of processes from one machine to another? Or am I confusing it with CONFIG_NAMESPACES?
The application of this is in Google's search engine cluster - where in USENIX 2004 it is mentioned that a single query can potentially passed through 1000+ physical machines (here the podcast online). So it will be useful if it is possible to migrate the entire processing from one machine to another. Sometimes this is useful, esp in scenario of gigantic transaction, where restarting the transaction from beginning again is expensive (search engine or terabyte database processing), so process migration across machines (or process duplication - from one machine to another, and continuing the execution at the point where it has been left off in the previous machine) will be useful. Comments? -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [EMAIL PROTECTED] Please read the FAQ at http://kernelnewbies.org/FAQ
