-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 John,
On 9/13/2011 5:51 AM, John Bass wrote: > In the event of a node failure, I'm assuming that there's no way to > recover from that and the failure will be visible to a client > application. Correct: no other node in the cluster can serve the response being generated by a dying Tomcat instance. As Pid points out, this isn't unique to Tomcat. > Similarly, if a node fails during a long running calculation, I'm > assuming that there's no way to persist that execution state. There's nothing that Tomcat does that would persist any state, unless your "long-running calculation" periodically saves it's state into the user's session, and you are using distributed or persisted sessions. If you have long-running tasks, I would encourage you to architect the code such that the state /can/ be saved somewhere trivial such as in an HttpSession or even a relational database, and that a replacement data processing thread can take-over and resume operation on a partially-completed job. If the initial node goes down, a second request that goes to another node can resume that operation in-progress without starting over again. - -chris -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk5xJJgACgkQ9CaO5/Lv0PBbrwCgofwAJWqCWEImqEvpDZE16QqX oLAAnjFJDJWeJBElIUaImqZbRrTS4wY/ =+cpD -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org