Whenever a proxy backend is configured with a list of URIs and the first server becomes unavailable for a while, all connections will need to try it before getting to another working one.
What about introducing an (optional) algorithm that decides, after an appropriate retry pattern, to rework the list of URIs? I understand that since the proxy backends make concurrent access to the list of URIs this may require further mutex protection; also, the algorithm may not be general enough, or too complicated for most of the applications, so the best approach would be to add some API (much like rebind) to allow applications to deal with this issue, leaving the current behavior as the default. p. Ing. Pierangelo Masarati Responsabile Open Solution OpenLDAP Core Team SysNet s.n.c. Via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it ------------------------------------------ Office: +39.02.23998309 Mobile: +39.333.4963172 Email: [EMAIL PROTECTED] ------------------------------------------
