We're logging the data to a central database from 2 relays. Is there a way for syslog-ng to buffer the data while the database server is rebooting or innaccessible? Here's our current setup. All machines forward the syslog msg to either a production or development relay which runs syslog-ng and then the messages are forwarded to a server running an oracle DB. You may wonder why we have 2 relays... well we have a few hundrer servers and workstations and we're just implementing the DB aspect of it. 15 Millions rows for 6months growing by 100K rows/day. on the DB server I have a custom written syslogd which buffers data in case the DB is down but the problem resides with syslog-ng interrupting the relay, for 60 seconds, of messages to the oracle DB server if the server dies or reboots. Is there someway for syslog-ng to relay missed messages? The only other way I can think of doing it is kinda messy... having syslog-ng dump to a FIFO pipe where a perl script could buffer the data in an array or file until the DB server comes alive again. We're using 1.4.14 -------------------- - Cloud of servers - -------------------- / \ / \ -------- ------- relays - prod - - dev - -------- ------- \ / \ / ------------- - oracle DB - ------------- Sylvain Martin