We're using syslog-ng 1.5.13 (yes, I know) on a bunch of diskless clients to send log messages to a central host. The main switch in our machine room had a hardware problem, and refused to pass packets for 20 minutes or so until we rebooted it. On two of my clients, syslog-ng subsequently stopped sending log messages to the central host after the switch recovered. Here's what was in the log: Mar 1 08:55:38 syslog-ng[13050]: io.c: do_write: write() failed (errno 9), Broken pipe Mar 1 08:55:38 syslog-ng[13050]: pkt_buffer::do_flush(): Error flushing data Mar 1 08:55:38 syslog-ng[13050]: Connection broken to AF_INET(loghost.example.com:514), reopening in 60 seconds Mar 1 08:57:53 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:00:08 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:02:24 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:04:39 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:05:38 syslog-ng[13050]: STATS: dropped 0 Mar 1 09:06:54 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:09:10 syslog-ng[13050]: Error connecting to remote host AF_INET(loghost.example.com:514), reattempting in 60 seconds Mar 1 09:25:39 syslog-ng[13050]: STATS: dropped 0 Mar 1 09:35:39 syslog-ng[13050]: STATS: dropped 0 The other host has a remarkably similar log. After 9:25:39, the only messages that made it to the central host are the "STATS: dropped 0" messages. Is this a bug that has been fixed in more recent versions? We've been meaning to upgrade the clients for a while (the central server is using a current syslog-ng), but haven't gotten to it yet.