FREMONT: Beginning ~07/01/2015 I began receiving intermittent errors as above ~45/720 connections per day, and seem to be slowly increasing, and only occurring on 2 of our 16 Linodes. I changed this connection from the private IP to the primary destination IP and this appears to not be seeing the failure. These ssh cron connections send snipets to the same system every 2 minutes, and the counts are reported by LOGWATCH.

The failing servers are on fremont793 and fremont736

There is nothing significantly different between these 2 systems and the other 14, however I have never seen a connection count difference from the expected 720/day of more than 1 on any of the others.

I added some additional logging to see when the errors happen (although since I am now talking on the primary IP may not see any) and while this particular activity is not critical, we do have other inter-system communications using the private IP that are.

Any suggestions or guesses about this would be appreciated.

In the long term, moving to IPv6 avoids the need for the special intra-datacenter addresses, and so you should probably do that at some point.

In the short term... have you opened a ticket with Linode support?

_________________

/* TODO: need to add signature to posts */