Malte

Thank you for the quick fix 👍

There shouldn't be any long-running connections to be honest (5s max). Our problem is: the stale haproxy processes still accept connections and cause 503 due to obviously now defunct instances.

I always get 4 because of nbproc. The issue remains even when I use nbproc 1 and hence only one PID.

Has anyone tried marathon-lb's reload command? https://github.com/mesosphere/marathon-lb/blob/master/service/haproxy/run ``` #!/bin/bash exec 2>&1 export PIDFILE="/tmp/haproxy.pid" exec 200

Found https://github.com/hashicorp/consul-template/issues/442 and https://github.com/golang/go/issues/13164 Could actually be related to Go. I just compiled bamboo with go 1.6 and will update this accordingly. BTW: Another reload script that I might try...

I have not seen this issue since upgrading to 1.6 Malte On Tue, Apr 19, 2016 at 08:16 imrangit [email protected] wrote: > @elmalto https://github.com/elmalto: were you able to resolve the...

Marathon used to be somewhat inconsistent with health checks. It also shuts down old tasks before the new one is healthy. Just two things to consider imho Timo Reimann [email protected]...

I use cassandra 3.0.8 and with cassandra-driver-core 3.0.7 and cassandra-all 3.10 but get the errors ``` java.lang.RuntimeException: Failed to list files in ... at org.apache.cassandra.db.lifecycle.LogAwareFileLister.list(LogAwareFileLister.java:77) at org.apache.cassandra.db.lifecycle.LifecycleTransaction.getFiles(LifecycleTransaction.java:561) at org.apache.cassandra.io.sstable.SSTableLoader.openSSTables(SSTableLoader.java:76) at...