Fix akka logging initialization timeout on startup

We've seen CDS internittently failing to startup b/c akka fails to
initialize its logging system - it times out after 5 sec. It seems
on startup threads may be busy enough to prevent akka's message
to inialize the logging actor to timeout. Interestingly, We're only
seeing this on RedHat. The timeout is configured via
logger-startup-timeout which defaults to 5 sec. Increasing it fixes
the issue. I made it really high, 5 min, since if it times out we're
dead in the water anyway.

Change-Id: Ic95d7298b9320f03f664e9f2b171f980546ca95d
Signed-off-by: Tom Pantelis <tpanteli@brocade.com>
(cherry picked from commit 94c2eb2becdb561032e62a2897b15f52bd8e3643)

diff --git a/opendaylight/md-sal/sal-clustering-config/src/main/resources/initial/akka.conf b/opendaylight/md-sal/sal-clustering-config/src/main/resources/initial/akka.conf
index a4d554c2bfba99751fbc58c7bbb963c6e03c8602..f542847fa139e814fb81424f0f22f33a5693cce7 100644 (file)
--- a/opendaylight/md-sal/sal-clustering-config/src/main/resources/initial/akka.conf
+++ b/opendaylight/md-sal/sal-clustering-config/src/main/resources/initial/akka.conf
@@ -11,6 +11,7 @@ odl-cluster-data {
   akka {
     loglevel = "INFO"
     loggers = ["akka.event.slf4j.Slf4jLogger"]
+    logger-startup-timeout = 300s
 
     actor {
       provider = "akka.cluster.ClusterActorRefProvider"
@@ -83,6 +84,7 @@ odl-cluster-rpc {
   akka {
     loglevel = "INFO"
     loggers = ["akka.event.slf4j.Slf4jLogger"]
+    logger-startup-timeout = 300s
 
     actor {
       provider = "akka.cluster.ClusterActorRefProvider"