hadoop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexandre Fouche <alexandre.fou...@cleverscale.com>
Subject logs full of state: diagnostics: "", exit_status: -1000,
Date Mon, 29 Oct 2012 15:32:45 GMT
Hi,

What is NodeStatusUpdaterImpl and what do these lines in /var/log/hadoop-yarn/yarn-yarn-nodemanager-....log
with exit_status=-1000 mean ? My nodemanager logs are full of it, and i cannot figure out
if it means something, or if these are just harmless or buggy messages. 346 lines out of a
job log of 1046 lines are these -1000 statuses

    2012-10-29 13:44:12,943 INFO org.apache.hadoop.yarn.server.nodemanager.NodeStatusUpdaterImpl:
Sending out status for container: container_id {, app_attempt_id {, application_id {, id:
6, cluster_timestamp: 1351504801306, }, attemptId: 1, }, id: 2, }, state: C_RUNNING, diagnostics:
"", exit_status: -1000, 

I do not see any error messages in the full log (which i posted here: http://pastebin.com/uNhWngrm
), and the job finishes successfully, but nevertheless appears extremely slow to "distcp"
a few small files. Something must be wrong but i can't figure out what

--
Alexandre Fouche
Mime
View raw message