zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jun Rao <jun...@gmail.com>
Subject too many client EventThreads
Date Tue, 25 Jan 2011 00:45:21 GMT
Hi,

I had a problem very similar to an old post:
http://www.mail-archive.com/zookeeper-user@hadoop.apache.org/msg01570.html

Basically, we get lots of
org.apache.zookeeper.ClientCnxn$SendThread<http://boverstr-md.linkedin.biz:7000/object/0xfffffd7f7a3f7080>threads
with long thread names like
"task-EventThread-EventThread-EventThread-EventThread". In our setup,
maxClientCnxns is the default 10 and we probably do have more than 10
clients in the same JVM trying to connect to the same ZK server
concurrently. What exactly happens in that situation? Do those extra clients
can't establish connections at all or do clients start to thrash? Also, we
do create new ZK clients in the watcher (when a session expires), but we
seem to close the old clients. Are there any other reasons that lots of
org.apache.zookeeper.ClientCnxn$SendThread<http://boverstr-md.linkedin.biz:7000/object/0xfffffd7f7a3f7080>threads
could be created? Thanks,

Jun

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message