Thanks Hong... Not sure if I hit on that one in JIRA or not when I searched,
appreciate the link. I can't move to 4.0 because this is a production
environment for us. When 4.0 hits stable release cycle I will definitely
move. We also have our domains on an NFS share for HA purposes, with a VIP
failover between nodes when needed. I was concerned that NFS could be a
problem given some of the exceptions, but I read some text in the Oracle docs
indicating NFS was a viable option for DAS HA. I do believe we had a past NFS
issue, and that may have caused some stale file handles to be recorded in the
domain directory. We are going to monitor for these types of files and
exceptions going forward, because at this time I think those events may have
been responsible for our odd deployment behavior. [quote=hzhang_jn]This seems
to be a similar problem as reported here:
http://java.net/jira/browse/GLASSFISH-18376 This issue was fixed in GlassFish
4.0 now. Are you able to move to GlassFish 4.0 to give a try there? Does this
affect how the application runs or the application still runs fine regardless
of this? No, redeploy is similar to undeploy/deploy, and CLI is similar to
GUI, they share similar code path in the backend.. - Hong[/quote]
--
[Message sent by forum member 'sspyrison']
View Post: http://forums.java.net/node/891558