"jesper.pedersen" wrote : a deployment should be an atomic operation.
Yep. But i think the issue is:
1) A OS copy operation is in progress
2) HDScanner wakes up and does a check for new file availability
3) OS says new file available, even though copy is not complete (i remember this on
Windows, not sure about other OS)
4) HDScanner starts processing the deployment while the file is still being copied
5) HDScanner blows up with exception since its a corrupt file.
And i guess, we will be able to reproduce this on JBossAS-4.x too.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4230449#...
Reply to the post :
http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...