Version: Labs - v3.0.0 (Stable 1526) running on ubuntu 8.04 JeOS VM with gobs of RAM, using iptables to redirect privileged ports. I don't know all the information that will be needed to solve this problem, but I'm happy to provide it.
The problem is like this.
When you start Tomcat fresh, alfresco runs like a champ. It works great in our 10 user environment with users coming and going via CIFS and WebDAV without trouble for at least 6 hours. Never a problem until about ~6 hours after a restart. Then, inevitably one user will have a problem navigating the directory tree, or saving or opening a file. Maybe they're mapped drive (implemented via CIFS in Windows) will fail to open. The following error shows up in catalina.out.
java.io.IOException: Connection reset by peer at sun.nio.ch.FileDispatcher.write0(Native Method) at sun.nio.ch.SocketDispatcher.write(SocketDispatcher.java:29) at sun.nio.ch.IOUtil.writeFromNativeBuffer(IOUtil.java:104) at sun.nio.ch.IOUtil.write(IOUtil.java:75) at sun.nio.ch.SocketChannelImpl.write(SocketChannelImpl.java:334) at org.alfresco.jlan.smb.server.nio.ChannelPacketHandler.writeBytes(ChannelPacketHandler.java:133) at org.alfresco.jlan.smb.server.nio.NetBIOSSMBChannelHandler.writePacket(NetBIOSSMBChannelHandler.java:199) at org.alfresco.jlan.smb.server.SMBSrvSession.procNetBIOSSessionRequest(SMBSrvSession.java:878) at org.alfresco.jlan.smb.server.SMBSrvSession.processPacket(SMBSrvSession.java:1415) at org.alfresco.jlan.smb.server.nio.NIOCIFSThreadRequest.runRequest(NIOCIFSThreadRequest.java:105) at org.alfresco.jlan.server.thread.ThreadRequestPool$ThreadWorker.run(ThreadRequestPool.java:141)
I know a timeout error isn't particularly helpful, but I don't know where else to start.
Sometimes this is the end of alfresco until a restart - other times, it continues to work for other users. What the heck is going on? Any ideas?
Let me know what information I can provide to help figure this thing out. There are people talking about dropping DMS all together for a simple file-share which I know they will regret, but this just isn't workable.
We had the same problem (also Ubuntu). We just stopped using CIFS.
Since then there are still java errors in the log but the system seems to runs fine. I had the impressions the java exceptions where the cause for the server slowing down after some time.
I don't know if it's connected but this post regarding java errors from a user Meredic really troubled me.
Quote: "Re: Mediawiki Integration with Labs (stable) by mederic on 04 Feb 2009, 17:24
Hello
I have looked at the Java source code to see what changed between the 3.0.0c and the 3.0.0stable. These are my conclusions:
In the 3.0.0c, modules were executed using system user credentials. This is not the case with 3.0.0stable: modules are executed using anonymous credentials, and as they are not allowed to, Java exceptions are thrown. I modified the classes "org.alfresco.repo.module.ModuleComponentHelper", "org.alfresco.repo.module.ModuleServiceImpl", "org.alfresco.repo.admin.registry.RegistryServiceImpl" and several XML files, rewriting the references concerning "AuthenticationComponent" objects which existed on the previous software releases.
It seems to work fine on my development server, but I advice you to wait for a new official release of the software to deploy on your production server, all the more since there is probably a good (security) reason why Alfresco developers removed system user credentials for modules launcher.
