Community
Participate
Working Groups
Build Identifier: RSE 3.2.0 A dstore connection to a linux host was activaed. The host threw exception and the connection dropped while letting the connection idle Reproducible: Always Steps to Reproduce: 1. Create dstore connection to a linux server 2. Start the server with the following command system("java -Xscmx5m -Xshareclasses:name=RSE%g,groupAccess,nonfatal -Xms10m -Xmx64m -DA_PLUGIN_PATH=\$A_PLUGIN_PATH -DDSTORE_SPIRIT_ON=true -DSPIRIT_EXPIRY_TIME=30 -DSPIRIT_INTERVAL_TIME=30 -Dfile.encoding=ISO-8859-1 -DDSTORE_TRACING_ON=true org.eclipse.dstore.core.server.Server $port"); 3. When the connection is activated and the SSL dialog pops up, check the checkbox so that the message will not show again. 4. Disconnect the connection and reconnect it again without expanding any filters 5. In a minute, the connection would drop. The following messages are found in the dstore trace Wed Aug 11 12:53:53 EDT 2010: Read timed out java.net.SocketInputStream.read(SocketInputStream.java:155) java.io.BufferedInputStream.fill(BufferedInputStream.java:229) java.io.BufferedInputStream.read(BufferedInputStream.java:246) org.eclipse.dstore.internal.core.util.XMLparser.readLine(XMLparser.java:388) org.eclipse.dstore.internal.core.util.XMLparser.parseDocument(XMLparser.java:525 ) org.eclipse.dstore.core.util.Receiver.handleInput(Receiver.java:149) org.eclipse.dstore.core.util.Receiver.run(Receiver.java:121) Finished Tracing Note: if a filter is expanded or the SSL message pops up, the problem will not be seen.
I tried used the same arguments you provided for starting up a server manually but I was not able to reproduce this. I'll have to check but it's possible I'm not hitting this because there may be hidden subsystems in the workspace I'm using.
I can constantly reproduce the problem. Even with the running server launcher.
The key to reproduce the problem is not to expand the filter and check the check box on the SSL dialog so that the dialog not showing any more.
*** Bug 323261 has been marked as a duplicate of this bug. ***
This bug is the same as bug 323261 and is related to bug 246826 in that it's the timeout case where it takes longer than IO_SOCKET_READ_TIMEOUT to send the data.
Created attachment 177844 [details] patch to use read timeout while awaiting initial keepalive response Samuel, could you see if this patch works for you? Masao, could you see if this patch resolves the issue in bug 246826#c6?
Thank you for the patch, Dave. It works fine for me.
The patch does not solve the bug 246826#c6. I reopened it.
Since this solves Samuel's problem but not Masao's and Masao has reopened the other, I'll just put in the fix for Samuel's here. Kevin, could you review the patch? Thanks.
Review +.
Thanks for the review. I've committed the change to cvs.
