thread starving in MultiThreadedHttpConnectionManager
-----------------------------------------------------
Key: HTTPCLIENT-675
URL: https://issues.apache.org/jira/browse/HTTPCLIENT-675
Project: HttpComponents HttpClient
Issue Type: Bug
Components: HttpConn
Affects Versions: 3.1 RC1
Reporter: Roland Weber
Assignee: Roland Weber
Fix For: 3.1 Final
Hi folks,
I might have found a bug in MTHCM. It has to do with removing
HostConnectionPool instances that have no more connections in them. That was a
fix for a memory leak we previously had. There are two cases where the pools
get deleted. One is in handleLostConnection: (excerpt)
...
if (hostPool.numConnections == 0) mapHosts.remove(config);
notifyWaitingThread(config);
...
Could this delete a pool in which there is still a thread waiting to get a
connection? If so, the thread would remain in the global pool. But even if it
is interrupted there, it would still use the old HostConnectionPool in which no
connection will ever become available again.
I suggest to change the removal check in both cases to:
if ((hostPool.numConnections < 1) && hostPool.waitingThreads.isEmpty)
What do you think?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
