My project uses HttpClient 2.0, which has frozen on a number of occasions bringing production servers down. I've submitted bug 30514 against this. However, I need to get this production issue resolved quickly, and would like some help regarding possible solutions / workarounds in the interim.
My current thought is to check for the state of the connections in the pool before attempting to call into it. The only way I have to do this is to check the current number of connections in use vs. the pool size. However, this is not 100% accurate. The quickest solution for us is to handle this outside the HttpClient code, so is there a way we can do the following externally: 1.Check the true state of the pool / connections in the pool (to see if it's frozen.) 2. If the cleaner thread is frozen (we don't know if this is the problem), can we restart it? 3. What are the performance implications in using a SimpleHttpConnectionManager and creating a new HttpClient per request? 4. Do you know if the pool will eventually free up again? For the longer term, do you have an ETA to the turnaround time for this bug? Thanks for your help, Neil Thier :: Trilogy :: work: 512-897-5074 This electronic transmission (and any attached document) is for the sole use of the individual(s) or entity(s) to whom it is addressed. It is Trilogy Confidential. If you receive this message in error, please notify me, and destroy the attached message (and all attached documents), immediately.
