Home Page › Forums › BizTalk 2004 – BizTalk 2010 › Memory leak on BizTalk server › Memory leak on BizTalk server
Hi Greg,
Thanks for the timely response. First of all, the following conditions were the ones that led me to conclude that the memory leak could be causing the BizTalk server from responding.
1. Performance graphs have indicated that the memory leak occurs once the BizTalk server has been restarted and continues till the next time it is restarted.
2. There are other Windows 2003 servers in the same environment and the memory graphs for them do not show any kind of leakage. If what your saying about the memory leakage is acceptable, then that trend should have been observed on all the other servers too. But that is not the case.
3. We do a BizTalk clean up of the databases and then reboot the server. On days when the server gets rebooted, the BizTalk server seems to have responded exceptionally well. And then when the same job is run for the next day we have experienced delays. We keep observing these delays till the day when BizTalk server is restarted again. On that day again the job response time was noted to be exceptionally well. Thats why we started attributing the memory issue to the responsiveness of the BizTalk application.
Also, with regards to the FILE adapter failures, I see the following message as a warning event in the event viewer logs of the BizTalk server.
\”The adapter failed to send the message to the send port \”D:\\AIApplicationHome\\BTSApplicationHome\\NICE\\temp\\requestinstance\\%MessageID%.xml\”. It will be resent after the retry interval specified for this send port.\” Event Id: 5786
The folder location is the place where we send the xml files to get stored. We initially thought that it could be a problem only with a particular xml file. So we tracked down the job and reran it again only to find that at the other instance the job ran fine without any FILE adpater warnings. We also observed that during the retry interval for this adapter failure, the FILE adapter just does not do any other processing even though there are other orchestrations which require it.
We arrived at the conclusion that the memory leakage was causing the FILE adapter failure which in turn led to the bad response ( or no-response) state of the system. We even found out a microsoft article which spoke of a memory leakage in the FILE adapter. The article can be seen at http://support.microsoft.com/?scid=kb;en-us;886226&spid=1444&sid=297
The article suggests that we use the latest service packs on the server. We did try that out and still found out that the issue was recurring.
Please let us know in case if we are missing anything or if our understanding is wrong.
Our sincere thanks for all the efforts.