terça-feira, 28 de setembro de 2010 21:47
I have a test environment of two standalone FW servers and one FW array server assigned to an EMS.
I applied update 1 to all. All had SP1.
All seemed to work after the Update 1 was applied in a couple of days I noticed that the server configured to the EMS was no longer creating reports. After some investigation I noted high CPU utilization. The highest service would go back and forth between the FW service and the SQL server service. I also noticed that logfiles were not being created instead was getting LLQ files created.Tried a number of things then tried reinstalling update 1. that seems to have fixed the problem. After a few minutes the LLQ file disapeared and the log files were created. CPU has settled down. I will see if reports are created tomorrow. I believe they will and I will report either way.
Todas as Respostas
quarta-feira, 29 de setembro de 2010 21:21
Still broken. will report back when I find a solution. Seems to be going from bad to worse. Seeing repetive access attempts to MS MRS site.
tried a number of things found on this forum. then uninstall TMG and reinstalled. Report worked this am. then there is the "but". The continuous access to the MRS started again. Thinking this volume of SSL accesses is overloading the reports process. If I stop the URL filter the MRS access stop. If I restart then the accesses resume. for the volume of access see line below that represents a 4pm to 1:30am next morn on a test server that no one is pointed to.
If no thoughts from the forum I'll call support.1SSL-tunnel1,161,10399.7 %
10/6/2010 - More testing. uninstalled tmg and rebuilt TMG thru sp1 install. Did not install update1. Server will now live and produce reports. SSL to MS MRS site still higher that other two TMG server 1200 request vs 130 and 160 a day. I believe the high connectivity to the MRS site is the problem. If I apply update1 the the ssl access to the MRS site goes up a maganitude and ultimately overloads the logs and the report summary fails in the evening. Please bare in mind this is observation and some amount of guess work.
I will continue to debug and report back