andybrucenet
12-08-2011, 01:48 PM
I have a problem that occurs regularly when Windows VMs restart - some of the point-to-point IP addresses within the Windows appliance are "broken" (cannot connect). I can reproduce this real easily simply by restarting multiple applications. The applications (and all the components) are running OK, it is just that the internal IP addresses generated by AppLogic for the point-to-point interfaces do not work.
Normally, if I stop the "broken" applications and restart serially, the problem goes away.
I actually have the problem up and running right now. I have a Windows VM that connects to a "db" appliance; this is simply a proxy appliance that connects to the real SQL Server database on my external network. I use that approach to avoid putting in IP addresses and to prepare for the day when I might want to run the SQL Server database within the AppLogic application. What I am observing is that a "ping db" from an affected Windows VM within the application works but that opening the the SQL Server port does not work (as though a firewall were blocking the port). Changing the running program in question to use the explicit IP address of the external SQL server, however, works fine.
On three other application instances that are identical to the failing application - the "db" appliance proxy is working fine. And, normally, when I restart the AppLogic application, the problem goes away.
Anyone else seen point-to-point connection failures within an AppLogic application between Windows components? I'm seeing this both with Windows -> Windows and Windows -> CentOS connections.
I'm happy to provide more details, but I do wonder first of all if anyone else is running into this problem.
Normally, if I stop the "broken" applications and restart serially, the problem goes away.
I actually have the problem up and running right now. I have a Windows VM that connects to a "db" appliance; this is simply a proxy appliance that connects to the real SQL Server database on my external network. I use that approach to avoid putting in IP addresses and to prepare for the day when I might want to run the SQL Server database within the AppLogic application. What I am observing is that a "ping db" from an affected Windows VM within the application works but that opening the the SQL Server port does not work (as though a firewall were blocking the port). Changing the running program in question to use the explicit IP address of the external SQL server, however, works fine.
On three other application instances that are identical to the failing application - the "db" appliance proxy is working fine. And, normally, when I restart the AppLogic application, the problem goes away.
Anyone else seen point-to-point connection failures within an AppLogic application between Windows components? I'm seeing this both with Windows -> Windows and Windows -> CentOS connections.
I'm happy to provide more details, but I do wonder first of all if anyone else is running into this problem.