DCOMdefficient
IS-IT--Management
Hello,
I hope this is a good group to post this one
I have been spending quite a few days trying to figure out why dcom stopped
working for no apparent reason. From my tests I have determined the problem
is in the OS. Tests include RPCPing and MS DCOM client/server test
application. I have also verified router ACLs at the facility level and have
worked with the developers. All dcomcnfg settings have been quadruple
checked - 3x.
When I stop and restart COM+ services dcom works for a few minutes and then
starts failing. Beginning yesterday it would start working for no apparent
reason for a few minutes and stop. The error reported by the app that uses
it is typically "The RPC server is unavailable.[800706ba]" The server had
been up only a couple of weeks when dcom inexplicably stopped working.
I have witnessed that the RPCPing enpoint search is successful when dcom
client/server test application fails
"CoCreateInstanceEx: The RPC server is unavailable<800706ba>"
Event logs will show the rather uninformative "DCOM was unable to
communicate with the computer a.b.c.d using any of the configured
protocols."
Here is a summary of the scenario:
Server1 - w2k3 hosting web sites. ms sql and running the web hosting
automation software (HC)
Server2 - w2k running remote DNS (SimpleDNS Plus)
Server3 - w2k running remote Email server (MailEnable)
HC uses dcom to communicate with Server2 and Server3 to provision the
various elements of web hosting accounts.
(Server2 used to serve the same purpose as Server1 until IIS became "full."
It communicated with Server3 in exactly the same way Server1 is supposed to
do now. Because server1 is not communicating I have switched back to server
2 as "primary" server).
What else can I look at or do? One of my theories is that for some reason
too many connections are created on port 135. Netstat doesn't seem to say
that but maybe I am wrong.
I hope this is a good group to post this one
I have been spending quite a few days trying to figure out why dcom stopped
working for no apparent reason. From my tests I have determined the problem
is in the OS. Tests include RPCPing and MS DCOM client/server test
application. I have also verified router ACLs at the facility level and have
worked with the developers. All dcomcnfg settings have been quadruple
checked - 3x.
When I stop and restart COM+ services dcom works for a few minutes and then
starts failing. Beginning yesterday it would start working for no apparent
reason for a few minutes and stop. The error reported by the app that uses
it is typically "The RPC server is unavailable.[800706ba]" The server had
been up only a couple of weeks when dcom inexplicably stopped working.
I have witnessed that the RPCPing enpoint search is successful when dcom
client/server test application fails
"CoCreateInstanceEx: The RPC server is unavailable<800706ba>"
Event logs will show the rather uninformative "DCOM was unable to
communicate with the computer a.b.c.d using any of the configured
protocols."
Here is a summary of the scenario:
Server1 - w2k3 hosting web sites. ms sql and running the web hosting
automation software (HC)
Server2 - w2k running remote DNS (SimpleDNS Plus)
Server3 - w2k running remote Email server (MailEnable)
HC uses dcom to communicate with Server2 and Server3 to provision the
various elements of web hosting accounts.
(Server2 used to serve the same purpose as Server1 until IIS became "full."
It communicated with Server3 in exactly the same way Server1 is supposed to
do now. Because server1 is not communicating I have switched back to server
2 as "primary" server).
What else can I look at or do? One of my theories is that for some reason
too many connections are created on port 135. Netstat doesn't seem to say
that but maybe I am wrong.