Host restart failure
Our QA environment has seen the growth of Host instances to over 35 now. Host instances in a 2-server cluster.
Lately we have been unable to start services. When we try to start them - we get: "The service did not respond to the start or control request in a timely fashion".The account logon is a domain account and not the local system account.
The only way we are able to get services to start is to stop other host instances.
I saw this Hotfix - but its only supposed to be for MSMQ. We are not running any MSMQ adpaters in our env.
What is thepractical limit of the number of Host Instances that BizTalk allows without this kind of issues (theoretical I believe is unlimited)? How can we mitigate this scenario?
Hi,
I've never seen a good answer to this one, but I have thought about quite a lot. There will be performance issues with creating a large number of host instances, even if they are mostly idle. Each host instnace will poll the message box database repeatedly to dequeue messages (you can see this activity if you start up SQL Profiller).
Couple of thongs you can try:
Check the load on your message box database using the perf monitor tools, see if it looks like you have a problem there.
Analyse the reasons for creating so many host instances, is there a way that this can be optimized to use fewer.
Regards,
Alan