Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I need to configure SQL or Windows 2008 Clustering to get HA with no packet loss.

I configure my SQLcluster, but when I restart one node, I lose one ping. I try to lose no ping so that my query will continue to execute with no timeout when the owner node cluster restarts.

Is it possible ?

share|improve this question

migrated from Nov 19 '09 at 3:37

This question came from our site for professional and enthusiast programmers.

it belongs to both... as a lot effort is both from programmer / dba and the server admin to have a good active active cluster setup. – Dani Nov 18 '09 at 21:00
for the client not to feel a node crush one of the solution is to have a service layer between the client and the database. the service might be or might not be aware of the cluster (and in case of failure submit the query again maybe ?) I am not sure what @Boivin ment about "my query continue to execute" – Dani Nov 18 '09 at 21:03
He means that even if a server is going down and a query is executing, it won't stop it from finishing. – esylvestre Nov 18 '09 at 21:26
up vote 4 down vote accepted

Clusters don't allow querries to continue to execute on failover. A cluster failover means the instance starts again from a cold state, runs pending recovery on all databases and puts them online on the new node. All clients are disconencted, all pending transactions rolled back, all running querries aborted. Clients have to reconnect and resume from the last commited transaction.

BTW this also applies to Mirroring. The only technology that abstracts transparently the failover event so that applications don't notice the incident is messaging based, eg. Service Broker, because all work is first commited locally and delivered asynchronously to the cluster/mirror, but this does not apply to the vast majority of applications.

share|improve this answer

do you have a secondary "heart beat" nic connecting the two servers ? the 2nd server need to know about the first server "problem" in order to "take over" and it means than it can't be on the milisec.... there might be 1 ping lost, if there is a heart bit cable between them, it might be detected earlier, also - if one server start processing a query and was shutdown (power) the query will be lost...

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.