This thread looks to be a little on the old side and therefore may no longer be relevant. Please see if there is a newer thread on the subject and ensure you're using the most recent build of any software if your question regards a particular product.
This thread has been locked and is no longer accepting new posts, if you have a question regarding this topic please email us at support@mindscape.co.nz
|
Hi, I am currently working out a custom ConnectionStrategy and Interceptor to allow retries after transient connection failures etc. I have noticed that the commands dealing with KeyTable (allocating new IDs etc) do not go via the ConnectionStrategy, is it currently possible to catch the opening of these connections? The actual command execution goes via the Interceptor which is good, as long as it got a good connection to start with... Thanks Barry |
|
|
Hi Barry, Yes the calls associated with the KeyTable updates are handled in an independent transaction. I can have a look next week at having this managed by the connection strategy for you. Ill post an update once Ive had a look into this more.
|
|
|
Hi Barry, Ive had a look into this and while we can convert the KeyTable code over to use the connection strategy it will introduce a number of new issues with concurrency due to it re-using the same underlying connection as the UOW in the standard case so we are going to shelve this for the time being. After looking into this though Im interested around what your use case is - given that a new short lived connection and transaction is used for manipulating the KeyTable what would be your need for a connection strategy to manage this?
|
|
|
Hi Jeremy, We're in a cloud environment with replicated database (AWS / RDS / MySQL) and are trying to figure out how to deal as best we can with time outs, fail overs etc. In the case that creating the connection for the key table requests times out, we'd like to be able to use the same retry logic we do for our other queries. Does that make sense? Cheers ~ |
|
|
Gotcha. Ill have a bit more of a think about how we might be able to support this as that sounds pretty reasonable.
|
|