VBucket Aware clients HA Question
Sat, 03/03/2012 - 22:45
So I was looking at all the docs on the various clients, and it appears there might be an HA limitation. All the clients take in 1 url for the VBucket Aware clients. However how do these handle when the server goes down. The most I can see is that as long as the client was already connected it knows about the other servers, but what happens if this is a new connection being made?
Also why do the clients all require /pools/default being added on to the URL? Looking at the code for the Python client it just drops it during the parse, but yet still requires it being there.
>> The most I can see is that as long as the client was already connected it knows about the other servers, but what happens if this is a new connection being made?
I know the java sdk will allow you to give a list of addresses to connect to so that if the first address fails you can try the next one in the list. I'm not sure how the other sdk's function, but if they don't allow you to do this then it is probably something we should add. Which sdk are you using?
>> Also why do the clients all require /pools/default being added on to the URL?
I can't remember the exact details here as to why we did this, but originally we thought we would have different ways of being able to access the rest api and therefore we needed to add the /pools/default. This idea was dropped so now having this extra stuff in the url isn't important any more, but hasn't been removed. I know my answer is kind of vague. If you would like more details I can find the exact reason for you. Just let me know.