HTTP errors can be split into a few categories: client ephemeral, server
ephemeral, server permanent, and globally permanent. You could argue
there is even more permutations. However, for simplicity, these errors
can be viewed as ephemeral and permanent.
The Glance XenAPI plugin has been updated to raise a RetryableError for
ephemeral and unexpected (i.e. errors that are categorized as neither
permanent or ephemeral) errors.
Additionally, an instance fault will be logged every time an error
occurs. This will serve as transaction history where every attempt is a
transaction on the state. If an ephemeral error occurs, there is a
retry, then a permanent error, the history of each error will be in the
instance_faults.
Deployers should configure the num_retries relative to the number of
api_servers. Right now, servers are put in random order and then cycled.
Trying the same server multiple times could cause unnecessary load and
delays. However, trying multiple servers, is ideal when a single server
is behaving badly or cannot reach another server it needs to communicate
with to fulfill a request.
Closes-Bug: 1380776
Change-Id: I267a5b524c3ff8a28edf1a2285b77bb09049773c
cc2a34ce05