During a publish cycle one of my data-centres is eventually failing transactions indicating they are not deployed.
Publishing a couple of hundred items is OK. When I push in a few hundred (I want around 200 thousand eventually!) then the success stops to increment and all transactions stop in waiting for deployment
until they eventually timeout.
Finer details:
- I'm using the
MIRROR
strategy in deployment. - Mirror_1 is a simple master and single worker
- Mirror_2 is an AWS Load-balancer endpoint
- The AWS load balancer points to two masters, with two workers
Now - and this may be very related: I did have to move a Finalize="true"
from the deployer config as I had the error noted below:
This strikes me as very odd given the XSD would appear to support it:
QUESTION:
Why can I not include the Finalize="true"
attribute?