Description of problem:
The hosted engine storage is currently limited in many ways (can not take master, can not be the first storage in the system) causing issue is several system flows.
We want to make this storage into a regular storage in the system and have the minimal amount of HE specific limitations.
The considerations of load on this SD should be moved to documentation.
We should try to define what to note when planning disk placement on this storage to not cause engine downtime as part of this effort.
(In reply to Yaniv Dary from comment #1)
> We should try to define what to note when planning disk placement on this
> storage to not cause engine downtime as part of this effort.
HE's domain should be the least favored target for anything the system determines automatically (master domain, memory volumes, etc). Other than that, disk locations are determined by the user.
Moving to SLA. Once we implement the new deployment, we expect it to be a regular domain (with the exception of blocking moving it to maintenance from the Engine - a separate bug exists). This bug is a tracker(?) for all activities needed to remove the special treatments throughout the code.
Why was this moved to 4.3?
Because it is too late to make sure all flows work as expected in 4.2.0. We will start removing the restrictions for 4.3 and consider backporting all that are working back to 4.2.z.
For now closing.
We will remove more limitation as requested from customer.