[Aoetools-discuss] Query on AoE initiator accessing target started locally over lo

[Nivin L. <niv...@gm...>]

Hi Experts,

      Could you please help with your comments for the below.



1)      Is all the code for AoE initiator/target located @
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/
linux.git/tree/drivers/block/aoe <http://%2522/>?


2)      Is it possible to cleanly stop the AoE target, ensuring the cache
is written to disk and any cleanup needed is done?


3)      I have a use case where I need to use AoE over network initially,
but later stop the remote target, restart target locally on loopback
interface. The storage that is accessed in both cases is the same, so no
issues with data being in sync. I have a few clarification requests in this
regard:

a.       Wanted to understand if the AoE model has specific optimizations
done to make this as close as possible to a regular local disk access.

b.      I see the Linux storage stack @ https://upload.wikimedia.org/
wikipedia/commons/3/30/IO_stack_of_the_Linux_kernel.svg <http://%2522/>, I
was wondering about the overhead involved in accessing storage from
application POV once I restart the AoE target locally.

                                                              i.      Overhead
from the perspective of additional LOC involved as compared to a local disk
access

                                                            ii.      Overhead
from the perspective of throughput possible as compared to a local disk
access

c.       Do you think further optimizations are theoretically possible to
take care of this specific scenario to make it just like a regular local
disk access?

4)      Is there a way to pause the initiator from trying to communicate
with the target once it is decided that the target has to be shut and
restarted locally?

a.       I would like to avoid having to drop and retry in this specific
case as it’s a controlled operation. I understand this might not be a use
case generally.

b.      Would be ideal if we can flush any outstanding requests, stop
further requests from being sent and queue it up, receive all the responses
so that there are no outstanding requests from initiator POV. This would be
a good point to shut down the remote AoE target and restart target locally.


5)      For use of AoE in production, I was wondering if there would be any
kind of metric which would help with details on the stability and
availability expectations and any feedback from current usage in production.



Thanks,

Nivin