allocate additional socket buffer memory but triggers
reclaiming pages) -> kernel fs driver (wedges because locks already held)
-Shaun
On Fri, Sep 16, 2022 at 4:01 PM Nikolaus Rath wrote:
> On Sep 14 2022, Shaun McDowell wrote:
> > It has been a few years since I've worked on it bu
Wanted to share a personal journey since myself and a partner went down a
similar route on a past project of using NBD + userland loopback + S3 for
block devices.
5 years ago we actually forked nbdkit and nbd-client to create async,
optimized versions for use privately as a userland loopback block
I've created a workable async framework with the latest nbdkit code and a
sample file plugin using aio to service requests. This allows for a single
worker thread to issue multiple requests in parallel via a callback
mechanism. Existing locking mechanisms are preserved and if an async thread
model
19, 2018 at 12:41:01PM -0500, Shaun McDowell wrote:
> [...]
>
> Thanks for the other email, I thought it was very interesting and I'm
> glad that people are looking at the performance of nbdkit seriously.
>
> > Our cbdkit was branched from 1.1.12 with some patches from
plugin_pread
plugin_pwrite
cbdkit_async_reply
cbdkit_async_reply_read
cbdkit_async_reply_error
On Fri, Jan 19, 2018 at 12:05 PM, Eric Blake wrote:
> On 01/19/2018 10:56 AM, Shaun McDowell wrote:
>
> > Limitation: The kernel will (with today's default settings) typically be
>
Hi,
We've been using a modified nbdkit (ours is cbdkit internally) for about
half a year now and since you guys appear to be working on a next version
of the API I wanted to go over some of the limitations we hit with nbdkit
that we think others may also hit for consideration into the batch of
cha
I went ahead and implemented 2 new hooks on the existing nbdkit_plugin
struct for async_pread and async_pwrite to get some testable numbers.
The test is set up with an all in memory block device (data being copied to
and from a std vector as simulated reading and writing).
Every read and write op
The concern is a client is blocked while processing a request. The nbdkit
server design requires a thread per request being processed regardless of
the number of connections or clients. We want to run 1000's of requests in
parallel without needing a thread at nbdkit layer per request in flight.
Ou
