From: Josef Bacik <jbacik@fb.com>
To: <axboe@kernel.dk>, <nbd-general@lists.sourceforge.net>,
<linux-block@vger.kernel.org>, <kernel-team@fb.com>
Subject: [PATCH 0/6] Lots of NBD fixes and enhancements
Date: Tue, 28 Feb 2017 11:57:05 -0500 [thread overview]
Message-ID: <1488301031-3199-1-git-send-email-jbacik@fb.com> (raw)
This is kind of a big batch of patches, but they all depend on eachother so it
was hard to tease out the fixes from the enhancements without making my life
miserable.
FIXES:
nbd: set queue timeout properly
nbd: handle ERESTARTSYS properly
The ERSTARTSYS one in particular is pretty awful as we can tear down a whole
device if a userspace app has a signal pending while submitting IO. This is big
and scary but I had a debug patch on top of this to randomly induce ERESTARTSYS
to make sure it was behaving properly.
ENHANCEMENTS:
1) Handle signle path failures gracefully. This is the first step to handling
reconnects gracefully, but for right now we can easily fall back on other
connections if we happen to lose one connection.
2) Ref counting and bdev change. This is in preparation for the netlink patch
and handling reconnects and the such better.
3) Netlink interface. Trying to add the nbd-control thing was controversial,
and I realized the more I wanted to do with monitoring and stuff I would need to
use netlink for anyway. With this new interface we can easily configure and
disconnect devices, dynamically add devices if we are past our pre-allocated
limit, and dynamically find free devices to use if we don't want to try and hunt
for a device. The userspace patch to utilize this is kind of rough but can be
found in my github tree
https://github.com/josefbacik/nbd
These have been pretty well tested, but I'd like to hear any thoughts on the new
interface. Thanks,
Josef
next reply other threads:[~2017-02-28 16:57 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-02-28 16:57 Josef Bacik [this message]
2017-02-28 16:57 ` [PATCH 1/6] nbd: handle single path failures gracefully Josef Bacik
2017-02-28 16:57 ` [PATCH 2/6] nbd: ref count the nbd device Josef Bacik
2017-02-28 16:57 ` [PATCH 3/6] nbd: stop using the bdev everywhere Josef Bacik
2017-02-28 16:57 ` [PATCH 4/6] nbd: set queue timeout properly Josef Bacik
2017-02-28 16:57 ` [PATCH 5/6] nbd: handle ERESTARTSYS properly Josef Bacik
2017-02-28 16:57 ` [PATCH 6/6] nbd: add a basic netlink interface Josef Bacik
2017-03-08 10:07 ` [Nbd] " Wouter Verhelst
2017-03-08 14:56 ` Josef Bacik
2017-03-08 20:44 ` Wouter Verhelst
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1488301031-3199-1-git-send-email-jbacik@fb.com \
--to=jbacik@fb.com \
--cc=axboe@kernel.dk \
--cc=kernel-team@fb.com \
--cc=linux-block@vger.kernel.org \
--cc=nbd-general@lists.sourceforge.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox