Re: CEPH/BSD status

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 28-9-2017 09:56, Mogamat Abrahams wrote:
> Hi,
> 
> Can we help you with testing/feedback on this one?

[ cc-ing this also to the ceph-dev list ]

Hi Mogamat,

Would be great. I'm currently rather occupied with my daytime job.
So the last few weeks haven't done much.
The FreeBSD report was rather late so some of the work has progressed a
bit. Yesterday 12.1.1 has been released. So my first concern is to build
and upgrade  a new luminous port package.

> Which of these are still open?
> 
>  1. Run integration tests to see if the FreeBSD daemons will work with a
>     Linux Ceph platform.

No serious testing work done. I have understood that a librbd on Mac
works against a Linux cluster. So I would expect it to work, but it
certainly needs to be given some thought on how to test the
interoperability of all components. And perhaps a few structural tests
need to be written and submitted.

>  2. Compile and test the userspace RBD (Rados Block Device). This
>     currently works but testing has been limitted.
>  3. Investigate and see if an in-kernel RBD device could be developed
>     akin to ggate.

These 2 have been done. there is a rbd-ggate in the package that allows
one to mount a RBD image on a device. Partition, format and just use as
regular FS. I'm using that, and it is not a wonder of blinding
performance (got slow testcluster hardware), but no problems.

>  4. Investigate the keystore, which can be embedded in the kernel on
>     Linux and currently prevents building Cephfs and some other parts.
>     The first question is whether it is really required, or only KRBD
>     requires it.

I have not yet found reasons to actually use this, but then I also do
not have a kernel-cephfs module. So that would need to be fixed/written
first. cephfs over fuse works, but that is with fuse and thus slow.

>  5. Scheduler information is not used at the moment, because the
>     schedulers work rather differently between Linux and FreeBSD. But at
>     a certain point in time, this will need some attention (in
>     src/common/Thread.cc).

Still open. And I do not have any information on how relevant this will
be for FreeBSD.

>  6. Improve the FreeBSD init scripts in the Ceph stack, both for testing
>     purposes and for running Ceph on production machines. Work on
>     ceph-disk and ceph-deploy to make it more FreeBSD- and ZFS-compatible.

Some work has been done ceph-disk zap can wipe a disk.
ceph-deploy has not had any attention at all, but I would expect most
basic users to take a book, of blog and then they'd need ceph-deploy.
I do all config work by hand, or use ceph-disk on pre-build ZFS volumes.

In the mean time Ceph-volume has entered the stage and has an lvm
component. Would be nice to have a matching ZFS component, perhaps even
working with ZVOLs.

>  7. Build a test cluster and start running some of the teuthology
>     integration tests on it. Teuthology wants to build its own libvirt
>     and that does not quite work with all the packages FreeBSD already
>     has in place. There are many details to work out here.

Started a few tries here, but still have not really goten anywhere.
Libvirt, Qemu and teuthology setup keep me running in circles.
It is needed to start running the more elaborate tests.

>  8. Design a vitual disk implementation that can be used with bhyve and
>     attached to an RBD image.

Still open. But would be a great thing for bhyve.
Now the problem/challange is that Ceph is GPL of one sort or the other,
and bhyve is FreeBSD-lic. And mixing is going to be hard since no GPL is
allowed in the FreeBSD base tree., so it needs to be done a package module.

But the module interface for different disk providers needs to be cut
out from the current bhyve code. If at all possible.
Then librbd-module package goes into ports and we have worked around the
GPL conflict.
OR
    Ceph needs to release librbd also under a more liberal license.
	which could be possible, but I do not know RH and/or ceph's
        possition on this.
OR
    It needs to maintained as a patch set to bhyve.
	But that will require continuous maintenance
	And will fast lead to bitrot

Hope this helps in getting the current status a bit more clear.
Alan Sommers is already working on getting AIO and bluestore working,
with promising outlook.

--WjW

> 
> -- 
> *Mogamat Abrahams*
> 
> */TAB IT Solutions cc/*
> Suite B, 2nd Floor, 16a Newmarket Street, Cape Town, 8000
> www.tabits.co.za
> 
> Cell: 0829238001
> Office: 0210071510

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux