From: Goldwyn Rodrigues <rgoldwyn@suse.de>
To: Peter Kieser <peter@kieser.ca>, neilb@suse.de
Cc: lzhong@suse.com, john@stoffel.org, linux-raid@vger.kernel.org
Subject: Re: [PATCH 00/24] Clustered MD RAID1
Date: Fri, 26 Dec 2014 16:49:10 -0600 [thread overview]
Message-ID: <549DE5E6.2060208@suse.de> (raw)
In-Reply-To: <549DE299.3010205@kieser.ca>
Hi Peter,
On 12/26/2014 04:35 PM, Peter Kieser wrote:
>
> On 2014-12-23 9:24 AM, Goldwyn Rodrigues wrote:
>> This md device can be used as a regular shared device. There are no
>> restrictions on the type of filesystem or LVM you can use, as long as
>> you observe clustering rules of using a shared device.
>
> This is very vague. I feel like this is a huge layering violation. LVM
> or filesystem on top of mdadm should not have to be aware of what is
> going on below.
How is this a layering violation? I said there are _no_ restrictions as
long you follow the clustering rules of using a shared device. This
device can be used as a regular device with the filesystem being unaware
of how the data is handled underneath. The only thing the md device
ensures is that the data (between mirrors) is consistent for the cluster.
By clustering rules I mean rules such as:
- Use a clustered filesystem
- Use cLVM (as opposed to usual LVM) on top of MD-RAID.
- Use local filesystems in active/passive mode
> Otherwise, I feel like other clustered filesystems that
> include RAID are better suited.
such as?
> As it stands, what are the benefits of
> using this over $CLUSTER_FS?
This is not a cluster filesystem. This is a RAID which is made cluster
aware for your shared storage. This provides redundancy at the device
level for a clustered filesystem. Your cluster shared storage is a
single point of failure (in case hardware-based raid is not used) and
this software-based solution avoids making it a single point of failure.
An example usage scenario (not limited to) could be using it across two
iSCSI devices to ensure higher availability.
--
Goldwyn
next prev parent reply other threads:[~2014-12-26 22:49 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-18 16:14 [PATCH 00/24] Clustered MD RAID1 Goldwyn Rodrigues
2014-12-23 17:24 ` Goldwyn Rodrigues
2014-12-23 21:47 ` John Stoffel
2014-12-26 22:35 ` Peter Kieser
2014-12-26 22:49 ` Goldwyn Rodrigues [this message]
2014-12-27 0:47 ` Marcus
2015-02-06 2:39 ` NeilBrown
2015-02-10 17:00 ` Goldwyn Rodrigues
2015-02-11 4:17 ` NeilBrown
2015-02-11 17:25 ` Goldwyn Rodrigues
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=549DE5E6.2060208@suse.de \
--to=rgoldwyn@suse.de \
--cc=john@stoffel.org \
--cc=linux-raid@vger.kernel.org \
--cc=lzhong@suse.com \
--cc=neilb@suse.de \
--cc=peter@kieser.ca \
/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;
as well as URLs for NNTP newsgroup(s).