[zfs-discuss] L2ARC and SLOG on HW RAID with writeback cache

Gordan Bobic gordan.bobic at gmail.com
Fri Apr 27 15:29:13 EDT 2018


On Fri, Apr 27, 2018 at 8:18 PM, Gionatan Danti via zfs-discuss <
zfs-discuss at list.zfsonlinux.org> wrote:

> Hi all,
> given a system with a proper, writeback-enabled RAID controller which can
> both create a classical array on some disks *and* passthrough some other
> disks, is it a good idea to place L2ARC and SLOG on the HW array?
>

If you have write-caching enabled, why not just skip SLOG all together? If
all your disks are hooked up to the same controller, all of your writes
will be rendered asynchronous anyway.


>
> My reasoning is:
> - using CentOS I would like to avoid ZFS on root, so I need some sort of
> mirrored system partition (ie: MDRAID or HW array);
>

Why? I have been running ZFS roon on CentOS for quite some years now.


> - an HW RAID1 mirror provide resiliency and very easy replacement of a
> failed disk (ie: disk replacement is completely transparent both for the OS
> and the L2ARC/SLOG);
>

Why do you think it's easier to replace a failed disk in hardware RAID than
using zfs replace?


> - a writeback-enabled RAID card give very low write latency, which is
> ideal for SLOG;

- data stay on SLOG for a short time only and are read on emergency
> recovery only, meaning the probability of an unrecoverable read error is
> very low;
>

Are you using a different controller for your bulk data? If not, what do
you think SLOG will get you that the writeback cache on the controller
won't get you without it?


> - by using SSD for the system disks, I can partition them for L2ARC use;
> - any failed/uncorrect L2ARC read will be trasparently redirected to the
> main pool;
> - the main pool has full ZFS protection against bit root and the likes.
>
> On the other hand, by using a software RAID1 array for the OS partition
> only, presenting the other SSD partitions directly to ZFS, I can stripe the
> L2ARC partition for added cache space (which is the solution I am commonly
> adopting). Moreover, I trust MDRAID way more the HW blackboxes.
>

Disable writeback cache, let ZFS take care of the lot, and never look back.
Using hardware RAID might seem appealing but you are going to regret it the
first moment you get bitten by bit rot that the hardware RAID didn't
protect you from.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.zfsonlinux.org/pipermail/zfs-discuss/attachments/20180427/1f005939/attachment-0001.html>


More information about the zfs-discuss mailing list