You are not logged in.
Pages: 1
I have a Samsung 830 and 840EVO
1x partition on each @ 200G (matched start and end sectors), remainder unpartitioned.
Partitions added to array on md0
The array is purely for VM Images, very few, very large files.
What filesystem would best be suited to this?
I was looking at XFS with a 64k block size, but not sure how to specify stride and stripe.
The Arch Wiki shows an example for ext4 on RAID0, but I don't know how to convert this to XFS (or other FS) syntax.
Am I on the right track to an optimal setup for the intended use?
Thanks! :-)
Offline
I was looking at XFS with a 64k block size, but not sure how to specify stride and stripe.
The Arch Wiki shows an example for ext4 on RAID0, but I don't know how to convert this to XFS (or other FS) syntax.
From ArchWiki XFS:
The default values already used are optimised for best performance in the first place. mkfs.xfs will detect the difference between single disk and MD/DM RAID setups and change the default values it uses to configure the filesystem appropriately.
Word of caution: RAID0 multiplies the chance of an SSD failure causing total loss. Personally, I'd bypass RAID/filesystem entirely and go with libvirt storage pools on LVM.
Last edited by sabroad (2022-10-12 12:06:09)
--
saint_abroad
Offline
JustSomeGeek wrote:I was looking at XFS with a 64k block size, but not sure how to specify stride and stripe.
The Arch Wiki shows an example for ext4 on RAID0, but I don't know how to convert this to XFS (or other FS) syntax.From ArchWiki XFS:
ArchWiki wrote:The default values already used are optimised for best performance in the first place. mkfs.xfs will detect the difference between single disk and MD/DM RAID setups and change the default values it uses to configure the filesystem appropriately.
Word of caution: RAID0 multiplies the chance of an SSD failure causing total loss. Personally, I'd bypass RAID/filesystem entirely and go with libvirt storage pools on LVM.
I'm wary of letting anything doing it's thing automatically. Part of the reason for using Arch in the first place :-)
As i'm using a large (512k) block size, on a (relatively) small storage device, with a specific use in mind, I wanted to do things manually, so at least it's my mistakes that are at fault when things don't work. Thanks for the XFS link though. After ages spent Googling, I never thought to come back to the Arch Wiki. Lesson learned.
And i'm good with RAID0 for this. It's only for my gaming and media VM's, with important stuff stored elsewhere. Nothing I can't afford to lose. Thanks though.
Offline
So, according to:
https://wiki.archlinux.org/title/XFS#St … _and_width
XFS *sometimes* detects RAID Geometry, and there's a (dead) link to calculating correct sunit values to specify manually.
Does anyone have any experience of this, and of using XFS on 2xSATA SSD's in RAID?
Is the maximum block size of 65526 going to cause any issues?
Any other recommendations?
Thanks!
Offline
There are some guides around, e.g. a quick google search gave https://ewen.mcneill.gen.nz/blog/entry/ … derations/
But - there's almost no point, if you're not a server where every tiny bit of performance counts you probably never notice any difference at all...
And RAIDs aren't static anyway. You can grow RAID and then whatever geometry you had before no longer matches the new situation.
So don't worry too much about it.
Last edited by frostschutz (2022-10-14 16:17:44)
Online
Pages: 1