[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-discuss] data consistency on LVM snapshots using writeback cac
From: |
Daniel Neugebauer |
Subject: |
Re: [Qemu-discuss] data consistency on LVM snapshots using writeback caching |
Date: |
Mon, 15 Jul 2013 21:26:19 +0200 |
User-agent: |
Mozilla/5.0 (X11; U; Linux x86_64; de-DE; rv:1.8.1.23) Gecko/20091004 Thunderbird/2.0.0.23 Mnenhy/0.7.5.0 |
> I wonder how you handle RAID for your LV because in my experience LVM
> RAIDs are very slow in comparison to mdadm RAID. I have issues like you
> describe for guests running on LVM RAID (not using raw LV, though).
I'm using mdraid around LVM, so e.g. /dev/md1 (RAID1 with 2 disks) is
being used as a PV.
I never bothered to try LVM RAID1 as I am happy with mdraid and storing
metadata on a third disk appears to be mandatory but seemed like too
much overkill.
Alignment should be okay as well. All disks involved on those servers
have 4k sectors but I paid close attention while partitioning and
initializing mdraid and LVM, so I am pretty sure sectors are aligned.
Even if they wouldn't be it should not have such a serious impact,
should it? It's only a few kB every few seconds coming from guests and I
already have incredibly high latency so I doubt it could have anything
to do with sector alignment.
Daniel