[pve-devel] Default cache mode for VM hard drives

[Stanislav German-Evtushenko]

On Wed, Nov 19, 2014 at 9:16 AM, Dietmar Maurer <dietmar at proxmox.com> wrote:

> > >>2. It is often happen to Linux VM swap space.
> > mmm,good point, I don't known how flush is working with swap.
>
> Swap space is inititialized on startup using 'swapon', and AFAIK there
> is no complex format which gets damaged if data is not synced? At least
> I never observed problems with swap after server crash


I can think about two cases at least when this is critical:
1) Live migration - if swap is corrupted live migration will fail.
2) DRBD consistency check - each time when the check find out-of-sync
blocks you would need to calculate if they belong to swap or data disk.
Also you will need to do disconnect/reconnect with shutting down one of
DRBD nodes in case of master-master in order to resync and reset "oos"
counter.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.proxmox.com/pipermail/pve-devel/attachments/20141119/62e7c61c/attachment.htm>