Boston Linux & UNIX was originally founded in 1994 as part of The Boston Computer Society. We meet on the third Wednesday of each month at the Massachusetts Institute of Technology, in Building E51.

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Discuss] Cloud-backup solutions for Linux?



On Mon, Sep 28, 2015 at 8:24 AM, Edward Ned Harvey (blu)
<blu at nedharvey.com> wrote:
>> From: Bill Bogstad [mailto:bogstad at pobox.com]
>>
>> > 2- Use a snapshotting filesystem like btrfs or zfs in the host, so the host can
>> replicate the guest storage to another location seamlessly.
>>
>> I don't see how this can work in a way that would be useful.
>> Filesystem snapshots of your emulated disk images by the host OS may
>> give you a single point in time copy, but they don't guarantee that
>> the copy is in any way consistent.
>
> This is one of my favorite modes of operation. I run a ZFS host, and have guest VM's residing in zvol's, which get snapshotted and replicated periodically to additional attached storage, and offsite and offline.
>
> If something happens, like the whole machine explodes or whatever, then I restore the guest snapshot, and power it on. The behavior of the guest machine is exactly as if the guest machine had been running and then suddenly the guest power was yanked or kernel panic or something. The guest storage device is a precise snapshot of what the guest storage would have looked like at the instant that the storage snapshot occurred.

Well, we are on the same page on that anyway.   It is exactly like
pulling the plug on a running system.

> If you're running an OS or some daemon that can't survive a power interruption, time to find a new OS or switch to a different daemon.

While some OSes/filesystems handle power interruption well at this
point, it seems to me that there are lots of apps/servers which do not
and which people still need to use.   Particularly in a VM environment
where you might be running legacy OS/app combinations because you
can't replace them, it seems to me that suggesting this method as a
generic way to backup VMs is not really appropriate.   Sure we should
all replace our old software systems with ones that use transactions
to protect against this kind of failure, but I don't think we are
there yet.

Bill Bogstad



BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org