linux-lvm.redhat.com archive mirror
 help / color / mirror / Atom feed
* [linux-lvm] snapshots documentation clarification
@ 2007-02-27 23:58 Dane Miller
  2007-02-28  1:04 ` Mike Snitzer
  2007-02-28  1:13 ` Dane Miller
  0 siblings, 2 replies; 3+ messages in thread
From: Dane Miller @ 2007-02-27 23:58 UTC (permalink / raw)
  To: linux-lvm

afternoon collector
	lvm reading
	GS update: single auth db
	check ebay

	bmi virtual machine
	zimbra schedule downtime: snapshot conversion
	zimbra security updates

Hi,

Section 3.8 "Snapshots" of the LVM Howto has a scary warning after the first paragraph (http://tldp.org/HOWTO/LVM-HOWTO/snapshotintro.html), with a caveat in the last sentence:
  "If the snapshot size equals the origin size, it will never overflow."

What exactly does this mean?  Does this imply that it's safe to use "long-term" snapshots whose size is equal to the origin's size, without worrying about effects of significant divergence from the original?  Long-term to me is 3yrs as the root fs of a (Xen domU) mail server.

Note that in this case, writes to the root fs will be limited to logs (/var/log) and software updates.  Mailboxes live on a separate, non-snapshot logical volume.

Thanks,

Dane

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [linux-lvm] snapshots documentation clarification
  2007-02-27 23:58 [linux-lvm] snapshots documentation clarification Dane Miller
@ 2007-02-28  1:04 ` Mike Snitzer
  2007-02-28  1:13 ` Dane Miller
  1 sibling, 0 replies; 3+ messages in thread
From: Mike Snitzer @ 2007-02-28  1:04 UTC (permalink / raw)
  To: linux-lvm

On 2/27/07, Dane Miller <dane@greatschools.net> wrote:
>
> Section 3.8 "Snapshots" of the LVM Howto has a scary warning after the first paragraph (http://tldp.org/HOWTO/LVM-HOWTO/snapshotintro.html), with a caveat in the last sentence:
>   "If the snapshot size equals the origin size, it will never overflow."
>
> What exactly does this mean?

It means there isn't any risk of the snapshot LV getting invalidated
because of overflow.  Reason being, the copy-on-write would only ever
fill (but _not_ overflow) the snapshot LV if the entire origin LV
changed.  Even if that were the case you wouldn't overflow because the
snap LV would accomodate such extensive change.

That said, I can't see how a long-term (3yr) _single_ snapshot is
useful but thats up to you.  Generally snapshots are done periodically
rather than maintaining a single snap for such a long period.  Do you
understand that LVM snapshots need a new snapshot LV for each snapshot
you create?  So if you want N snapshots, without the risk of overflow,
during the 3yr period you'd consume N*(size_of_origin_LV) space.

Mike

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [linux-lvm] snapshots documentation clarification
  2007-02-27 23:58 [linux-lvm] snapshots documentation clarification Dane Miller
  2007-02-28  1:04 ` Mike Snitzer
@ 2007-02-28  1:13 ` Dane Miller
  1 sibling, 0 replies; 3+ messages in thread
From: Dane Miller @ 2007-02-28  1:13 UTC (permalink / raw)
  To: LVM general discussion and development

Dane Miller wrote:
> afternoon collector
> 	lvm reading
> 	GS update: single auth db
> 	check ebay
> 
> 	bmi virtual machine
> 	zimbra schedule downtime: snapshot conversion
> 	zimbra security updates

oops.  cut-'n-paste snafu here.  now the world knows my todo list for
this afternoon :-\  sorry...

Dane

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2007-02-28  1:13 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-02-27 23:58 [linux-lvm] snapshots documentation clarification Dane Miller
2007-02-28  1:04 ` Mike Snitzer
2007-02-28  1:13 ` Dane Miller

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).