We've been told that a fully flex-vol'd aggregate can cause issues with guaranteed reservations due to metadata shortages.
Seems the aggregate will randomly turn off volume reservations if it has no room for aggregate-level metadata, though I'm not entirely certain of what that is made. We've been bitten by it three separate times that I know of, possibly more.
What really hurts is they recommend 5% overhead for this, but with a default 10TB aggregate that's 500GB of metadata overhead. Ouch.
I *think* the original concept for aggregate snapshots was part of Prospero, though I'm not aware that this went anywhere.
Jeff Kennedy QCT Engineering Compute 858-651-6592
-----Original Message----- From: owner-toasters@mathworks.com
[mailto:owner-toasters@mathworks.com]
On Behalf Of Paul Letta Sent: Thursday, January 31, 2008 7:25 AM To: toasters@mathworks.com Subject: Aggr's at 100% with FlexVols
Hi, I know when traditional vols are in use, you don't want to let them get too close to full because performance will be affected.
But what about using FlexVols and Aggrs ?
I have a few Aggrs that are at 100%. But the FlexVols in those Aggrs are mostly under 50%.
I had a NetApp support person tell me that its bad to have Aggrs at 100%, even if the FlexVols are not full.
What's the answer here. Does having Aggrs at 100% because the
FlexVols
contained in them are sized to fill the Aggr present a performance
issue
?
I would think that as long as the flexvols are not close to full, its
ok
to have the aggr's at 100%.
Thanks,
Paul