Actually that sounds very similar, at first this affected only a few
Qtrees & Home folders on the two volumes. Toward the end of yesterday
afternoon I thought the filer had sorted itself out, this morning it
started again around 09:30Hrs and was affecting pretty much all our
qtreees & home folders. The problem became apparent yesterday around
13:00Hrs and dropped off around 16:30Hrs, starting again around 09:30hrs
this morning but worse! So there may be some correlation between filer
workload and this problem.
What DFM was reporting and initially alerted one of the guys to there
being a problem, was that about a dozen Qtrees were, according to DFM,
set at 50Mb. The quotas file had them set to 31.5Gb. When a asked the
filer for a quota report it itially told me that the quotas were set to
50Mb as well, after a quota off followed by a quota on I again queried
the filer for a quota report, this time it reported the quotas correctly
BUT still users couldn't write to their Qtrees or Home folders. I even
forced the cluster partner to take over the disks and rebooted the filer
having the problem but that dind't resolve it either.
Iain
-----Original Message-----
From: Adam McDougall [mailto:mcdouga9@egr.msu.edu]
Sent: 15 March 2005 21:29
To: Iain Barnetson
Cc: toasters(a)mathworks.com
Subject: Re: Quotas
An issue that may be related that I noticed, if I perform a quota
resize, it silently drops random users to a default user quota I have
specified.
If I run quota resize again, it complains about the default user entry.
Quotas remain confused like this until I perform a full quota off/on,
then quotas seem to operate perfectly fine for me.
On Tue, Mar 15, 2005 at 06:17:15PM -0000, Iain Barnetson wrote:
Of two filers updated to Data OnTap 7.0.1 Sunday morning, one has now
got a Quotas problem.
Users are getting disk full errors when trying to write to Qtrees or
Home folders despite have copious amounts of free space. The volumes
and
disks have space. The Qtrees are on one traditional volume and the
Home
folders on another traditional volume. When the quotas are off there's
no problem.
Have checked the filer is reading the quotas file correctly by doing a
rdfile /etc/quotas and comparing to the physical quotas file and it's
exact.
Have tried switching on and off the quotas, and also rebooting the
filer
(it works for MS boxes!)
Have a ticket open with NetApp but so far no resolution.
Anyone experienced anything similar or any ideas on a resolution?
Iain
T: +44 1224 776913
E: iain.barnetson(a)halliburton.com