I've found checking out the messages file on a routine basis a good idea. Helps find bad disks esh's or other things that have gone to the dark side.
And if you have enough filers, I find dfm to be a useful system. Unless you have over 300 filers.
-Blake
On 1/11/07, Glenn Walker ggwalker@mindspring.com wrote:
I'd recommend the scheduled WAFL reallocates to help defrag.
I'm curious about the fact they question the last time it was wafl_ironed - this looks for and fixes corruption, not poor layout.
I'd start by monitoring performance - stats output\sysstat and statit. Perfstat (from the now site) is a good tool to collect a lot of data, but unless you've been working on filers for a number of years, it will be mostly gibberish (though some output is similar to what you'd find on unix systems).
Glenn
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Nick Silkey Sent: Thursday, January 11, 2007 5:11 PM To: toasters@mathworks.com Subject: Preventative/proactive filer maintenance
Greetings,
I inherited a production F810 some time ago. It has since been replaced with a newer 3020C. During the 810->3020 data migration, I had severe NDMP enumeration latency. NetApp support inquired about various details on the 810, such as the last time it was wafl_ironed, etc. Indeed, the 810 had spun in production for years with minimal admin intervention. Apparently this contributed to the performance hit we took during our data migration.
My question to the list is what are some 'best practices' for caring-and-feeding of a 7.x filer, ala preventative steps to take within OnTAP in order to prevent said problem from happening again in the future. Preferably steps to take on a live filer without interrupting service (impacting read/write/throughput ok, downed filer not).
Much appreciated.
-- Nick Silkey | silkey@ece.utexas.edu Senior Operating Systems Specialist Electrical & Computer Engineering The University of Texas at Austin ENS 526W | 512.475.8284 | 0x35EB31E2