Mailing List Archive: Re: Service Usage information in Autosupport e-mails (fwd)

I got this reply to my query. Thanks to Britt for the fast reply with this
information.

--
Dylan Northrup
northrup@loudcloud.com
Unix System Administrator
EDS Automated Operations

---------- Forwarded message ----------
Date: Wed, 16 Oct 2002 10:19:25 -0400 (EDT)
From: Britt Bolen <bolen@netapp.com>
To: Dylan Northrup <northrup@loudcloud.com>
Subject: Re: Service Usage information in Autosupport e-mails

On Tue, 15 Oct 2002, Dylan Northrup wrote:

> I've tried looking around and haven't been able to find what the precise
> information is that's contained in the 'SERVICE USAGE' section of a netapp's
> auto-support e-mail that filers send out. Here's some example output so you
> guys know what I'm talking about:
>
> ===== SERVICE USAGE =====
> Service statistics as of Sun Oct 6 08:46:15 GMT 2002
> System (UP). First recorded (72944725) on Wed Jun 14 02:20:50 GMT 2000
> P 12, 3626258, 3620527, Wed Jul 26 13:17:09 GMT 2000
> U 2, 25042, 18089, Tue Oct 1 20:36:50 GMT 2002
> NFS (UP). First recorded (72944721) on Wed Jun 14 02:20:54 GMT 2000
> P 12, 5922405, 5916260, Tue Aug 22 02:59:22 GMT 2000
> U 2, 25118, 18130, Tue Oct 1 20:37:31 GMT 2002

This is the condensed output of the availtime command.

Try running 'availtime full' (with a wide terminal!) it gives a more
verbose output.

Basically it tells you how many planned (P) and unplanned (U) downtimes
the filer has had, how long those downtimes lasted, and when and how long
the longest down time was.

feel free to forward my response back to the list.

Britt

--
Britt Bolen, bolen@netapp.com, Sledgehammer Engineering, block head.

"Aah, for the days when aviation was a gentleman's pursuit - back before
every Joe Sweatsock could wedge himself behind a lunch tray and jet off
to Raleigh-Durham." --- Sideshow Bob

northrup@loudcloud.com (Dylan Northrup) writes
>
> I got this reply to my query. Thanks to Britt for the fast reply with this
> information.
>
[...]
>
> This is the condensed output of the availtime command.
>
> Try running 'availtime full' (with a wide terminal!) it gives a more
> verbose output.
>
> Basically it tells you how many planned (P) and unplanned (U) downtimes
> the filer has had, how long those downtimes lasted, and when and how long
> the longest down time was.

Ah ha! YAUC [Yet Another Undocumented Command]

Now at last I understand what that mysterious file /etc/.avail is for. One
can even guess roughly how it is encoded by comparing it with the "availtime"
command output.

Due to a small [umm...] accident this file got removed on my filer early
this year, and didn't reappear until the next reboot. I imagine that is
why "availtime" thinks we had nearly 42 days of unplanned downtime then. :-)

Chris Thompson
Email: cet1@cam.ac.uk