You miss that it could take 5 minutes, in the situation of a core dump.
This is a rather rare ocurrance however.
-----Original Message----- From: Bennett Todd [mailto:bet@rahul.net] Sent: Friday, July 21, 2000 10:29 AM To: Bruce Sterling Woodcock Cc: toasters@mathworks.com Subject: Re: Linux-NFS list recommended I post here -- RE: NetApp versus Linux
2000-07-20-16:25:44 Bruce Sterling Woodcock:
+++ NetApp (F720) Boots in under 2 minutes due to the use of NVRAM even after an improper shutdown. [...]
2 minutes is really unlikely unless you're just using vanilla NFS and haven't generated a core. Think more like 5 minutes (still impressive).
That is such a bummer!
I remember only a couple of years back, the proud boast was that you could kick the plug out of a netapp, and no more than 45 seconds after plugging it back in, it'd be live serving NFS data once again. I even remember a salescritter saying (I think on this list) that they'd had customers take that boast and try and feed it to 'em, with a line like ``Ok, let's try it, if it's not back in 45 seconds by my watch, you're outa here'', and stood by with confidence while the box passed the test.
When did the time increase to 2.6-6.6 times that once-proud 45-second figure?
At 5 minutes, that means Netapps are no longer booting significantly faster than generic servers.
-Bennett
2000-07-21-13:45:38 jeff.mohler@netapp.com:
You miss that it could take 5 minutes, in the situation of a core dump.
This is a rather rare ocurrance however.
So 5 minutes may be rare. But Bruce also said that 2 minutes is unlikely quick.
This is a radical change from the old line of 45 seconds is so reliable that account reps will happily bet a contract on it. A sad change, too.
-Bennett
I guess I've also suffered from too many crashes which dumped core, rebooted, then failed NVRAM replay, crash, reboot, then "Previous shutdown was dirty", dump core, reboot, then "Recomputing Parity", etc. :)
Bruce
-
Bennett Todd -
Bruce Sterling Woodcock -
jeff.mohler＠netapp.com