I have an F740 running ONTAP 6.2. Something odd seems to be happening to the value of "options ip.tcp.sack.enable" - it toggles spontaneously between "off" and "on" once an hour for no apparent reason! This seems to happen at the same time that /etc/registry is rewritten with a new "state.config.refresh.last" value.
Can anyone else running 6.2 or 6.2R1 confirm that they see the same effect?
[ For amusement only: I run a weekly job to check that all the options are what I expect and to report any differences. I noticed an oddity (now explained by the above) when I updated the reference values after upgrading 6.2 over 6 weeks ago, but didn't investigate at the time. The jobs always run at the same time of day, and so there have been an even number of toggles between them ... but the "once an hour" turns out to be more like "one hour plus one second", and the slow drift has apparently now caused there to be an odd number of toggles during the last week! This is what I guess from looking at /etc/registry in the snapshots, at least. ]
Chris Thompson Email: cet1@cam.ac.uk
On Sun, May 12, 2002 at 11:55:26PM +0100, you wrote:
I have an F740 running ONTAP 6.2. Something odd seems to be happening to the value of "options ip.tcp.sack.enable" - it toggles spontaneously between "off" and "on" once an hour for no apparent reason! This seems to happen at the same time that /etc/registry is rewritten with a new "state.config.refresh.last" value.
Can anyone else running 6.2 or 6.2R1 confirm that they see the same effect?
yes, I see this on our F740 (6.2): how curious!
Jonathan Hughes
jrh@dmu.ac.uk (Jonathan Hughes) writes:
On Sun, May 12, 2002 at 11:55:26PM +0100, you wrote:
I have an F740 running ONTAP 6.2. Something odd seems to be happening to the value of "options ip.tcp.sack.enable" - it toggles spontaneously between "off" and "on" once an hour for no apparent reason! This seems to happen at the same time that /etc/registry is rewritten with a new "state.config.refresh.last" value.
Can anyone else running 6.2 or 6.2R1 confirm that they see the same effect?
yes, I see this on our F740 (6.2): how curious!
Tim Thompson & Sreenivasa Potakamuri from NetApp have been in touch with me about this problem. There's now a bug number (71658) for it.
It can be cured by removing the line "options.ip.sack.enable=off" from /etc/registry.local on the filer. (NB no "tcp" in that name: the problem is caused in some way by the aliasing of ip.sack.enable and ip.tcp.sack.enable.)
Chris Thompson Email: cet1@cam.ac.uk