toasters

toasters@lists.teaparty.net

3 participants
13514 discussions

Re: Console Break (was Y2K...)
by Ketil Kirkerud Elgethun 23 Apr '99

23 Apr '99

jrg(a)acm.org (James R Grinter) writes: > For Solaris 2, I've been told that there exists modified serial driver > that - when the OS is running - will prevent that happening. This is in fact the standard serial/console driver in Solaris 2.6 (and 7). You'll have to set a variable in /etc/default/kbd - basically you have the option to either -react to BREAK (or L1-A, if you have a keyboard attached) - this is the default -ignore the BREAK (or L1-A). We're running quite a few (50ish) Suns, all connected to a linux box running conserver (serial hardware being Stallion cards). We've had no problems with this so far - and we _do_ have breaks enabled, to be able to reset a machine that won't let us in for some reason. Another thing : I'm thinking of attending the Netapp202 course (it fits nicely in with my plans for Usenix Tech.'99) - I've been running Netapps for a few years now, but wouldn't mind some more "inside" information on the secret bits. Is anyone willing to recommend this course ? Finally, a small tidbit for those among you who get a thrill out of (large) numbers: We've got an F540 (current uptime: 404 days) - this box has sustained 9900 (!) NFSops/sec for 20 hours. Id' say that on average it has sustained approx 3000 NFSops/sec over the last 6 months... NetApp needs to do something about the "total nfs ops counter" - it's rolled 15 times since the last reboot - nfsstat -l gives some really interesting output : [output snipped] 195.225.3.37 www2.sol.no NFSOPS = 1385724634 (4294966978%) :-) -- ---Ketil Kirkerud Elgethun, SOL System

2 1

Console Break (was Y2K...)
by scw＠seas.ucla.edu 23 Apr '99

23 Apr '99

> > I just wish there was a way to enable breaks from console. > > By "enable breaks" do you mean the "BREAK causes the OS to jump to the > PROM monitor" stuff that SunOS (including 5.x) does? > > We don't detect BREAK or respond to it, but there's probably no reason > we couldn't do that. ARGGG! NO! NO! NO! Disaster waiting to happen. A failed terminal, or one that is turned off or has its power line unplugged generates a BREAK -well strictly speaking it's not really a BREAK but most implimentations of UART see it as a break. This used to happen to our Admin folks every once in a while with their SUN. Remember the users with their torches and pitchforks. > I don't know if our serial port is set up so that unplugging a terminal > (or terminal server) and plugging it back in looks like a BREAK; that's > one of the suckages on at least some Suns. > ----- Stephen C. Woods; UCLA SEASnet; 2567 Boelter hall; LA CA 90095; (310)-825-8614 Finger for public key scw(a)cirrus.seas.ucla.edu,Internet mail:scw@SEAS.UCLA.EDU

4 4

Re: Console Break (was Y2K...)
by tkaczma＠gryf.net 23 Apr '99

23 Apr '99

On Fri, 23 Apr 1999 scw(a)seas.ucla.edu wrote: > Clearly you don't work at a university, where faculty members walk > into the machine room and grab any old piece of equipment that they want No, I work in an environment where dumb tubes are a greatly coveted pieces of equipment, since we don't have too many to go around. Still, pulling a serial cable doesn't cause the machine to halt in our environment. I don't care that you want this functionality disabled. If you do you'll never enable it in the variable. I have roughly 15 NACs in 6 different buildings, some of them 12 miles away. > In any case OUR machine room is protected against faculty intrusion, > and AIX doesn't have that problem anyway. But Sun's implementation of > BREAK detection is severly broken ( 0v != SPACE ) or is it MARK? Like I said, pulling the cable off doesn't do it, plugging the cable in doesn't do it. Turning the tube on while connected to a Sun does it. Hence, the tube is the broken piece not the Sun. Tom

1 0

size of NFS ops counters (was RE: Console Break (was Y2K...))
by Hackworth, Brian 23 Apr '99

23 Apr '99

: From: Ketil Kirkerud Elgethun [mailto:ketilk@a.sol.no] : NetApp needs to do something about the "total nfs ops counter" - it's : rolled 15 times since the last reboot - nfsstat -l gives some really : interesting output : : : [output snipped] : 195.225.3.37 www2.sol.no NFSOPS = 1385724634 (4294966978%) For the next major release of ONTAP, we'll be increasing the size of many of the internal counters, including the NFSOPS counters, to 64 bits. This should help for a while. - Brian Hackworth

1 0

RE: Y2K: Any one rolled a filer forward & backwards? (&courses)
by Page, Ruth 23 Apr '99

23 Apr '99

Page,> I don't see training revenue as NetApp's Page,> major benefit from customer training; Hitz's response and mine crossed yesterday, but we were consistent on this point... >Also, it seems like NA could reduce the cost of the >class by holding classes in Santa Clara instead of >shipping NetApps and trainers around the country. Most of our classes are at our corporate office. We hold field classes when/where we get customer requests. It is more expensive for NetApp, but as long as we have a minimum # of students, we can strike a balance of our costs and the savings for the customers (and the benefits of offering training to customers who will not come to California). The most cost effective training route in the field is by taking advantage of the multiple student discount, (new as of about Dec.) the web registration discount and the multiple class discount. It's pretty cost effective. We also offer classes at customer sites (minimum # of students required), which many customers are choosing. Ruth... -----Original Message----- From: Jay Soffian [mailto:jay@cimedia.com] Sent: Thursday, April 22, 1999 9:16 PM To: ruth.page(a)netapp.com Cc: mds(a)gbnet.net; beepy(a)netapp.com; toasters(a)mathworks.com Subject: RE: Y2K: Any one rolled a filer forward & backwards? (&courses) "Page," == Page, Ruth <ruth.page(a)netapp.com> writes: Page,> I don't see training revenue as NetApp's major benefit from Page,> customer training; rather (I see the benefit as) the Page,> customers' increased proficiency with the product. Page,> Residual benefits include fewer calls to TS, the likelihood Page,> of more purchases and referrals (result of customer Page,> satisfaction) and more face time with the customer for Page,> feedback/input. We just had this argument with our NA salesperson. If NetApp really believes this, why do they charge so much for the training courses? Why not include a coupon with each NetApp sold that is good toward one training class? Don't want to give a class away? The value of the coupon could be based on the cost of the filer purchased. NA should at least offer the course at something close to cost. Afterall, it benefits NA as much as it benefits the customer, as you just said. Also, it seems like NA could reduce the cost of the class by holding classes in Santa Clara instead of shipping NetApps and trainers around the country. Half the people that attended the class here in Atlanta recently had to fly in anyway, so it doesn't appear you are saving the customer much money by holding regional classes. j. -- Jay Soffian <jay(a)cimedia.com> UNIX Systems Administrator 404.572.1941 Cox Interactive Media

1 0

Re: Console Break (was Y2K...)
by jrg＠acm.org 23 Apr '99

23 Apr '99

On Fri 23 Apr, 1999, fubar(a)mathworks.com (Kendall Libby) wrote: >I wish Sun would implement that. I've got a bunch of sun's on a >terminal server and have never had a problem unless the cisco is >powered off. That's acceptable; all the sysadmins know it, and no >one else is allowed in the machine room (*cough*). Had a similar >situation when I worked for an ISP. (a bit off topic, but..) For Solaris 2, I've been told that there exists modified serial driver that - when the OS is running - will prevent that happening. Variously, I've been told that it's something Sun Consulting have available, if you ask, to it being something that a particular large Wall Street firm and big Sun (Solaris) user wrote. It switches the signal away from a break to a combination key set, I believe. Alternatively, it's possible to modify a serial cable to accomodate a long break but avoid the short one you get from the disconnection of a cable or death of a terminal server. A third option is to get a terminal server that doesn't do that - I've direct experience of the Lightwave ones (already mentioned here today) and they do prevent it happening if they are powered on or off. James.

1 0

RE: administrivia
by Robert Palermo 23 Apr '99

23 Apr '99

Hey Kendall, Maybe your body has been taken over by aliens and you only think you're running this list. Du'Oh !!! > MIME-Version: 1.0 > Content-Transfer-Encoding: 7bit > Date: Fri, 23 Apr 1999 13:18:13 -0400 (EDT) > From: fubar(a)mathworks.com (Kendall Libby) > To: "Kinney, Roger" <roger.kinney(a)netapp.com> > Cc: "'fubar(a)mathworks.com'" <fubar(a)mathworks.com>, toasters(a)mathworks.com > Subject: RE: administrivia > X-Bike: 95 VFR750F > X-Animal: waterbuffalo > X-Retort: Bite me, fanboy > > >>>>> "roger.kinney" == Kinney, Roger <roger.kinney(a)netapp.com> writes: > > roger.kinney> I believe that's unsubscribe dl-toasters <your email > roger.kinney> address> > > Um, no. I run the toasters list, and it's most definitely "toasters". > > K.

2 1

Re: Y2K: Any one rolled a filer forward & backwards?
by Jim McCoy 23 Apr '99

23 Apr '99

Jason Kelleher writes: >mwalters(a)netapp.com: >>ONTAP 5.3 has an optional (license required !) "snaprestore" feature [...] > > And if you don't have the license can you still call support and > have them walk you through the undocumented procedures to do this > from a boot floppy You really don't even need to boot floppy (search for the the "floppy-boot?" setting in the boot system) if you are in a hurry. The prev_cp works for 5.x as well but it is a bit of a hack to do this sort of a restore (OTOH I doubt that the "snaprestore" is anything more than a wrapper over prev_cp which masks out the checkpoints stored in the NVRAM log.) > they should've spent more money on books and less on leather jackets Come on, those were pretty sweet jackets! :) (and the logo was rather subdued for a corporate freebie.) Last I heard they decided to pocket the money they were spending on the jackets and the class docs are still very weak. They should also put the class docs onto the web somewhere because there are times when the docs aren't handy and I need to remember if it is 26/7 or 28/7 that fries my disks... jim mccoy Yahoo! Inc.

4 6

Dump core after running CIFS setup
by Mike Ball 23 Apr '99

23 Apr '99

Has anyone ever had a "Dump Core" while running CIFS Setup? We had deleted the C$ and recreated C$ to point to a qtree which we created on vol1. This is the only thing out of the ordinary which was done on the filer before rerunning CIFS Setup. I am currently running Data Ontap 5.12 on a F720 Thank you, Mike Ball

1 0

administrivia
by fubar＠mathworks.com 23 Apr '99

23 Apr '99

With the flurry of unsubscribe requests going to the list lately, I thought I'd take this opportunity to remind everyone how to get off this list correctly. *** If you ever want to remove yourself from this mailing list, you can send mail to <majordomo(a)mathworks.com> with the following command in the body of your email message: unsubscribe toasters <your email address> *** There is a list archive (searchable, even) available on the web at http://teaparty.mathworks.com:1999/toasters/ K.

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters